# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
Windows traditionally limits many file operations to `MAX_PATH` (260
characters). Deeply nested plugin packages can exceed this limit during
`.whl` extraction, leaving files missing and causing the plugin to fail
to load.
This change adds a shared extended_path helper that prepares paths for
Windows file operations using the extended-length prefixes `\\?\` or
`\\?\UNC\`. The helper is used for ZIP archive I/O and extraction
destinations, including plugin wheel extraction. This avoids relying on
the Windows long-path setting, which changes machine-wide settings.
Paths remain unprefixed outside file operations. Code that bypasses the
shared ZIP helpers may still require separate long-path handling.
See [Microsoft's documentation on path length
limitation](https://learn.microsoft.com/en-us/windows/win32/fileio/maximum-file-path-limitation?tabs=registry)
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
Tested on Windows 11 by using a portable installation of OrcaSlicer and
placing it in some really long filepath length e.g.
`C:/somename/<240 characters long folder path>/OrcaSlicer` and then
running and trying to install any plugin locally or on OrcaCloud.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
# Description
Opening the texture picker decoded every texture in the expanded
sections on the GUI thread in a single frame, each one twice at full
size. This stalled the UI on large libraries, and the full-size decodes
stayed in memory afterwards.
- Picker thumbnails are loaded lazily: only tiles inside the visible
scroll area are considered, at most 2 uncached decodes per frame, and
the rest fill in over the following frames. Deferred tiles draw as a
placeholder.
- Those follow-up frames are requested with `request_extra_frame()`.
Marking the canvas dirty from inside the render did not schedule another
idle pass, so thumbnails waited for the next mouse event.
- The selected texture is still centred on open, even when it is off
screen. The scroll reveal runs before the visibility check so that tile
is loaded on the next frame.
- Library validation no longer decodes pixels. The new
`png::is_decodable_8bit_png()` reads only the PNG header and accepts
8-bit grey or RGB/RGBA, which is what `decode_png()` and
`decode_colored_png()` take.
- Picker thumbnails are decoded without entering the decode cache
(`decode_height_texture(layer, keep_decoded = false)`). Only the 128 px
thumbnail is kept per library entry; the full-size decode is made
transiently and freed.
Tradeoff: the header-only check does not read pixel data, so a PNG with
a valid header but corrupt or truncated pixel data now passes validation
and fails when the texture is loaded. The previous full decode caught
that case.
Each thumbnail decode still runs at full size on the GUI thread, one at
a time. Peak memory per decode is unchanged; a streaming downscale would
be needed to reduce it.
# Screenshots/Recordings/Graphs
Old:
https://github.com/user-attachments/assets/03038532-396f-48c2-898c-aa06ef83e5f5
New:
https://github.com/user-attachments/assets/49fdc31a-ce1b-470a-88a7-4c6319f03b81
## Tests
Not yet run. Suggested checks:
- Open the texture picker with a library of many textures, and do not
move the mouse: tiles should fill in over a few frames without a visible
stall.
- Open the picker with a layer whose texture is far down the list: it
should be centred and loaded.
- Import a valid 8-bit grey and an RGB PNG, plus an unsupported one
(16-bit, palette, or grey-alpha): only the supported ones should be
accepted as-is, and the rest converted on load.
- Watch memory while scrolling through a large colour library: it should
not grow by the full decoded size of every texture shown.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Marking the canvas dirty from inside the render does not schedule another
idle pass, so deferred thumbnails waited for the next mouse event. Asking
for an extra frame does.
The thumbnail is downscaled on upload, but the full decode it was made
from stayed in the decode cache for as long as the library entry lived.
Thumbnail decodes now skip the cache, so only the 128 px thumbnail is
kept per library texture.
* Fix TabCtrl clipping titles and indicator dots
Button measures a label plus icon without the icon gap that render draws,
so icon tabs were ellipsized and the indicator dot was cut off. Widen the
horizontal padding so the measured width covers the gap.
* Re-layout publish dialog when extruder switch labels change
Setting the indicator bullet on a switch option changed its width without
laying out the page, leaving the switch distorted until the next layout.
With ooze prevention on, an extruder that has no layers left is set to 0 °C when it is parked, instead of its standby temperature. On IMEX printers where several filaments share one heater, the heater is only turned off once every filament on it is done.
The clearance preview shown while dragging sorted instances with a comparator that is not a strict weak ordering, which aborts with libc++ debug hardening. The preview now uses object-list order, with each object's copies in their last validated print order, which is the order print validation uses. The height check is now in SequentialPrintClearance.hpp, with unit tests.
Opening the texture picker decoded every texture in the expanded sections
on the GUI thread in one frame, each one twice at full size. Only the
visible tiles now load, at most a couple per frame, and the rest fill in
over the following frames.
The library validation no longer decodes the pixels: a header-only check,
is_decodable_8bit_png(), accepts exactly what the two decoders take.
# Description
Adds a fifth Python plugin capability type, `ResourceTexture`, which
contributes
height-map textures to the Texture Displacement picker. A plugin returns
a
directory of images; they appear as selectable textures in the picker,
grouped
under one collapsible section per pack. Disabling or uninstalling a pack
removes
them again, and projects that already use one of its textures keep
working,
because a layer holds its own image bytes.
**The plugin-facing contract** is one method:
```python
class WoodPack(orca.resource_texture.ResourceTexturePluginCapabilityBase):
def get_name(self):
return "Wood Pack"
def get_texture_dir(self):
return os.path.join(os.path.dirname(__file__), "textures")
```
The host scans one non-recursive level of that directory, accepting the
same
formats a hand-picked image already accepts (`.png/.jpg/.jpeg/.bmp`),
and decodes
them through the same `load_texture_image_data()` the picker uses for
the user's
own images, so a pack texture and a hand-picked one behave identically.
**Distribution is `.whl` only for v1.** `install_plugin` copies exactly
one
file, so a `.py` plugin cannot ship sibling images; a wheel is a single
file and
the loader already extracts it on demand.
**Files of interest**
- `src/slic3r/plugin/pluginTypes/resourceTexture/` — the capability, its
trampoline and its pybind base, mirroring the Pages type.
- `src/slic3r/plugin/host/PluginResourceTextures.{hpp,cpp}` — host
registry,
free of wx and GUI includes so it can sit beside `PluginManager`.
- `src/slic3r/GUI/TextureLibrary.{hpp,cpp}` — the third texture source,
the
`pack_name` provenance on an entry, and a generation counter.
- `src/slic3r/GUI/Gizmos/GLGizmoTextureDisplacement.cpp` — the picker.
- `src/slic3r/plugin/PluginFsUtils.{hpp,cpp}` —
`collect_resource_texture_files()`.
**Behaviour change worth a reviewer's attention:** the picker popup is
now a
fixed size that scrolls rather than sizing itself to its content, and it
gains
one collapsible section per pack. The sizing applies to every user, not
only
those with packs installed. With no packs installed the picker otherwise
behaves
exactly as before.
**Not in scope for v1:** folder or `.zip` install for non-programmers,
and
preset-referenceable pack textures (which would need a plugin-backed
`ConfigOptionDef::plugin_type`).
# Screenshots/Recordings/Graphs
https://github.com/user-attachments/assets/b96787cf-b475-4698-826d-5db8b9f386f1
## Tests
New suite `tests/slic3rutils/test_plugin_resource_texture.cpp`, 19
cases:
- Validator: a file inside the root is listed; a parent-directory escape
and an
absolute directory outside the root are refused; a symlink pointing
outside is
skipped; files in a subdirectory are not listed; non-regular entries are
skipped; the extension whitelist is case-insensitive and rejects
near-misses;
the listing stops at 64 files; a missing directory yields an error.
- Listing and lifecycle: a pack's images appear and clear on disable and
unload;
an empty pack contributes nothing; two distinctly named packs each stay
one
contiguous run; a second pack declaring a loaded capability name is
refused;
reinstalling with different bytes at the same path changes the decoded
content.
- Hostile overrides: `get_texture_dir()` returning `data_dir()` or
`resources_dir()` is refused and the pack loads with no textures; a
raising
override still loads the pack.
- Decoding: PNG and BMP from a pack decode through the same path as the
same
files picked by hand.
- Persistence: a project keeps a pack texture after the pack is
uninstalled,
bytes unchanged across a re-save.
- Wheel metadata: a minimal resource-texture wheel is accepted.
Known gaps, for transparency: the wheel path is covered only at the
metadata
gate, not through a full extraction-and-import lifecycle; there is no
test for a
registry callback after interpreter shutdown, as there is no seam for it
without
reaching into `src/`; and `.jpg` decode parity is covered by the listing
rather
than by a real fixture, since no JPEG encoder is available to synthesise
one.
Manually exercised in-app on Linux with a locally built pack wheel (five
colour
PNGs): install, enable, textures listed under the pack's own section,
then
disable and uninstall remove them. The new suite itself has not been run
locally — CI is the first execution.
* Linux: only disable WebKit compositing for WebKitGTK older than 2.42
The default Wayland path set WEBKIT_DISABLE_COMPOSITING_MODE=1 whenever the
session had both DISPLAY and WAYLAND_DISPLAY, i.e. any Wayland session with
XWayland. That forces WebKitGTK accelerated compositing off and also disables
threaded scrolling, so the Device tab web view (Mainsail/Fluidd) lagged unless
the app was launched with GDK_BACKEND=x11, the only path that skipped the
variable. Disabling compositing also turns off threaded scrolling.
The XWayland compositing bug the fallback worked around is fixed in WebKitGTK
>= 2.42, so apply it only for older WebKit. WebView::NeedsCompositingWorkaround()
reads the runtime WebKitGTK version; the X11 opt-in path is unchanged.
Behavior is unchanged on WebKit < 2.42, on pure X11, and on native Wayland
without XWayland. The wxWidgets skill reference is updated to match.
* Linux: use WebKit's shared-memory transport for the NVIDIA DMABUF renderer
The Device tab web view is slow on NVIDIA proprietary drivers even with
accelerated compositing left on: since the DMABUF renderer became the
default in WebKitGTK 2.42, the driver cannot back it, and web views fall
back to a slow path. Gating the old WEBKIT_DISABLE_COMPOSITING_MODE
fallback on WebKit < 2.42 is the wrong axis, because the failure is
version-independent from 2.42 on.
Force the shared-memory transport with WEBKIT_DMABUF_RENDERER_FORCE_SHM=1.
That keeps compositing and threaded scrolling and skips only the hardware
path. Also set WEBKIT_FORCE_DMABUF_RENDERER=1, which clears the isNVIDIA()
early return that Debian/Ubuntu/openSUSE patch into
rendererBufferTransportMode() and which otherwise returns before the
FORCE_SHM check is reached.
Both are set before the X11-vs-Wayland split so the workaround also covers
XWayland, and only when WebKitGTK is new enough to honour FORCE_SHM
(>= 2.44, WebView::SupportsForceSharedMemory()).
The AppImage launcher exported the superseded
WEBKIT_DISABLE_DMABUF_RENDERER=1, which is read first, leaves the transport
set empty and so disables compositing outright; replace it with the same
pair. The wxWidgets skill reference is updated to match.
* Linux: drop the AppImage WebKit DMABUF exports; correct the skill notes
The launcher cannot gate on the runtime WebKitGTK version, so exporting
WEBKIT_DMABUF_RENDERER_FORCE_SHM/WEBKIT_FORCE_DMABUF_RENDERER there would
clear a distro isNVIDIA() guard on hosts older than 2.44, where FORCE_SHM
is ignored, re-exposing the broken hardware path. The binary already sets
both with the correct gate, so remove them and say who owns the workaround.
Also correct two claims in the skill reference. An empty transport set only
means "no accelerated compositing" from 2.44 on; 2.42/2.43 still fall back
to the Wayland/X11 renderers, removed in the same release that added
FORCE_SHM. And threaded scrolling is off by default in 2.44/2.45 (on from
2.46), so FORCE_SHM keeps compositing rather than scrolling. Document the
module-level detection cost and the WEBKIT_DMABUF_RENDERER_FORCE_SHM=0
opt-out.
* Flatpak: force WebKit's shared-memory transport instead of disabling DMABUF
The entrypoint exported WEBKIT_DISABLE_DMABUF_RENDERER=1 for NVIDIA. WebKit
reads that variable first and leaves the transport set empty, which turns
accelerated compositing off, so the Device tab still lagged for NVIDIA
Flatpak users, and the binary could not override it because its setenv is
non-replacing.
The runtime's WebKitGTK is vanilla, so there is no distro isNVIDIA() guard
to clear and WEBKIT_DMABUF_RENDERER_FORCE_SHM=1 alone is the right switch:
compositing stays on and only the hardware DMABUF path is skipped.
* Flatpak: drop the WebKit NVIDIA export; the binary owns the workaround
/proc/driver/nvidia/version is readable inside the Flatpak sandbox, so the
binary's NVIDIA branch already sets WEBKIT_DMABUF_RENDERER_FORCE_SHM with
the correct runtime gate and the entrypoint's own export is redundant.
Verified with:
flatpak run --command=cat com.orcaslicer.OrcaSlicer /proc/driver/nvidia/version
which prints the NVRM version (driver 615.71.09) from inside the sandbox.
* Linux: say what the XWayland compositing gate actually tests
DISPLAY and WAYLAND_DISPLAY are both set in a native Wayland session as
well, because GNOME and KDE run Xwayland as part of the session and export
DISPLAY for it whether or not this process is an X client. The pair means
XWayland is available in the session, not that GTK picked X11 for this
process, so "(an XWayland session)" reads as more than the condition
establishes.
Say what the pair means, and name the case the fallback actually covers:
the X11 backend inside a Wayland session, which is what the forced
GDK_BACKEND=x11 above produces in a build without EGL.
Comment only; behaviour is unchanged.
* Linux: apply the XWayland compositing fallback only where GTK picks X11
The default path set WEBKIT_DISABLE_COMPOSITING_MODE whenever DISPLAY and
WAYLAND_DISPLAY were both set, but a native Wayland app inherits both, so a
build with EGL disabled accelerated compositing — and with it threaded
scrolling — while running on Wayland, where the XWayland bug does not apply.
The only default-path case that really runs through XWayland is the forced
GDK_BACKEND=x11 used when wx lacks EGL, so set the variable there. The
backend cannot be re-checked later: that branch exports GDK_BACKEND itself,
so the force and a user's GDK_BACKEND=x11 opt-in, which deliberately keeps
compositing, would be indistinguishable.
Behaviour is unchanged for WebKitGTK >= 2.42, where the gate already skipped
the variable, and for builds without EGL. A build with EGL on WebKitGTK <
2.42 in a Wayland session no longer loses compositing and threaded scrolling.
The wxWidgets skill reference is updated to match.
---------
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
* Fix unreadable text in wiki and video guide tooltips
* Use a keyboard icon for the gizmo shortcut button
* Fix blank tooltips on the Move and Scale gizmo combo boxes
* Fix external spool gradient and multicolor rendering
* Display AMS gradients and multicolors in color popup
* Preserve AMS gradient selection through color popup
* Match popup selections by full filament color data
* Align AMS Lite gradient geometry with BambuStudio
* Align AMS color palette with BambuStudio presets
* Match Bambu circular color rendering without background masks
* Preserve AMS gradient colors when applying filament settings
* Match AMS preview gradient swatches to the solid color shape
The texture picker's Download from OrcaCloud shortcut asks for a specific page. Re-running the
session handoff to reach it flashed the sign-in interstitial every time; instead an already-open
pane is navigated straight to the page URL, keeping its session. The handoff is still used on the
first open, where a session has to be established.
set_cloud_mode() gains an optional page_url for this. The homepage nav passes none, so its ordinary
re-open keeps the page the pane already holds (guide §8/§10).
* Let Z contouring slice objects that have several instances
The contouring step refused any object with more than one instance. The
instances of a PrintObject differ only by their XY shift, so the mesh is now
placed with the first instance of the PrintObject being contoured.
* Say in the Z contouring comment how instances map to print objects
* Check that Z contouring follows each instance as it is turned
The texture picker lists every plugin pack as an indented, collapsible child of one
"Resource Texture" section, which now always renders so the new button is reachable even
before any pack is installed. A pack whose name carries a " by ..." suffix shows that part
on a smaller, dimmed second line, keeping the name line short.
The button switches to the Home tab and opens the OrcaCloud pane on the plugin hub filtered
to resource textures. A native open now also drives the homepage shell (nav highlight and the
cloud-mode layout), and the requested page is remembered so a sign-in re-seed keeps it.
Also fixes the PluginsDialog fuzzy-search script path, which pointed at a directory the file
is not in.
Space opened the Speed Dial and Ctrl+P the Preferences while the text tool's
box or a sketch value field was being typed into. Those fields are ImGui drawn
on the GL canvas, so the canvas is the focused wx window and focus_keeps_space()
cannot see them; the CHAR_HOOK handler took the chord first and the field never
saw the key.
Skip the global shortcut dispatch while ImGui reports that one of its items owns
the keyboard and the focused window is not a wxControl.
# Description
Adds School House Spools PLA and PETG to the global Orca Filament
Library. Both are standard PLA and PETG, so Bambu PLA Basic / PETG Basic
served as the baseline for tuning values. The one intentional difference
is PETG nozzle temperature (252°C vs Bambu's inherited 255°C default),
based on our own testing.
# Screenshots/Recordings/Graphs
N/A
## Tests
Ran scripts/assign_vendor_setting_ids.py to generate setting_id values,
and scripts/orca_extra_profile_check.py --vendor="OrcaFilamentLibrary"
--check-filaments --check-materials to validate. No new errors or
warnings introduced (pre-existing unrelated errors in other vendors'
profiles were not touched).
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
* Define validated color picker selection contract
* Integrate web color picker with gradient and alpha controls
Co-authored-by: yw4z <ywsyildiz@gmail.com>
* Apply shared Orca theme to web color picker
* Host color picker in a validated WebView dialog
* Persist color picker favorites in user preferences
* Integrate localized web color picker with filament actions
* Add color picker browser regressions and verify direct includes
* Fit color picker content in a borderless floating panel
* Anchor color picker panels to their triggering controls
* Use accent foreground for selected color picker switches
* Integrate web color picker into generic color fields
* Use web color picker for AMS custom colors
* Use shared web color picker from Linux sidebar menu
* Use web color picker when adding texture import material
* Preserve alpha in filament color picker actions
* Clean up color picker asset whitespace
* Use application translations for color picker validation
* Preserve existing formatting in color picker integration
* Match SpeedDial rounded window handling on macOS
* Align color picker host behavior with SpeedDial
* Consolidate color picker implementation into its dialog
* Remove native color picker fallback
* Limit color picker interfaces to production consumers
* Use shared WebView translation field name
* Reapply anchored color picker placement after GTK show
* update
* fix increment buttons not working on spin inputs
* Use thumb color from its own channel
* init gradients with shifted color for second color
* fix hsl sliders cannot move freely and lock downs other one
* unify save button and color preview & increase stored color count to 32
* switch to hsl based calculation and fix plot and slider drag events
* update
* add support for multi colored gradients
* Update test_filament_color_picker.cpp
* add filtering color palette and store sidebar and combo state
* open dialog with calculated size
* apply shadow to dialog on windows
* add tooltip for gradient add color button
* Restore draggable fixed-size color picker frame
* Fix color picker sidebar sizing in WebKit
* Fix invisible default filament color button on macOS
* Remove duplicated color defs for color picker
* Remove duplicate HTML color picker frame
* Match color picker tests to current theme roles
* update icons
* use focus color on background of filter box
* align gradient switch with hex input
* fix initial window sizing
* Preserve caller color order when drawing filament icons
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
The wizard's loading page waits for HandleStudio to be called with
userguide_profile_load_finish and jumps to its target page on it; if the
call never arrives, a fallback timer in the page jumps after 180 s.
GuideFrame started the profile load from the first wxEVT_WEBVIEW_NAVIGATED
event and ran HandleStudio in the page as soon as the data was ready. That
event is the navigation's commit, before the page's scripts have run, and
with the wizard's derived cache the data is ready about 100 ms later.
Since the web view is created with an empty URL so the translated strings
can be injected before the first load, the first navigated event is the
commit of about:blank, so the data is ready before the real page has even
begun loading and the call reaches a page without HandleStudio. WebKit
drops the error and nothing retries, so on macOS the wizard often sits on
"Loading..." for three minutes before the printer list appears.
The loading page now reports in from its OnInit, the way the other wizard
pages request what they need, and the host sends the message once both
the data is ready and the page has reported; every instance of the page
reports again, so one that replaced an earlier one gets its own message.
Fix Move gizmo alignment icon scaling
Apply the native Windows DPI factor when sizing Move gizmo alignment icons.
This matches the sizing path used by other 16 px gizmo controls,
including support paint, seam, fuzzy-skin, and MMU painting tools.
`ninja-build` and `webkit2gtk4-devel` no longer exist on Tumbleweed. Use `ninja`, and use `webkitgtk3-devel` for WebKit. It is the GTK3 build and ships the `webkit2gtk-4.1.pc` that `src/slic3r/CMakeLists.txt` requires.
The `; first_layer_bed_temperature` and `; first_layer_temperature` compatibility lines were always taken from filament 1. Anycubic's GoKlipper reads the bed line to heat the bed, so a plate printed only with another filament got the wrong temperature.
The bed line now follows `bed_temperature_formula` (highest used filament or first printing filament). The nozzle line uses the first printing filament. Both values are computed before the by-object loop.
* Remove unused CMake modules
FindAlienWx.cmake and FindPerlEmbed.cmake came over from the Slic3r Perl
build. Nothing calls find_package(AlienWx) or find_package(PerlEmbed).
LibFindMacros.cmake was only included by those two modules.
bin2h.cmake lost its last caller when the avrdude sources were removed
in e6fcbfae07.
* ci: build changes under cmake/
The Build all path filters did not include cmake/, so a change to a
find module or to ClangClShowIncludes.cmake ran no build on a pull
request or on a push to main.
* fix: pin the Python that builds CPython on Windows
CPython's build needs a Python to run its own build scripts. Without a local
Python 3.10-3.13, find_python.bat installs the newest pythonx86 from nuget.
That became 3.15.0 on 2026-10-09, and 3.15 removed
SourceFileLoader.load_module, which 3.12's makeopcodetargets.py calls:
AttributeError: 'SourceFileLoader' object has no attribute 'load_module'
Download Python 3.13.16 from nuget.org as its own dependency, with a hash
(pythonarm64 on ARM64 hosts, pythonx86 elsewhere), and pass it to PCbuild as
PYTHON, which find_python.bat takes before any other Python. PYTHONHOME and
PYTHONPATH are dropped from that step, since PCbuild runs the scripts
without -E. DEP_PYTHON_HOST points at an existing Python 3.10-3.13 instead,
for machines that cannot reach nuget.org.
Drop 01-windows-nuget.patch. It fixed the dead nuget.exe download link in
find_python.bat, and that fallback is never reached now. Existing deps trees
keep working: with no patch command, ExternalProject skips the patch step
instead of re-applying the patch to patched sources.
* fetch externals separately
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
* Option: Show/ Hide inactive plates
* Hide inactive plate volumes by bounds
Switch inactive-plate volume hiding to use geometric overlap with the current plate bounding box instead of object and wipe tower ID mapping. This also guards against a missing current plate and keeps any active volume visible when its transformed bounds overlap the current plate footprint.
* Hide inactive plates per instance and keep them unpickable
Decide what is shown per object instance, from its footprint over the
current plate, so the parts of an object are never split; a selected
instance stays visible.
Keep the hidden state between frames instead of restoring it after each
one. GLVolume::on_hidden_plate records what this update hid, so it only
shows that again and leaves the volumes gizmos hide alone, and the raycasts
made while handling the mouse see the same scene as the frame. Nothing runs
while the inactive plates are shown.
Deactivate the raycasters of the other plates, whose icons still took
clicks where nothing was drawn, move the axes to the current plate, and
limit the option to Prepare, as Preview already shows the current plate
alone.
* Follow the selected plate with the camera while the others are hidden
Selecting a plate in Prepare while the inactive plates are hidden moves
the camera target over it and keeps the angle, distance and height, so
the plate comes into view where the previous one was. A zoom to all
plates, as after loading a project or adding a plate, frames the current
plate instead.
* Add a compact plate switcher while inactive plates are hidden
The hidden plates can no longer be clicked, and the object list gets
crowded on large projects. While they are hidden in Prepare, a column of
numbered buttons in the top left corner, a compact counterpart of the
Preview plate toolbar, selects the plate to show. It highlights the
current plate, dims the empty ones and gives the plate name in the
tooltip.
* Simplify the inactive plates update and the plate switcher layout
Decide and apply each volume's state in one pass instead of filling a
per-volume vector first, and hoist the switcher's loop-invariant colors.
Keep the height the bottom-left corner takes next to the constants it is
made of, with the canvas toolbar's button spacing named, rather than
repeating that layout in the switcher.
* Rename the option to Focus mode
The canvas menu entry reads Focus mode, lit while the other plates are
hidden, and is off by default, so nothing changes until it is turned on.
The app config key is focus_mode. Turning it on zooms to the current
plate, so the plate fills the view.
* Show plate thumbnails in the plate switcher
The switcher shows the thumbnails of the Preview plate toolbar instead of
numbered buttons, with rounded corners concentric with the window's.
Plater::plate_toolbar_image_version() lets it refresh them without
clearing the Preview toolbar's dirty flag.
* Don't draw the sinking contours of inactive volumes
GLVolume::render() skips an inactive volume, but its sinking contour was
still drawn, so an object on a hidden plate left its outline behind.
* Show the plate switcher's scrollbar only on hover, as in Preview
The plate switcher now shares the Preview plate toolbar's scrollbar: the
same width and colors, shown only while the mouse is over the column or
drags the bar, with the mouse wheel scrolling it either way, and tighter
window corners while it shows. Both toolbars draw it through the same
helpers, so they cannot drift apart; the Preview toolbar behaves as before.
Drop the frame style the switcher no longer uses since it draws
thumbnails instead of buttons.
* Make the plate switcher's thumbnails cheaper and match the Preview toolbar
The thumbnail version moves after almost every edit, and the switcher
redrew both thumbnails of every plate each time. It now redraws only the
current plate's, the only one edited while the others are hidden, and the
other plates' once their thumbnails are invalidated, through the new
Plater::update_plate_thumbnails().
IMToolbarItem::set_thumbnail() uploads a plate thumbnail without keeping
a copy of its pixels, a megabyte per plate, and replaces the item setup
the Preview toolbar and the switcher each repeated. The switcher's hover
border takes the Preview toolbar's color.
---------
Co-authored-by: Ian Bassi <ian.bassi@outlook.com>
* Export G-code statistics as JSON
Add an export of a plate's G-code statistics for scripts and
spreadsheets that compare prints or settings: print time, travel,
filament and feature totals, the same per layer, and the filament of
each object instance and its supports. The times are the preview's
estimates, summed the way the preview sums them.
The GUI writes <name>.statistics.json next to an exported G-code when
"Export statistics with G-code" is enabled in Preferences (Control >
Slicing). The CLI writes plate_N.statistics.json to the output directory
with --export-statistics, and exits with -25 when it cannot.
The per-instance masses now carry the object's name and the instance's
index, so the export can name them. The vertices the processor adds
inside a move for the actual speed view repeat the move's extrusion, so
the filament sums skip them.
* Move the statistics export preference to the Developer tab
"Export statistics with G-code" now sits in Developer > Settings
instead of Control > Slicing.
* Store the G-code statistics inside exported 3MF files
With the preference on, an exported sliced 3MF file holds each plate's
statistics as Metadata/plate_N.statistics.json, next to its G-code. The
print job sent to a printer and the autosave backup leave them out.
The CLI does the same with --export-3mf, and writes the separate file
only next to the G-code in --outputdir, as without it the G-code is a
temporary file. The file is written in binary mode, so it reads the
same on every platform and matches the copy in the 3MF.
* Address review of the statistics export
- Number each object instance by its position among its model object's
copies. A copy of another rotation or scale, and every copy on a belt
printer, slices as a print object of its own, where it was always
instance 0; a copy after a non-printable one shifted too.
- Take the filament totals from the processor's statistics, which count
the flush as the preview's filament table does, and list each
filament's model, support, prime tower and flush use. The summary's
cost is now filament_cost, as the preview's cost adds the time cost.
- State when the version goes up.
- Drop the preference's default, which get_bool already gives.
Object and copy ids were only assigned in set_object_info(), which runs only for Klipper, Marlin and RRF with Exclude objects on. Every other printer, Bambu included, printed an uninitialized object id and labeled every copy as copy 0. The numbering now runs on every export, so the labels are stable and unique on all printers. Output for Klipper, Marlin and RRF with Exclude objects on is unchanged.
The label test covers Marlin, Klipper and Repetier, plus a BBL row that checks each copy gets its own M624/M625 skip block.
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
When a CLI export rendered the no-light thumbnail but no plate thumbnail, the no-light loop set the plate thumbnail's status flag. That skipped the fallback that copies the plate thumbnail from disk, so re-exporting a 3mf that has plate_N.png but no plate_no_light_N.png left plate_N.png out of the package. The loop now sets no_light_thumbnail_status instead.
# Description
Follow-up to #13373, which added Align/Distribute to the object context
menu. This fixes the problems found in its review and brings
BambuStudio's align and distribute controls to the Move gizmo.
Align and distribute now work when all copies of an object are selected,
on parts of rotated, scaled or mirrored objects, and on objects with
several copies, whose other copies used to lose the part move. Aligning
a single object arranges its solid parts and leaves modifiers and
negative volumes in place. Moves in Z are offered only when Auto-Drop
would not undo them. Commands that would do nothing are disabled, each
operation is a single undo step, and other plates keep their slicing
result. The same fix stops the automatic drop to the bed after a move,
rotate, scale or mirror from also moving copies of other selected
objects.
The Move gizmo gets BambuStudio's align row. It aligns or distributes
the selected objects or parts, or aligns them to their object or to the
plate, and previews the target while you hover an icon. While the Move
gizmo is open, Ctrl+click adds objects to the selection, and Alt+click
adds or removes parts of the selected object.
No change to slicing output or project files.
# Screenshots/Recordings/Graphs
https://github.com/user-attachments/assets/44c308c3-faff-4f51-a3b7-88f264f03743
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Print::apply() decided the shrinkage compensation from the existing PrintObjects and their regions. On the first apply() they do not exist yet, so the first slice was not scaled. The CLI calls apply() only once, so filament_shrink and filament_shrinkage_compensation_z had no effect there.
The compensation is now decided before the PrintObjects are built, from the filaments the model uses, and stored in Print. When the filaments have different shrinkage values, every filament assigned to a feature, height range, modifier, support, tool change or the wipe tower counts, even if it does not end up printing. If they disagree, compensation stays off with the existing warning.
Supports, seam, fuzzy skin and texture displacement loaded paint data with the default max_ebt, so a 3mf with a paint state above what the gizmo paints read past the 3-entry color table. Each gizmo now passes its highest paint state as max_ebt, as the MMU gizmo already does, and higher states load as unpainted.
has_restore_data() signals "this backup is in use, leave it alone" by setting
origin to "<lock>". Its caller treats anything else as permission to delete
(Plater.cpp, the EVT_RESTORE_PROJECT handler):
try {
if (originfile != "<lock>") // see bbs_3mf.cpp for lock detail
boost::filesystem::remove_all(last);
}
catch (...) {}
So every "return false" that leaves origin empty is a delete instruction, and
three of them fired without anyone having established that the backup was
spent.
lock.txt exists but cannot be read. load_string_file() was outside the try, so
the exception left has_restore_data() altogether. The handler is queued and has
no try around the call, so this took the application down. Fixes#15670.
lock.txt reads but holds no pid. A crash or a full disk mid-write leaves a
zero-length lock.txt, because save_string_file() truncates before it writes.
boost::lexical_cast<int>("") throws into the catch, which returned false with
origin still empty, and the caller deleted the backup. Also #15670.
The .3mf probe could not answer. A path that cannot be probed at all reports
false exactly like a missing one, so an unreachable backup looked spent. The
error_code now separates the two: ENOENT and ENOTDIR mean "not there" and stay
collectable, any other error means the question was never answered.
All three now claim the lock. Declining a restore is recoverable; deleting a
backup nobody could examine is not.
What deliberately did not change: a readable directory with no lock and no
project still comes back with origin empty and stays collectable, as do a lock
held by another process and a directory that does not exist. Claiming the lock
whenever anything looked amiss would strand every spent backup on disk, so
those three cases are tests rather than prose.
Tested in tests/libslic3r/test_3mf.cpp. Against the previous behaviour the new
cases fail as they should -- the two uninterpretable locks come back with an
empty origin, and the unreadable one escapes as
basic_ios::clear: iostream error
while the three collectable cases pass unchanged, before and after.
Regenerate the template after rebasing onto main, which brought the 14
strings of the object Align/Distribute menu (#13373), and translate
them in all 22 catalogs. Each language reuses its catalog's existing
"align" wording, with one consistent pattern for the edge, center and
distribute items. Entries whose wording needs a native speaker's check
carry review notes.
Labels that were wrong or contradicted the glossary:
- Bridge flow ratio and internal bridge flow ratio use the glossary's
Flow ratio term (ca, de, fr, ja, nl, ru, sv, zh_CN, zh_TW), with the
tooltips that quote them.
- tr Line width was "layer width"; tr Modifier was the verb "change";
hu Rectangle was "square"; pl No sparse layers meant "layers without
purging"; uk Auto backup was slang.
- Wall order labels made consistent (de, fr, it); th Migrate User
Presets uses the glossary's preset term; lt keeps Pressure Advance in
English per the glossary and translates Adaptive.
- Role based wipe speed now says it follows the extrusion role (de, ru,
pl, tr, zh_CN, zh_TW).
- ru Line infill pattern is "Линия", matching the glossary.
Translations that only covered part of the source or had typos: de
external bridge flow tooltip, zh_TW role-based wipe tooltip, ru wall
type tooltip option names, fr apostrophes, uk final period. es entries
from the last pass now use usted, the catalog's register.
Terms aligned with the glossary: purge/flush, prime tower, wipe and
flushing volume (ko), flow ratio (ca, ja), layer, shell, travel, walls,
plates, raft and fuzzy skin (th), wipe (eu), bridge (de).
Machine-translated entries that use the unsettled Prime tower / Wipe
tower and Tool (printer) terms carry review notes (cs, fr, ja, nl, tr,
uk).
The design said every pattern other than the trapezoidal and adaptive
ones outlines its centerlines with multiline_fill(). 3D Honeycomb skips
it and draws its own parallel copies of each wave, which never cross.
small_support_perimeter_speed declared outer_wall_speed as its ratio_over
parent, while GCode.cpp and the tooltip use the support or support
interface speed. The option is a per-nozzle vector, so ratio_over is
only metadata here and no output changes.
The prompt that clumping detection needs a prime tower was nested inside
the smooth timelapse check, so disabling the tower with clumping
detection on and traditional timelapse gave no prompt; the problem only
surfaced as a warning after slicing. The prompt now appears whenever
clumping detection is on. The smooth timelapse prompt is unchanged, and
answering it with No still skips the clumping one.
extruder_printable_area linked to the Extruder offset Position section,
which describes a different setting; the wiki now has an Extruder
printable area section. The two compatibility conditions linked to their
parent sections instead of the Printer Condition and Process Condition
sub-sections that describe them.
The wiki now has sections for options that were linked to a page top or
not linked at all: the polyhole sub-options, accel to decel, extrusion
rate smoothing on external features, max bridge length, the tree support
brim, seven prime tower options, Flush into this object and clumping
detection. Point each link at its section, and follow the corrected
spelling of the wall transitioning threshold angle heading.
- Minimize wall height angle acts on slopes steeper than the angle, not
shallower (ContourZ.cpp).
- The wipe tower's default wall type is Rib, not Rectangle; extra flow
for purging can't make lines narrower (minimum 100%); the max purge
speed caps the sparse infill speed on sparse layers and also the outer
walls.
- Interlocking depth is no longer ignored without a maximum width: every
other layer is cut to that depth unless beam interlocking is on.
- Draft shield: skirts are placed around brims now, so the note about
intersecting brims is removed.
- Infill/wall overlap percentages are relative to the inner wall line
width, not the sparse infill width.
- Bridge line width 0 uses the nozzle diameter for thick bridges.
- Fully overhanging walls use the 75% overhang speed limited by the
bridge speed; there is no 100% speed. The bridge speed tooltip no
longer mentions the removed classic overhang mode.
- First layer speed covers walls and brim; everything else uses first
layer infill speed.
- Inner/Outer/Inner is recommended over Outer/Inner, not against it.
- Minimum wall length is a multiple of the wall width, not mm, and One
wall threshold is in the Walls and surfaces group.
- Small perimeter speeds can't be set to zero (min 1), so "Set to zero
for auto" is removed.
- Label typo: Role based wipe speed.
The translation template is regenerated; it also picks up strings that
other recent changes added without regenerating it.
Linux webviews lost the BBL-Slicer, dark|light and BBL-Language tags
with the stock WebKitGTK User-Agent, so the bundled pages stayed light
and dropped every message to the host. The tags come back on Linux, now
appended to the string WebKitGTK reports for itself instead of the Mac
one. Windows and macOS are untouched.
A fifth Python plugin capability type, ResourceTexture, contributes
height-map textures to the Texture Displacement picker. A pack returns a
directory of images; the host validates it once at load and lists the
images under a section of their own.
- New capability type ResourceTexture, type string "resource-texture",
exposed to plugin authors as
orca.resource_texture.ResourceTexturePluginCapabilityBase with one
required method, get_texture_dir().
- The directory is resolved once on the load worker, under the GIL, and
cached as a plain string: the picker runs inside the ImGui frame, where
a trampoline call would block on the GIL behind a load/unload worker, or
throw past EndPopup.
- Validation resolves the directory and requires containment in the plugin
root, re-checked per file so a symlink inside the pack cannot resolve
outside it. Containment is the only security control here and it guards
exfiltration: layer bytes are embedded in every saved .3mf, so a pack
that could return ~/.ssh/id_rsa would leak it through any shared project.
- New host registry PluginResourceTextures, filled from both the plugin
and the capability lifecycle callbacks plus a startup sweep. It holds
strings only and never a capability reference, so nothing outlives
Py_Finalize.
- The picker lists pack textures read-only, one collapsible section per
pack, in a fixed-size scrolling popup whose height follows the window. A
generation counter invalidates the cached listing, and each gizmo's
cached thumbnails, when the set of packs changes.
- Tests in tests/slic3rutils/test_plugin_resource_texture.cpp cover the
validator (parent-directory and absolute escapes, an escaping symlink,
the extension whitelist, the 64-file cap), the listing and its
load/disable/re-enable/unload lifecycle, hostile get_texture_dir()
overrides, decode parity with hand-picked files, and a .3mf round trip
with the pack uninstalled.
No persistence work is needed: a layer carries its own image bytes and 3MF
writes them into the archive, so uninstalling a pack leaves existing
projects intact.
Behaviour change worth noting for review: the picker popup is now a fixed
size that scrolls, rather than sizing itself to its content, and it gains
one collapsible section per pack. That sizing applies to every user, not
only those with packs installed.
* Add Align/Distribute objects on the print plate
Ports the Align/Distribute feature from BambuStudio 2.5 (AGPLv3,
same license as OrcaSlicer). Adds a new GLGizmoAlignment helper
that operates on the GLCanvas3D selection, twelve thin Plater
wrappers, and an Align/Distribute submenu attached to the same
five context menus BambuStudio attaches it to: object menu, text
part, BBL part, and the single/multi-volume context menus.
GLCanvas3D::do_move gains a defaulted force_volume_move parameter
so the helper can move multiple selected volumes within a single
object even when selection mode is Instance. GLGizmosManager gains
check_object_located_outside_plate, which switches the active plate
to the one containing the selection before align/distribute runs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* fix build errors
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Everything after `--` is appended, unchanged and last, to the cmake
configure line of the deps and OrcaSlicer builds, so a custom compiler,
compile flags or linker can be passed without adding an option to the
script for each one. The arguments are kept as an array, so a value with
spaces such as -DCMAKE_CXX_FLAGS="-O2 -march=native" survives, and they are
forwarded into the -g container. A stray argument without `--` aborts with
a message instead of reaching cmake as a path.
The existing DEPS_EXTRA_BUILD_ARGS and ORCA_EXTRA_BUILD_ARGS hooks are
unchanged and now mentioned in the usage text.
* add new repositories and new push
* fix bug about setting_id
* fix bug and complete the test
* Fix Goofoo profiles so the bundle loads and passes profile checks
Repair the OrcaFilamentLibrary index entry that broke JSON parsing, regenerate the hand-written filament ids as Orca ids, and bump the library version. Point each printer's default filament and print profiles at presets that exist, normalize the two presets that needed it, rebuild both indexes, correct the ABS-GF/PA-GF/PA6-CF filament types, align the GP1/Velo+ preset names with the library's @ spacing, and fix the casing and resolution of the model cover images.
* Drop the copied Custom Klipper printer from the Goofoo bundle
Goofoo shipped Custom's MyKlipper machine model verbatim, so the Add Printer list offered "Generic Klipper Printer" twice and the printer type resolved its model_id by vendor order. The Goofoo machines now inherit the shared fdm_klipper_common directly, carrying over the per-variant layer heights the removed base supplied, matching how the other Klipper vendors wire theirs. Each machine's resolved settings are unchanged.
* Align the Goofoo preset names with the profile conventions
The machine models and variants already followed the <Model> / <Model> <nozzle> nozzle shape, but the processes put the model in the quality slot and used the bare vendor as the @ target (0.20mm GP1 @Goofoo). They are now named like every other vendor's: <lh>mm <Quality> @<Vendor> <Model> (0.20mm Standard @Goofoo GP1), with the machines' default_print_profile and the index updated to match. The four leaves also inherit the shared fdm_process_klipper_common directly instead of the copied Custom "0.20mm Standard @MyKlipper" base, which is removed; GP1 and TurboM gained the layer_height 0.2 that base supplied, so every preset's resolved settings are unchanged. Removed the orphaned Generic Klipper Printer cover and the unused Turbo-M texture, and renamed TurboM's bed texture to the -texture.svg form the other two models use.
* new Goofoo machine and clean profile
* Restore library filaments dropped by merges from main
* Give the Goofoo Velo+ Tool and 0.2 nozzle fitting processes
---------
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
* fix: login crash when network plugin is unavailable
* fix: directly use !config_version.empty() for the control block
* Use if/else for the empty plug-in version check
* Fix multi-device task list using a deleted network agent after plug-in reload
---------
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
The filament list groups presets by vendor: hovering a vendor in the main
popup opens a side popup with that vendor's materials. On GTK the side popup
holds the popup grab, and gtk_main_do_event hands every GDK_SCROLL of the popup
chain to the grab widget. GDK names the event for the window under the cursor,
so the side popup received the wheel events raised over the main list and
consumed them: the main list could not be scrolled while a vendor was hovered.
The problem shows on both X11 and Wayland, because the rerouting happens in
GTK, not in the compositor.
The side popup now inspects the event's GDK window and, when the event belongs
to the main popup, replays it there. The replay goes through
gtk_propagate_event(GetConnectWidget()): wxGTK connects its mouse handlers to
the inner client widget of a popup (a wxPizza child of the outer GtkWindow), so
a signal emitted on the outer GtkWindow reaches no handler at all.
The widget reference in .claude/skills/orca-wxwidgets records the mechanism and
the connect-widget fact.
The build height check in Print::validate() compared object layers that already include the Z shrinkage compensation against printable_height divided by that compensation. That counted the compensation twice, so objects that fit were rejected. The check also left out the raft, so an object whose top plus the raft is above printable_height passed validation.
The check now adds object_print_z_min to the top layer and applies the compensation once. The shrinkage message is shown only when the object fits without the compensation. Belt printers apply the compensation to the raw model height, since their height does not come from the layers.
The constructor re-declares `namespace fs = boost::filesystem;`, shadowing the
identical file-scope alias the rest of the file already uses. Removing it
leaves every `fs::` in the file resolving to that one declaration.
No functional change.
The startup probes for download_path and last_backup_path used the throwing
boost::filesystem overloads, so a path that could not be probed at all, such as
one whose parent is not searchable, raised filesystem_error instead of falling
back. init_download_path() runs from the GUI_App constructor inside GUI_Run()'s
try, which left the application showing a message box and exiting with no window
and no log file, the throw preceding the log sink. has_restore_data() runs from
the queued EVT_RESTORE_PROJECT handler, whose try starts after the call.
Five probes in those three functions take the error_code overload instead.
Nothing branches on ec, and readable or missing paths are unaffected because both
overloads return the same value for them. A new test covers has_restore_data()
against a backup directory whose parent has no search permission, and skips on
Windows, as root, and where the mode change is refused or not enforced.
- Slicing pipeline plugin linked to others_settings_plugin_picker, a page
that doesn't exist. It now opens the Slicing Pipeline section of the
plugin types page.
- Infill shift step linked to the Cross Hatch pattern, which doesn't use
it. It only applies to Cross Zag and Locked Zag, so it now opens Cross
Zag.
- Fuzzy skin and Fuzzy skin generator mode were crossed: the generator
mode opened the section that documents Fuzzy skin's values. Each now
opens its own section.
- The process Change extrusion role G-code had no wiki link; it now has
one, like the printer and filament versions.
- The hidden Frequent page linked Wall loops to the top of the walls page
instead of its section, which made the wiki sync write its metadata
twice.
Orca sent `X-Api-Key: 88888888` to Moonraker whenever the printer preset's API Key was empty, which Moonraker rejects with HTTP 401 since Arksine/moonraker@93f5d9c. The preset's API Key is now the only thing that sets the header, and an empty key sends none.
Logging out of a printer now marks it revoked rather than only clearing its code, since an empty code is a valid Moonraker binding. Discovery events are delivered through the network agent and dropped if their printer agent was replaced while queued.
* Precise Seam: new full-coverage policy for Seam Blocked
A Seam Blocked modifier that covers a whole perimeter now blocks it, like a
perimeter painted red all round, instead of being skipped. It overrides
painting and lower-priority modifiers, like Enforced and Neutral. A new
warning names the modifier: "Seam Blocked modifier "<name>" of "<object>"
covers an entire perimeter". It is shown only when a perimeter stays
entirely Blocked after all weak modifiers. The full-containment warning now
covers only Center, Left and Right.
* Precise Seam: fix seams on perimeters blocked all round
- SeamComparator::is_first_not_much_worse() compares the types first, so two
Blocked candidates fall through like Neutral ones. On a perimeter blocked
all round, by a whole Seam Blocked modifier or by red painting, Random put
the seam on the closing vertex of every layer and Aligned stopped aligning.
- The whole-perimeter Blocked warning flags only the highest whole zone, the
one in effect. A lower whole zone it overwrites is no longer reported, and
the candidate scan runs only when that zone is Blocked.
- The warning msgids take the type name from the menu's "Seam Blocked" msgid.
- PreciseSeamWarnings::set() shares the load-before-store idiom, and every
weak segment records its modifier.
- Tests: a whole Blocked perimeter gets the seams of an unmarked one under
every seam position; overwritten whole Blocked zones are not reported.
---------
Co-authored-by: Ian Bassi <ian.bassi@outlook.com>
The ubuntu:24.04 base image has a built-in ubuntu user with UID 1000, so the Dockerfile skips creating the host user and `docker run -u "$USER"` cannot find it. Run the container with the numeric UID and GID instead, and set HOME and the working directory to the bind-mounted home.
The time estimate added 260 s for every G29 on non-Bambu printers. Prusa MK4, MK4S and CORE One start G-code splits bed leveling across six G29 commands, so those printers showed about 26 extra minutes. The 260 s is now added once per G-code processing run. Bambu printers still only count a G29 inside an M622 J1 block, and their estimates are unchanged.
Fixes#15088
* Keep User Filament Presets on Printers Their Parent Listed in 2.4.2
User presets store only their diff, so they inherit the parent's
compatible_printers. Profile changes since 2.4.2 renamed and narrowed many
system filament presets, so user presets made against them disappear from
printers they were made for and the saved selection silently falls back to
another filament at startup.
orca_profile_tool.py legacy-compat writes, per current system filament
preset, the printers it covered in an older release and no longer lists,
kept only where a preset with the same filament_id still covers the printer
and they share an extruder variant. The loader gives these printers to user
presets that inherit their parent's list, without writing any file; saving
a preset detached, cloned or with its own list drops them.
* Ask Before Migrating User Filament Presets to Their Legacy Printers
Instead of giving user presets their parent's 2.4.2 printers in memory,
detect the user presets that inherit their parent's compatible_printers
and are missing printers from the legacy map, and offer to migrate them
at startup or from Help > Migrate Filament Presets.
Migrating first backs up the user folder to
filament_preset_migration_<datetime>_backup.zip in the configuration
folder, and migrates nothing if that fails. It then writes the printers
into each preset's compatible_printers, patching only that key so the
preset keeps the parent name it was saved with and older releases can
still resolve it. Cloud-synced presets are marked for upload.
No asks again at the next launch; "Don't remind me again" stops the
startup prompt.
* Key the legacy filament preset map by the parent's 2.4.2 name
* Check the legacy filament preset map in CI
* Limit the 2.4.2 filament preset migration to presets that need it
Only presets that still name their parent as 2.4.2 did and have no printer
list of their own are offered, and a preset changed after it was found is
left alone. The startup prompt appears once, after upgrading from 2.4.2 or
earlier, as a standard dialog that keeps "Don't remind me again"; Help >
Migrate Filament Presets is unchanged. Migrating no longer overwrites an
unsaved printer-list edit or marks the project modified.
* Ask on Every Start and Take the User Presets Lock Once When Migrating Filament Presets
The startup prompt no longer depends on the version the previous run
recorded: whoever still has user filament presets that lost printers is
asked, including users already on a 2.5.0 nightly or alpha. Answering No
asks again at the next start; "Don't remind me again" stops it.
Migration takes the user presets lock once, around the backup and every
file write, instead of once per preset. A lock another instance holds is
reported once ("another OrcaSlicer instance is saving presets") with
nothing written, rather than as one failed preset after another once the
lock's cool-down kicks in. Detection records each preset's file and the
list it will get, so the writes need no collection lock while the user
presets lock is held, the order InstanceLock requires; the loaded presets
are updated afterwards.
* Run the Legacy Compatible Printers Map Check in check_profile.sh and .ps1
CI's "Check profiles" job checks resources/info/legacy_compatible_printers.json
against the v2.4.2 profiles, and its failure message points at these scripts
as the local reproduction, so they run the same check: legacy_compat, right
after profile_tool. The v2.4.2 tree comes from git archive of the tag, is
cached in the work dir like the other downloads, re-extracted with --refresh,
and fetched from upstream first when the checkout lacks the tag.
* Back Up Before the User Presets Lock and Merge Migrated Printers Into the Live List
The lock guard is meant for a few file operations, and zipping a large user
folder is not one, so the backup is taken before it and the guard covers the
writes only; each write re-reads its file, so a save that slipped in before
is still respected.
The in-memory update adds the lost printers to the list the preset holds at
that moment rather than replacing it with the one found at detection, and
leaves a pending cloud sync state alone: a download waiting to be saved kept
its newer list and its "save" flag, which "update" would have cancelled.
The lock message no longer blames another instance alone, since the lock is
also refused for a while after a miss and when the lock file cannot be
opened.
---------
Co-authored-by: SoftFever <softfeverever@gmail.com>
Auto-arrange centered the pile on `best_object_pos`, so on printers with an off-center position (A1 mini, H2C, H2D, H2D Pro, X2D) a large pile could end up partly off the bed. The final alignment now always clamps the pile to the bed, and the `clamp_to_bin` flag is removed. Belt printers already clamped, and centered alignment is unchanged.
The check against fixed parts and excluded regions now tests the position the pile is actually moved to.
Fixes#15727
CurlGlobalInit() probed the CA bundle with the throwing fs::exists overload, which
returns false only for a missing path and throws on EACCES. An existing but
unreadable cert path then killed the GUI at startup instead of falling through to
the CA_BUNDLES list. Use the error_code overload at both probe sites.
Fixes#15653.
Print::validate sizes the pre-slice prime tower from Print::extruders(),
which lists every mixed filament slot under its own id, so the estimator
planned one purge block per slot. The tool ordering resolves a mixed slot
to one physical component on every layer, with or without colour
sublayers, and the generated tower only purges those. The GUI preview and
its placement clamp already expand the slots, so the tower was drawn
inside the bed and then refused as partially outside it: with 8 mixed
slots at 0.08 mm layers the estimate was 114 mm for a tower that prints
at 64 mm.
The estimator now expands mixed slots to their components itself, so
validation, the GUI and the CLI placement size the same tower.
* Add orca-translations Claude skill and PO tool
Add a .claude skill that guides translation passes over the localization/i18n gettext catalogs. It covers the workflow: run gettext, settle new glossary terms, translate in batches, verify, and report.
Includes translation-rules.md, which condenses the placeholder, line-break and glossary rules plus the guidance on review notes.
Also adds scripts/po_tool.py (standard library only). It lists pending entries with glossary hints, validates and applies translations, adding '# AI Translated' and '# Needs human review' comments, and rewrites only the entries it changes. It can also check catalogs, look up existing translations, suggest glossary candidates and update the wiki glossary TSV.
* Let po_tool apply translations by entry id
pending now gives each entry a short id, so a translation file can name
entries by id instead of repeating the English msgid. apply accepts either.
A "review": "" item clears an earlier review note when overwriting an
entry, and check gives each msgfmt run its own temporary output file so
parallel checks do not collide.
* Fill all empty translations
Regenerate the template with run_gettext --full and translate every empty
entry in the 22 catalogs (7,089 entries) with the orca-translations skill,
following the localization glossary, including the new belt printer,
IDEX/IQEX and plugin terms. Every new entry is marked "# AI Translated";
122 also carry "# Needs human review" with what to check.
* Show entry ids in po_tool lookup and search translations
lookup now prints each entry's id, so an entry found there can be
rewritten with apply --overwrite without repeating the English, and
--msgstr searches the translations instead of the msgids, which finds
every entry that still uses a term being changed.
SKILL.md describes how to change an established term: check the
catalog's usage first, update the glossary cell, then rewrite only the
entries where the term has that meaning.
* Fix mistranslated terms and align catalogs with the glossary
Correct labels that the glossary had copied, and the entries that used
them:
- ca: Raft was rendered as the Brim term; it is now "Rai". Exhaust fan,
Monotonic line, Unretract, Slice, Lateral Honeycomb and PA Tower
labels fixed.
- cs: Lightning infill is "Blesk", not "Osvětlení" (illumination).
- sv: Retraction is "Retraktion", not "Reduktion"; Unretract,
Lagerhöjd, Delkylningsfläkt, Utsugsfläkt, Aligned Rectilinear and the
Preset label fixed.
- de: Brim is "Rand" and Skirt is "Umrandung", as most entries already
had them.
- lt: brim entries used the Skirt word "apvadas"; they now use "kraštas".
- es: Sparse infill is "relleno de baja densidad", the catalog majority.
- eu: nozzle wipe is "garbiketa", not purge.
- th: the project plate is "เพลต", distinct from the bed.
- uk: Quarter Cubic, Latin "PA" in PA Tower, and Belt purge tower.
- pl: Cross Hatch, Tri-hexagon, Volumetric speed and Raft labels.
- zh_TW: Brim stays "Brim", as the existing labels have it.
- tr: "Prime Tower" read "Prime Minister Tower".
- ko: the plugin "Capabilities" label read "performance".
- Build Volume labels fixed in tr, de, cs and eu.
All 242 rewritten entries are marked "# AI Translated".
* Make more UI text translatable and fix pluralization and context
Audit the sources for text that translators could not translate
correctly and fix it.
Mark text that reached users in English only: IDEX/IQEX role names,
plugin permission prompts and errors, 3MF/preset loading errors, G-code
placeholder parser errors, calibration and device messages, notification
texts, file dialog filters, and the config validation messages.
Replace sentences built from translated fragments, such as the prime
tower warnings, "left"/"right" nozzle names, and "Filament " + number,
with single msgids that have placeholders, so translators can reorder
and inflect them.
Use plural forms for texts with a count ("%1% mixed filaments",
"%d objects", "%1% triangles") instead of "(s)" and mismatched verbs.
Fix the swapped singular/plural pair in "The following presets inherit
this preset".
Add translation contexts where one English word has several meanings
(Subscribed/Local, Standard, Pause, View, Step, Model, Camera, Flow,
Filter, Others, Prepare, the Design tools). Show "Size" instead of
"Build Volume" in the Cut tool, which displays the object size.
Fix code that compared translated text with English (the Edit G-code
dialog group, the plugin registration failure check), and add the
bundled web pages' missing string-table entries. Use the single
character for degrees Celsius, say "G-code", and describe the head
filament map without internal jargon.
Install the libslic3r translate callback before the presets load, so
their load errors are translated, and list the 28 source files that had
marked strings but were never extracted into localization/i18n/list.txt.
* Suggest earlier translations and list fuzzy entries in po_tool pending
pending now attaches the closest earlier translation of a similar
English text (from obsolete or translated entries), so a reworded or
split string keeps its wording, and it lists fuzzy entries by default,
since users never see them. A fuzzy entry also carries its stale
translation as previous_translation.
* Fix wording of the non-boolean operands error
The logical operation error said "operators" where it means the
operands the operation was applied to.
* Translate the newly translatable strings in all languages
Regenerate the template after the translatability changes and fill every
new or changed entry in the 22 catalogs (about 12,800 entries): the
strings that became translatable, the plural forms, the entries split by
context, and the previously fuzzy singular entries that became plurals.
Every new entry is marked "# AI Translated", and about 360 also carry a
"# Needs human review" note saying what to check. Terms follow the
localization glossary. Earlier translations of reworded strings were
reused where the meaning was unchanged. Existing entries were not
changed, and obsolete entries are kept.
* Document the source translatability audit in the orca-translations skill
Add a reference describing how strings reach translators, what to look
for (untranslated text, spliced sentences, missing plurals, ambiguous
words), what not to translate, where a context cannot be carried, and
what to do after the edits, plus a scanner that lists untranslated
user-facing string literals.
* Document how to write translatable strings in AGENTS.md
Tell developers to keep each message complete in one msgid: use
placeholders instead of concatenating translated fragments, plural forms
for counts, and contexts for ambiguous words, and to list new source
files in localization/i18n/list.txt. Add a matching code review item.
* Translate the strings added since the rebase
Regenerate the template against the current main and translate the 26
new entries per language: the Fix Model commands and messages, and the
texture displacement image adjustment panel. Each new entry is marked
"# AI Translated"; existing entries are unchanged.
* tidy
A 3mf saved without a plate thumbnail, such as a CLI export with no OpenGL context or a backup, still listed /Metadata/plate_1.png and plate_1_small.png in _rels/.rels. lib3mf rejects a package whose thumbnail relationship points at a missing part. The writer now records which plate thumbnails it wrote and only declares those.
A tooltip appeared only on the next pixel of mouse movement after its wait
had finished. ImGui hides a window on the first frame it is submitted, to
measure it before showing it, and does so every time a tooltip reappears
because those windows are recycled (the Popup|Tooltip branch of Begin()).
So the frame a tooltip is decided on is never the frame it is seen on, and
one carrying no animation asked for nothing further - leaving it in its
hidden measuring frame until some other event happened to produce the next
one. Before the hover delay this never showed, because the tooltip was
submitted on every mouse-move frame and the movement supplied the second.
It now asks for frames until it has been drawn twice in a row.
that finishes on its own
The hover delay only finished while the mouse was moving: a pointer resting
on a control produces no events, so nothing redrew the canvas and the wait
never reached its end - the tooltip appeared on the next pixel of movement
instead. It asked for its frames with schedule_extra_frame(), which only
arranges a timer whose event an idle event is assumed to follow, and
on_idle renders nothing unless the canvas was already dirty when it ran.
The wait and the animations now both use request_extra_frame(), which ends
on_idle with evt.RequestMore() and so obliges wx to deliver the next idle;
it is what 3DBed's fade uses. Asking for frames as fast as they come is
safe now that which frame is drawn comes from the clock rather than from
the frame count - a redraw that arrives early draws the same frame again.
Nothing explains a control while the control is being used, so no tooltip
appears while a slider is being dragged, a value typed into one, a combo
open or a layer card dragged. The pointer sits on the very thing being
edited, so a tooltip there covered the model exactly while the user was
watching it change. The full wait comes back afterwards rather than the
short one: having just let go of a slider is not a reason to be shown a
picture of it.
AnimatedTexture::texture_id() loses its time-to-next-frame out-parameter,
which nothing asks for any more.
The animations crawled when left alone and raced while the mouse moved.
They were timed by ImGui::GetTime(), and nothing in OrcaSlicer ever sets
io.DeltaTime, so it keeps imgui's default of 1/60 per frame: imgui's clock
counts frames, not seconds. On a canvas that renders only when something
asks it to, that is not a clock. Playback now reads a steady_clock, so a
frame is due when it is due whatever the canvas is doing.
An animation also restarts when its tooltip opens again, not only when a
different one opens: a gap in the frames it was drawn on is what that
looks like from here.
Tooltips now wait for the pointer to rest on a control for 0.45 s. The
panel is dense and a tooltip lands on the model beside it, so with no wait
a sweep from the top of the panel down to Bake flashed a dozen large panels
over the scene. After one has just been up the wait drops to 0.1 s for 0.7
s, so reading several settings in a row stays quick - but not to zero, or
that same sweep would flash again once one tooltip had been seen. A pointer
resting on a control produces no events, so the wait asks for the frame it
will appear on itself.
rest of the panel
The animations advanced one frame per mouse movement and stood still
whenever the mouse did. schedule_extra_frame() only guarantees an idle
event at the moment the next frame is due; GLCanvas3D::on_idle() returns
without rendering anything unless something marked the canvas dirty by
then, and nothing had. set_dirty() on the gizmo is what on_idle reads,
through GLGizmoBase::update_items_state(), to find out there are frames
still to draw.
Thirty-odd more pictures, covering the paint tools, brush size, the
connected-area fill, the five views, the five projections, subdivision,
feature-adaptive refinement, remeshing, hard edges, slivers, geometry
smoothing, filament mixing and denoise.
The generator grew three more drawing languages beside the shaded relief,
each picked by what the setting is rather than by taste: a flat mesh of
triangles where the subject is the mesh or a tool, a section where the
subject is where the surface sits relative to where it was, and a flat ramp
of greys where the subject is which filament a spot ends up as. Lighting a
colour illustration or looking down on a vertical change would both add a
second thing moving between two pictures meant to differ in one.
Mapping is one ball with the checker laid on it five ways, which is the
whole of what choosing a projection does - except Unwrap, which is not a
projection onto the model and so is drawn as the flat layout it produces.
Ten more tooltips get a picture: Depth, Tile size, Repeat/Mirror, Invert,
Smoothing, Edge fade and Blend as stills, Depth, Rotation and Midlevel as
animations. All drawn by scripts/gen_texture_displacement_tips.py from the
same arithmetic the feature uses, one motif throughout, so that whatever
differs between two of them is the setting and nothing else.
Midlevel is drawn in section instead of as a relief. Shading reads slopes,
so a height field moved up or down as a whole shades exactly as it did
before - and where the relief sits relative to the surface is the entire
meaning of that setting. From the side, against the bare surface, it is the
one thing visible.
A single picture is now capped at 10 em tall and centred in its column,
rather than drawn at the tooltip's full width, where a square one came out
as tall as the text beside it was long.
The panel is pinned to the right edge of the canvas, so a tooltip left to
follow the mouse covers the very control it explains. panel_tooltip()
places every tooltip in the gizmo to the left of the panel instead - or of
the texture picker, while a tooltip is coming from inside that popup - the
same placement the picker itself already uses.
A tooltip can now carry pictures under its text: stills, or GIFs that loop
while it is open. Several settings in this panel describe a shape rather
than a number, and a sentence about a shape is read twice and understood
once. Files live in resources/images/texture_displacement and are named by
a tooltip; one that is missing is skipped, so a tooltip without its art yet
is exactly the text tooltip it was before.
AnimatedTexture decodes a file once and uploads a GL texture per frame -
wxImage for stills, wxGIFDecoder for GIFs, whose frames are composited here
with the disposal method each one asks for. An animation restarts when its
tooltip opens and asks the canvas for a redraw timed to its next frame.
Tile and Rotation are wired up as the first two; scripts/
gen_texture_displacement_tips.py draws their relief from the same
arithmetic the feature uses. docs/HLSD/tooltip-illustrations.md covers the
rest.
WebFreak001/deploy-nightly still targets Node.js 20, so every main build
logged a deprecation warning per deploy step. The action has had no
release since 2024 and its Node 24 PR is unmerged.
The local action in .github/actions/deploy-nightly uploads the new file
under a temporary name and renames it once the upload finishes, so a
failed upload keeps the existing nightly asset. It also fails on a missing
file before deleting anything, retries transient API errors, times out
stalled transfers, handles concurrent runs, and finds the release by its
nightly-builds tag. The macOS validator DMG deploy now runs only when that
DMG was built.
* Carry existing colour across a bake instead of reindexing it
Baking, painting colour, then baking again left colour debris on surfaces
nothing had painted and destroyed what the earlier bake had laid down.
TextureDisplacementBakeJob::finalize() deserialized the volume's pre-bake
mmu_segmentation_facets into a TriangleSelector built on the post-bake mesh,
on the assumption that the bake is topology-preserving. That holds for the
classic path, which moves vertices and leaves the topology alone, but not for
the one-run pipeline, which rebuilds the mesh. TriangleSelector::deserialize()
only checks that the stored triangle ids are in range, and a bake almost
always grows the triangle count, so every stale id passed that check and
landed on an unrelated face.
Measured on a 26 mm slab baked twice at 0.19 mm with a 750 k budget: 7580
triangles ended up carrying paint neither bake produced, as far as 31 mm from
any colour that was actually laid down, while the first pass's colour was lost
where the second bake left the surface alone.
TriangleSelector::remap_painting() already transfers paint between meshes by
position, and ModelVolume::restore_painting() uses it for the same reason, so
the bake job now uses it too. It runs in process(), on the worker thread: it
builds an AABB tree over both meshes and took around a second on the case
above, which would have frozen the UI in finalize(). The bake's own colour is
still applied last, so it wins where it painted and the rest is kept, which is
what makes successive bakes accumulate rather than overwrite. If the volume is
painted while the bake is running, a timestamp check falls back to the live
mask so that stroke is not dropped.
The out_triangle comment in TextureColorRequest claimed the bake is
topology-preserving, which is what justified the original indexing; it now
says which mesh the colours belong to.
The regression test bakes twice, checks both results are closed, and checks
the first bake's colour survives the second at the positions where it was
laid down and nowhere else.
* Carry the paint across with transfer_facet_paint
The branch used TriangleSelector::remap_painting(), which pairs a source
paint triangle with a target triangle overlapping it to within
TriangleCursor::tolerance. That is right after a remesh, which leaves the
surface where it was, and wrong after a bake, which has moved it by the
relief depth - most of the paint finds no partner. transfer_facet_paint()
asks the other question: every triangle of the new mesh takes the state
painted at its foot point on the old one, which is exactly where it came
from however deep the relief.
It also does all four paint channels rather than colour alone, and it runs
where the volume is read, so the snapshot and the timestamp that guarded
against the user painting while the bake ran are no longer needed: the
annotation is read at commit time and is by definition current.
The function and its call site are byte for byte what the model-UV branch
carries, so the two merge in any order and in either order produce one
mechanism rather than two fighting over mmu_segmentation_facets. The
regression test is unchanged except for the call it makes.
---------
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
# Description
An image laid onto a model came out upside down under the box
(tri-planar) projection, and on three of a cube's six faces it was also
mirrored left to right.
The sampler treats `v = 0` as the height map's top pixel row, but every
projection built `v` from world +Z, so `v` grew upward. The mirroring is
a separate defect: both faces sharing a world plane got the same `u`.
Note: re-baking an existing project now places its texture upright, so
it will look different from before. Geometry already baked into a 3mf is
unaffected.
# Images
<img width="2420" height="1655" alt="image"
src="https://github.com/user-attachments/assets/dcf7f62f-0f51-4966-97e4-d0386eb454c2"
/>
# Tests
6 tests added in `tests/libslic3r/test_texture_displacement.cpp`,
covering all six cube faces and every mapping mode, plus a full bake of
a known asymmetric image.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
GCodeProcessorResult declares a copy assignment, which suppresses the
implicit
move assignment, so `*result = std::move(processor.extract_result())`
binds to
the copy and duplicates the whole moves array: 96 bytes per move,
measured at
147.7 MiB in one allocation for a 597k facet model at 0.08 mm, with the
source
staying alive until the export returns.
Add the move assignment. It assigns exactly the same 39 members in the
same
order as the copy, including the ones the copy deliberately leaves
alone, so a
target that carries its own filament maps and nozzle type keeps them. A
move
constructor is intentionally not added: the type holds a std::mutex, so
it is
neither copy nor move constructible today and a partial one would leave
ten
members uninitialised.
reset() now releases the two print sized vectors rather than clearing
them. Its
callers are the paths that discard a result, so the memory went back
only at the
next slice.
ViewerImpl::reset() does the same for the two vectors sized by the
print, which
is what keeps a discarded preview resident, and counts m_vertices_colors
in the
reported CPU memory, where it was missing.
The texture-displacement gizmo pushes pane state to the UV editor from the
preview rebuild that runs on every frame of a slider drag. Tile size is the
only one of the layer card's sliders whose value appears in that pane state,
so dragging it changes the "N.N mm tile" label, the panel relays out, and the
UV canvas receives a synchronous wxEVT_SIZE. Its on_size() flushes a paint
with Update(), and render() then makes the app's single shared wxGLContext
current on the UV drawable in the middle of the 3D canvas's frame.
On X11/GLX that out-of-turn context activation is fatal: the drawable the
3D canvas later swaps no longer belongs to the current context, the server
returns BadMatch, and GDK turns the unhandled X error into g_error(). EGL on
Wayland reports only a status, so the same theft silently corrupts the frame
instead of aborting - which is why the guard is not conditioned on the
display backend.
Add GLCanvas3D::is_frame_in_progress(), backed by a counter that
_render_frame() maintains through its existing scope guard, and let the UV
canvas skip the synchronous paint while a frame is on the stack, re-queueing
it with Refresh() so nothing is dropped. Rebind the context with
_set_current() before SwapBuffers() as well, so a surface that steals it by
some other route - a modal GL dialog a gizmo opens from its panel - cannot
turn the swap into the same fatal error.
The cloud upload sent only the keys that differ from the parent, so a
variant layout equal to the parent's was left out. On download the
loader sized every per-variant value by the missing layout, cut it to
one value and kept the parent's values for the other nozzles and nozzle
volume types. On dual-nozzle printers the right nozzle's settings
silently reverted to the parent's after a sync.
Upload the variant layout always, as the local save already does, and
read copies stored without it against the parent's layout when they
carry one value per parent variant.
Opened or drag-dropped G-code files ignored `; MANUAL_TOOL_CHANGE` comments, so the preview showed every move on the first filament. Only `apply_config(const PrintConfig&)` read `manual_filament_change`, and loading a file goes through `apply_config(const DynamicPrintConfig&)`. Read the option there too, initialize the flag to false, and add a regression test.
Debug builds asserted on startup because these calls raise GL errors on a core-profile context. Both draws already use shaders, so removing the calls changes nothing on compatibility contexts.
* Add CGAL-based mesh repair, drop Netfabb remnants
Add repair_mesh() in MeshRepair.cpp, a CGAL-based repair that replaces MeshBoolean::cgal::repair(). It welds cracks, cancels coincident triangles, fixes orientation, fills holes (including planar annular regions), removes local self-intersections, and unites crossing shells with libigl's exact boolean while following the nonzero winding rule used by slicing.
FixModelByCgal now repairs each volume in place without splitting it into parts, and reports whether the mesh was modified so unchanged meshes keep their painting and ids. The texture import repair prompt no longer depends on the Windows SDK, and the FIX_THROUGH_CGAL_ALWAYS conditionals are removed. The commented-out Netfabb/WIN10SDK build code is deleted.
libslic3r_cgal now builds with CGAL_LINKED_WITH_TBB and links TBB. Add docs/HLSD/mesh-repair.md and unit tests in test_mesh_repair.cpp.
* Parallelize repair while preserving separate shells
Repair connected mesh units and shells independently in parallel, keep crossing shells separate, and limit self-intersection remeshing to local defects. Update mesh-repair documentation and tests for the revised behavior.
* Make mesh repair deterministic and report its progress
Flags written from parallel loops were std::vector<bool>, whose neighbouring
bits share a word, so shells could be kept or dropped at random; they are now
std::vector<char>. Holes of up to 100 edges are filled with the search over all
triangles: the search restricted to a 3D Delaunay triangulation picks between
equally good fills differently from run to run.
Shells of fewer than four faces are dropped, as the volumes of two triangles
back to back need not cancel exactly, and a shell that self-intersection removal
changed is split again so the pieces it cut off are dropped too.
repair_mesh() takes a progress callback in place of the cancel callback, and
Fix Model reports each stage. The object's bounding box is now invalidated even
when a cancel or an error stops the repair after some volumes were replaced.
The edge-use sort, union-find and box-overlap code shared by several stages is
factored out, the topology analysed for welding is reused to build the work
units, the orientation pass checks only the edges it cut, and the output is
written from the shells without joining them first.
* Fix Split to parts crashing on flat pieces and staying offered after a repair
ModelVolume::split() added every piece to the object and then deleted the
pieces without a convex hull. When the first piece was flat, as often in a
triangle soup, that deleted the volume being split while split() and its
caller kept using it. A part made only of flat pieces lost all its volumes,
and a soup added thousands of pieces, saving a backup of the object for each
one it deleted. Pieces without a convex hull are now dropped before any joins
the object, the volume being split keeps the first remaining piece, and a part
with no solid piece is left whole. Other parts of the object without a convex
hull are no longer deleted as a side effect.
ModelVolume caches whether its mesh can be split, and set_mesh() kept the
cached value, so after Fix Model, Simplify or Emboss replaced a mesh, Split to
parts could still be offered for a single piece. set_mesh() and reset_mesh()
now clear the cache.
* Import STEP surface models with Split compound enabled
Splitting a compound kept only its solids, so a surface model, which OCCT
reads as a compound of loose faces, gave nothing to mesh and the import failed
with "Meshing of a model file failed or no valid shape". Each shell outside a
solid now becomes a part too, and loose faces stay together as one part, as
they do without splitting.
* Add Fix All for plates and re-enable --repair on the command line
ModelVolume::repair() applies repair_mesh() to a part and replaces the mesh
only when the repair changed it. Fix Model, the new Fix All and Fix All Plates
items of the plate context menu, and the command line's --repair option, which
had been commented out since the Win10 SDK repair was dropped, all use it.
With Keep painted feature after mesh change off, Fix Model cleared the painting
of every part of an object before repairing. Now only a part whose mesh changes
loses its painting, so fixing a whole project does not wipe it. The
notification counts the models that needed no repair instead of listing them as
repaired, and objects that did not change are not refreshed.
A filament_start_gcode list shorter than the filament count made the Bambu start-of-export path throw std::out_of_range. Use get_at() like the other filament_start_gcode reads, and add a regression test.
stapler fetches the notarization ticket from Apple's CloudKit server,
and a dropped connection there (NSURLError -1005, exit 68) failed the
whole macOS universal job twice on 2026-10-12, after notarization had
already been accepted. Rerunning the job passed, but repeats the full
build and both notarizations.
Wrap both stapler calls in the existing retry helper already used for
hdiutil in this step.
GLGizmosManager can be called while its gizmo list is empty, either at startup before the canvas has loaded the gizmos or after their init() failed. reset_all_states(), open_gizmo(), open_gizmo_by_shortcut() and activate_gizmo() now return early in that case. This fixes a Debug-build assert when a new project resets the gizmos at startup, and an out-of-bounds read in release builds when a gizmo shortcut key is pressed after gizmo init failed. activate_gizmo(Undefined) still returns true, since no gizmo is open. The reset_all_states() guard was first proposed in #16074 by kenrap.
Co-authored-by: kenrap <101818165+kenrap@users.noreply.github.com>
The reason recorded for CLI_NO_SUITABLE_OBJECTS (-50) assumed a 3mf being uploaded. The CLI also slices STL, OBJ and STEP input, so the reason now states what both -50 sites check: the plate is empty, or no object on it is both printable and fully inside the print volume. The exit code is unchanged.
Adds tests/cli/test_cli_no_suitable_objects.sh.
Pop the FramePadding style var before end() in the move and rotate gizmo windows. It was pushed inside the window and popped after End(), which failed ImGui's style stack check in Debug builds.
The CLI's arrange no longer rotates objects unless --allow-rotations is passed, which matches the GUI's default. STL, OBJ and STEP inputs are arranged even without --arrange, so a part that only fits the bed when rotated now exits with -50 unless --allow-rotations is given. The CLI prints that reason to stderr.
FanMover could split a move inside change_filament_gcode to place a delayed fan command. It computed the split point from the last position the G-code set, but on Bambu printers the firmware moves the head during the T command, so the inserted waypoint sent the head to a spot the template never visits (#15789).
Both toolchange paths, set_extruder() and the Type1 append_tcr(), now wrap the template output in ;_FAN_MOVER_SKIP_START/END tags when fan speed-up or kick-start is on. FanMover keeps fan commands inside them where they are and drops the tags from the output. multifilament_config() also sizes the per-filament options that the wipe tower reads by index.
This fixes the H2C start G-code template by commenting the two
`========== record data ==========` separator lines.
Those lines are currently emitted as raw text inside
`machine_start_gcode`. Since they are not comments or G-code commands,
they can make the generated start sequence invalid or firmware-dependent
when the template is used. The fix preserves the separators as comments.
Fixes#16218.
Validation:
- Parsed the modified JSON profile with Python `json.loads`.
- Checked both `record data` separator lines now start with `;` after
unescaping the G-code string.
# Description
Re-slicing no longer leaks layers in two places. Tree supports dropped
the support layers above the topmost overhang without freeing them. And
an object that had sliced itself, then became identical to another
object (for example a copy whose original lost a per-object setting or
modifier), leaked its own layers when it switched to sharing that
object's slices; it now frees them once it takes the shared layers over.
If slicing is cancelled before that point, they still leak, as before.
Tree support also stops copying its collision and avoidance areas on
every lookup, and releases its working cache once supports are generated
instead of keeping it until the next slice. Elephant-foot compensation
now uses each object's own slice closing radius, where the first object
sliced in a session used to fix the value for every later slice.
Slicing output changes only on layers with more than one region printed
with elephant-foot compensation and no raft, and only where an object's
slice closing radius differs from the value the session had fixed.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
Adds a regression test for an object that switches to sharing another
object's slices. The `fff_print` suite passes.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
# Description
Fixes highlights in the texture displacement gizmo that stayed on screen
after the state behind them changed, or were drawn where they could not
be seen. Erase whole model, Select whole model and Capture current view
update the paint highlight at once instead of on the next right-click. A
stroke released outside the 3D view ends on the next mouse move, in
every paint gizmo, rather than leaving the brush stuck in its pressed
colour. The layer textures stop showing an image that has been replaced.
Selecting islands in the UV editor switches a Normal view to Fast, where
their highlight is visible instead of buried under the displaced relief,
and Normal comes back once no island is selected unless another view was
picked in the meantime.
No change to slicing output; the changes are confined to the paint
gizmos and the UV editor.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The island highlight was hidden under the Normal view's relief. Normal comes back once no island is selected, unless another view was picked in the meantime.
The segment draw switched to an element buffer of 8-bit indices. GCN 1/2
GPUs cannot read 8-bit indices in hardware, so the driver converts them,
and on a Mac Pro 2013 FirePro under Windows Boot Camp the toolpaths come
out as scattered spikes. 16-bit indices are supported everywhere and keep
the eight shader runs per segment.
Erase whole model, Select whole model and Capture current view update the
highlight at once. A stroke released outside the 3D view ends on the next mouse
move, in every paint gizmo, instead of leaving the brush in its pressed colour.
The layer textures no longer show an image that was replaced.
* Wait for the PR's Deps Cache in clang-tidy When deps/ Changes
* Fail the clang-tidy Deps Wait When Listing Caches Keeps Failing
* Wait for Deps Through ci_deps_cache.sh in clang-tidy
Reuse the macOS deps wait for Linux: the job names, workflow and cache
ref are passed in, and a new 'own' mode waits for the PR's own build
when it changes deps/. A PR that keeps deps/ now waits for a running base
build with the same deps, and warns when it falls back to older ones. The
wait ends as soon as the deps job fails.
Every slider in the texture displacement panel can now be given an exact
number, either by Ctrl+clicking the track or via a small edit icon at
its right end. Both routes reuse ImGui's own temp-input mode.
Each slider also gets a tooltip with its description and its range, and
the panel's `?` popup lists both ways in.
Draft: description to be refined.
## General
Nothing about texture displacement was written to the `.3mf`, so a
painted relief did not survive closing the project.
Everything the user set up is saved now, so a project reopens with the
relief intact and still editable.
### What is saved
- The layer stack: every per-layer setting (depth, midlevel, smoothing,
tiling, projection blend mode, colour) and the per-volume options
- Each layer's texture, as a file in the archive (why? because user
might want to move .3mf to another computer where custom textures are
missing, or he saves .3mf then deletes textures, and returns back to
.3mf)
- The painted region, one mask per layer slot
- The UV work: LSCM seams, island placement, per-vertex overrides
- Editor state: the selected layer and Standard/Pro, so the panel
reopens where it was left
### What is not saved
- The preview mesh. It is a second copy of the model in the file, for
something the app rebuilds quickly on reopen.
- Panel docked/undocked, which goes to the application config instead
## Format
- `<volume id>.json` - the layer stack and the per-volume options
- `<volume id>_<slot>.png` - each layer's texture, referenced from the
JSON through `path_in_3mf`
Paint masks ride along as one per-triangle attribute per slot
(`paint_texture_0..7`), beside
the attributes the existing paint tools already write.
## Compatibility
Both ways:
- an older build ignores what it does not know and still opens the
project, losing only the relief
- a newer build finds nothing in an older project and leaves the volume
as today.
They were appended at the end of test_texture_displacement.cpp, where every
other open texture-displacement branch also appends, so the four of them
cannot be merged in any order without a conflict in a file none of them
really disagree about. Same tests, moved next to the patch-frame test they
belong with.
SysInfoDialog, JumpPointSearch and WebUpdatePlugin are in no CMake list on any platform, and nothing includes ConfigExceptions.hpp, wxinit.h or Utils/Profile.hpp. MultiTaskModel holds only empty namespaces. Also drop the empty GUI_App::system_info() and comments that still mention SysInfoDialog.
TriangleMeshDeal.hpp reused Timer.hpp's guard and CloneDialog.hpp reused
SingleChoiceDialog.hpp's, so including both headers of a pair silently
dropped the second one. Use #pragma once instead.
Its inline suggestions rarely apply: clang-tidy adds a missing include
to the include block at the top of the file, which a PR seldom touches,
and GitHub only accepts review suggestions on lines in the PR's diff.
* Add a /bot fix-tidy Command That Pushes clang-tidy's Missing Includes
A maintainer comments /bot fix-tidy on a PR. A read-only job runs
scripts/run_clang_tidy.sh --fix on the PR's merge commit and returns the
patch; a second job, which never runs the PR's code, accepts it only when
it adds nothing but #include lines to C++ files the PR already changes,
then commits it on the PR head and pushes with CLANG_TIDY_BOT_TOKEN.
* Post clang-tidy's Fix as Review Suggestions Instead of Pushing It
A failed clang-tidy check now runs once more with --fix and uploads the
patch as the clang-tidy-fix artifact; the check still fails. /bot
fix-tidy downloads that patch for the PR's latest commit, checks it adds
only includes to C++ files the PR changes, and posts one review: a
one-click suggestion for each include on a line the PR's diff shows, and
the whole patch with a git apply command. Nothing is pushed and no PR
code runs in the bot, so the PR's author can run it too and no token is
needed.
* Make the clang-tidy Fix in Its Own PR-Scoped Workflow on Request
clang_tidy.yml is back to main's version. clang_tidy_fix.yml runs on
every PR push but skips its job on the first attempt; /bot fix-tidy
re-runs it, so run_clang_tidy.sh --fix runs in the PR's own context
with a read-only token and the PR's cache, never main's, and the PR's
author can ask for it. When that re-run finishes, the bot posts the
patch as review suggestions as before. When the clang-tidy step fails,
the bot also comments asking the author to run the script or the
command, edits that comment on later failures, and deletes it once the
check passes.
* Give the clang-tidy Bot Write Access to PRs It Answers
* Tell /bot fix-tidy Users How to Start a Fix Run When There Is None to Re-Run
A project no longer reads texture files from outside its archive, and a
malformed value or layer slot drops only the texture layers instead of
failing the whole load. Copies of an object that share a mesh keep their
own texture paint when saved, and projects without texture paint no
longer use extra memory to load.
# Description
With "Support multi bed types" enabled on a Qidi printer, the
first-layer bed temperature was always taken from the Smooth PEI / High
Temp Plate setting, whichever plate type was selected.
The slicer computes the right value (`first_layer_bed_temperature` in
the G-code footer follows the selected plate), but the machine start
G-code of these presets hardcoded `[hot_plate_temp_initial_layer]`:
- `fdm_qidi_x3_common` (base of X-Max 3, X-Plus 3, X-Smart 3, X-Plus 4,
Q1 Pro)
- `Qidi Q1 Pro 0.4 nozzle`
- `Qidi X-CF Pro 0.4 nozzle`
They now use `[bed_temperature_initial_layer_single]`, which follows the
selected plate type, as most other profiles do. The Qidi bundle version
is bumped to `02.04.00.17`.
With "Support multi bed types" disabled, these printers use the default
plate (Smooth PEI / High Temp Plate), so the temperature is unchanged
there.
# Screenshots/Recordings/Graphs
N/A — profile change only.
## Tests
`python3 scripts/orca_profile_tool.py check` passes.
Sliced on Qidi X-Max 3 0.4 nozzle with "Support multi bed types"
enabled, `0.20mm Standard @Qidi XMax3`, `Generic PLA @Qidi`, with a
different bed temperature set for each plate type (first layer / other
layers).
Before, with first layer set to 50–55 °C: `M190 S54` for every plate
type.
After, a two-layer print:
| Plate type | Set (first / other) | `M190` at start | `M140` at layer 2
|
| --- | --- | --- | --- |
| Cool Plate SuperTack | 51 / 61 | 51 | 61 |
| Cool Plate | 52 / 62 | 52 | 62 |
| Textured Cool Plate | 53 / 63 | 53 | 63 |
| Engineering Plate | 54 / 64 | 54 | 64 |
| Smooth PEI / High Temp Plate | 55 / 65 | 55 | 65 |
| Textured PEI Plate | 56 / 66 | 56 | 66 |
🤖 Generated with [Claude Code](https://claude.com/claude-code)4
On Linux and macOS the app follows the system theme: at startup it
overwrites the saved dark_color_mode with the system appearance. The label
colours and the cached dark mode flag were already set from the saved value
by then, and the later "changed" check reads the mode only after the
overwrite, so it never fires. For that session every dialog that maps its
colours through UpdateDlgDarkUI keeps the old mode, e.g. white dialogs on a
dark main window after switching the desktop to dark, or on a fresh install
on a dark desktop.
Refresh the label colours and the flag when the system theme differs from
the saved mode.
# Description
## Summary:
Adds favorite filament support to the AMS material selection dropdown.
Users can now click a star icon next to a selectable filament in the AMS Materials Setting dialog to mark it as a favorite. Favorited filaments appear in a new `Favorite filaments` section at the top of the dropdown, while all compatible filaments remain available in the `All filaments` section.
This PR also adds generic right-side action icon support to Orca's custom dropdown widget:
- Adds optional action icons to `DropDown::Item`.
- Adds a new `EVT_DROPDOWN_ITEM_ACTION` event for clicks on the action icon.
- Relays the action event through `ComboBox`.
- Adds `ComboBox::RefreshDropdownIfOpen()` so callers can refresh an already-open dropdown without changing normal `ForceDropdownOpen()` behavior.
The AMS favorites implementation:
- Uses existing bundled star assets:
- `score_star_dark` for non-favorited filaments
- `score_star_light` for favorited filaments
- Stores favorite filament IDs in app config under `ams_favorite_filament_ids`.
- Stores stable filament IDs instead of display names
- Uses existing C-style string escaping helpers for persistence.
- Keeps Bambu/RFID/read-only slot behavior unchanged
No new external dependencies.
**Note:** `AMSMaterialsSetting` now uses Orca's custom `ComboBox` on macOS as well, because the favorite-star UI depends on the custom `DropDown` item renderer. I was able to test this on Windows, but macOS/Linux smoke testing is still needed and would be appreciated.
# Screenshots/Recordings:
https://github.com/user-attachments/assets/41e881be-d4de-4a7d-8f99-4297567adef2
<img width="494" height="36" alt="Screenshot 2026-05-30 103432" src="https://github.com/user-attachments/assets/e7c34ed3-db7a-4705-a1cd-4bfe7ac1616a" />
## Tests
Tested on Windows:
- Built OrcaSlicer locally
- Opened the Device tab with Bambu Lab printer with AMS
- Opened the AMS Materials Setting dialog from an editable AMS slot
- Confirmed clicking a star toggles favorite state.
- Star clicks do not select the filament row.
- Dropdown remains open after toggling a favorite.
- Confirmed favorited filaments appear under `Favorite filaments`.
- All compatible filaments remain visible under `All filaments`.
- Confirmed Filament still selected as before
- Confirmed favorites persist after closing and reopening the dialog.
- Confirmed favorites persist after restarting OrcaSlicer.
- Confirmed the favorite IDs are saved in `OrcaSlicer.conf` under `ams_favorite_filament_ids`.
- Bambu/RFID/read-only filament slots continue using the existing read-only display behavior.
Not tested:
- macOS
- Linux/Wayland
These should be tested, `ComboBox` behavior not tested on macOS.
[How to Download Pull Requests Artifacts for Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Five Plater::TakeSnapshot guards in GUI_ObjectList were unnamed temporaries, so snapshots were allowed again as soon as the snapshot was taken. Naming them keeps the suppression until the edit is done, as everywhere else.
The importer assigned pick_file.empty() and pattern_bbox_file.empty(), a bool, to the string fields, storing a one-character string instead of the path.
An image laid onto a model came out upside down under the ordinary box
(tri-planar) projection, and on three of a cube's six faces it was also
mirrored left to right, which reads as a 180 degree rotation.
DecodedHeightTexture::sample() defines v = 0 as the height map's first
pixel row, i.e. its top. Every projection built v from world +Z, so v
grew upward and the image landed bottom row up. The one path that was
already right is the projective "project from view" matrix, which
measures window y downward.
Both halves of the box basis were wrong:
- v is now negated on all three world planes, so the top of an image
sits at the top of the model.
- u now carries a sign for the side of the plane the face is on, chosen
so that u x v == -normal. Without it the two faces sharing a plane get
one u and one of them is mirrored. The sign flips exactly where that
plane's blend weight (the fourth power of a normal component) reaches
zero, so the blended tri-planar sum stays continuous and the hard axis
switch the old code removed does not come back.
The same v convention was missing from the other mapping modes:
cylindrical (which also had a left handed frame, right x up rather than
up x right, so its wrap ran mirrored), spherical, the affine branch of
view projection, and the flat chart fallback of the LSCM unwrap, which
now reuses the box basis so a cube face unwraps to the orientation
tri-planar would have given it.
The basis table is exported as box_projection_axes() and the gizmo's
adjust_tangent_basis() calls it instead of restating it, so the
placement ring and offset handle cannot drift from the bake. All four
preview shaders get the matching change, including projection_axes(),
whose tangent and bitangent have to flip with the coordinate or the
relief shading inverts.
Re-baking an existing project now places its texture upright, so it will
differ from how it looked before. Geometry already baked into a 3mf is
untouched, so nothing changes until the user re-bakes.
Verified by measuring the sampled texture row and column against world
position for every mapping mode on all six faces of a cube, and by a
full bake of a quadrant coded image, which lands each quadrant within
1e-3 mm of its expected depth.
Every slider in the texture displacement panel now takes an exact number,
either by Ctrl+clicking the track or through a small edit icon at its right
end. Both routes end in ImGui's own temp-input mode: the icon only asks ImGui
to focus the slider, which is what turns it into an input box, so the parsing,
the clamping and the Enter/click-away behaviour stay ImGui's rather than
becoming a second implementation.
The mechanism lives in one helper, typed_row(), with a float and an int form
over it. float_row() and int_row() are thin wrappers on those, and the six
sliders that previously built their widget inline (brush size, smart fill
angle, edge fade amount, remesh sharp angle, refine resolution and the uniform
triangle budget) now go through the same path.
Each slider carries a tooltip with its description and its range. The range is
derived from the bounds and the format string the slider is already given, so
no row states them twice and the tooltip cannot drift from the track. The
tooltip allows hovering while disabled, since several rows are drawn inside
disabled_begin(), where a greyed slider is exactly when the explanation is
wanted.
The panel's "?" popup gains an entry for each of the two ways in, appended to
every tool's shortcut list and shown on their own when no painting tool is
selected, where that popup was previously empty.
Clearing per-export state called unique_ptr::release(), which gives up ownership without deleting. Every export uses a new GCode, so the pointer is still empty there today and nothing leaks; reset() states the intent and would free the object if a GCode were ever reused.
* ci: wait for the base branch's macOS deps instead of rebuilding them
* Hold Pull Request macOS Builds Until a Runner Is Free
A pull request's macOS build now waits in one repo-wide line, from a
Linux job, until the macOS runners every active Build all run holds or
still needs leave two free. A run is let in whole: its first arch takes
both runners and its second goes straight through, since letting arches
in one at a time deadlocks with every runner held by a run waiting for
its other arch. Push, nightly and manual runs reserve two runners until
their macOS work is done, and a pull request run keeps its two until its
universal build and tests finish, so jobs that never pass the line are
accounted for. Queued jobs count as busy. API errors and a four-hour
limit let the build in rather than block it.
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
* Count Pending Runs That Hold macOS Runners
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
* Show the macOS Admission Checks Live in the Job Log
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
* Let a Labelled Pull Request Go Ahead of the macOS Line
A pull request labelled macos-priority when its run starts waits in its
own line under the same rule, and the normal line counts every priority
run still waiting as holding two runners, so the next free runners go
to it. It still waits for runners in use and for main's reserve.
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
* Let Each macOS Arch In on Its Own Without Holding an Idle Runner
Letting a pull request run in whole deadlocks when its second arch
reaches the line behind another pull request: the run holds two
runners, the pull request in front waits for them, and the second arch
never gets to check. Each arch now waits on its own. A run with one
arch let in holds a runner only until that arch's app build is done,
and a run with both holds two until its universal build and tests
finish, so whoever is first in line always gets in eventually.
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
* Stop Holding Runners for a Pull Request's Universal Build and Tests
Holding two runners from a pull request's second arch until its
universal build and tests finished kept one idle through the slower
arch's build. Replaying the week of 10-03 against the real queue, that
idle capacity pushed the median pull request from under three hours to
over sixteen. Those jobs take minutes, so they now queue like any other
job, and a pull request holds a runner only for an arch that is still
building. Each arch then always needs exactly one runner.
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
* Count Runs That Move Between Status Lists
A run is listed as pending while any of its jobs waits in a concurrency group, so it moves to in_progress the moment its last gate leaves the line. A check made in those seconds read the in_progress list before the move and the pending list after it, missed the run, and let in a second arch the limit had no room for. Runs created in the last day are now listed by creation time as well, and the run doing the check is read directly.
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
* Log API Calls and the Rate Limit Left on Each macOS Admission Check
The calls per check grow with the number of active runs, and the token's
hourly limit is shared by the whole repository, so each check now logs how
many calls it made and what the last response says is left.
* Give the macOS Admission Jobs Fixed Names
A skipped job is listed under its unevaluated name, so every Linux and
Windows leg showed the label expression. Split the waiting job into a normal
and a priority job with fixed names, sharing their steps through a YAML
anchor.
---------
Co-authored-by: raistlin7447 <kris.austin@gmail.com>
#16284 closed the Slice-plate hover popup only when the Slice button opened
the filament grouping dialog. The dialog also opens from the plate's filament
grouping icon, the preview legend, slicing notification links and the send
dialog, and a popup still shown on those paths keeps the mouse capture on
macOS and starves the modal dialog of clicks. Dismiss it in
FilamentMapDialog::ShowModal() instead, so every path is covered.
The popup could also stay up on its own on macOS: an enter event stopped its
close timer, and when the timer fired with the cursor inside it returned
without re-arming, so without a later leave event the popup never closed. On
macOS the timer now keeps running while the popup is shown and closes it once
the cursor is outside. The popup also stops and frees its timer when destroyed.
* Complete Simplified Chinese translations
* Fix Simplified Chinese 3D view, camera and extruder terms
* Use Chinese CAD view names on the navigator cube
* feat(preset): pack bed texture and model into printer bundles
Export copies bed_custom_texture and bed_custom_model into the .orca_printer zip under bed/. The manifest lists them in printer_asset. Import extracts each file into the bundle folder and rewrites the preset to that path. A missing file is skipped. A path that leaves the bundle directory is rejected. An old bundle with no printer_asset imports as before.
* Add includes flagged by clang-tidy misc-include-cleaner
* Gate bed asset packing on an export checkbox and key printer_asset per printer
The Export Configs checkbox can leave a custom bed out of a shared printer bundle. printer_asset is keyed by each printer_config zip entry, with files under bed/<n>/, so a later multi-printer bundle does not need a second manifest. The unshipped flat printer_asset object is not read.
---------
Co-authored-by: Rodrigo Faselli <162915171+RF47@users.noreply.github.com>
Hi OrcaSlicer Team,I have completed the Thai localization for OrcaSlicer (over 7,000 lines).Please note: This is a "Hybrid Translation" designed specifically for educational purposes. Instead of direct literal translations (which often confuse beginners due to technical jargon), I have included brief explanations of the mechanical and physics principles directly within the UI strings.Example: For "TPU is not supported by AMS", the translation briefly explains why (because the material is soft, rubber-like, and will jam the long tubes).This approach was developed for the EDrobot project in Khon Kaen, Thailand to help STEM students and beginners understand 3D printing concepts immediately without a steep language barrier.[cite: 8] We focus on empowering students to learn robotics and 3D design effectively.[cite: 8] I understand this makes some strings longer than standard localizations, but it has proven incredibly helpful for beginners. I would be honored if you consider adding this to the official release as the Thai language option. I am happy to maintain this file for future updates.Thank you for building such an amazing slicer!Best regards,Chetsit Srichan (EDrobot)
The start G-code of fdm_qidi_x3_common, Qidi Q1 Pro 0.4 nozzle and
Qidi X-CF Pro 0.4 nozzle heated the bed to hot_plate_temp_initial_layer,
so with "Support multi bed types" enabled every plate type got the
Smooth PEI / High Temp Plate temperature. Use
bed_temperature_initial_layer_single, which follows the selected plate.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Description
Adds Anker's own AnkerMake PLA+ filament line to the Anker vendor
bundle: **Basic, Matte, Metallic, Silk, Glitter**. Until now the bundle
only had `Generic PLA+ @Anker`.
Each filament follows the existing Anker layout. A `@Anker base` profile
(`instantiation: false`, inherits `Generic PLA+ @Anker base`,
`filament_vendor: AnkerMake`) holds the settings. An instantiated
`@Anker` profile holds only `compatible_printers`. `filament_id` and
`setting_id` were minted with `scripts/orca_profile_tool.py
generate-id`, and `Anker.json` was re-indexed with `update-index`. The
Anker vendor version goes from `02.04.00.05` to `02.04.00.06`.
## Source of the values
The values come from the AnkerMake filament profiles bundled with
**eufyMake Studio** (Anker's PrusaSlicer 2.6 fork, AGPL-3.0, source at
https://github.com/eufymake/eufyMake-PrusaSlicer-Release), in
`profiles/Anker-ini/AnkerMake base/base.ini`. OrcaSlicer is also
AGPL-3.0.
Conversion notes:
- PrusaSlicer keys are mapped to their Orca equivalents. For example,
`bed_temperature`/`first_layer_bed_temperature` go to `hot_plate_temp*`
and `textured_plate_temp*`, `fan_below_layer_time` goes to
`fan_cooling_layer_time`, `slowdown_below_layer_time` goes to
`slow_down_layer_time`, and `extrusion_multiplier` goes to
`filament_flow_ratio`. Only keys that differ from the inherited `Generic
PLA+ @Anker base` chain are written.
- Where Studio sets retraction but leaves deretraction unset (Metallic,
Silk), `filament_deretraction_speed` is set equal to the retraction
speed. That matches Prusa's "0 = same as retract" meaning.
- **Max volumetric speed is new.** Studio ships
`filament_max_volumetric_speed = 0` (uncapped) for these filaments, and
Orca has no equivalent. The caps were derived from the peak flow that
Studio's own M5C "Fast" process pushes: Basic/Matte/Glitter 22 mm³/s,
Metallic 20, Silk 13. Please treat these as the least-validated numbers
in the PR.
- Studio has pressure advance disabled for these filaments, so no PA
values are carried over.
- Colour variants are not included because they differ only in colour.
`default_filament_colour` uses each family's base colour.
- `compatible_printers` follows Studio's `compatible_printers_condition`
for each family, limited to the 0.4/0.6 printers that `Generic PLA+
@Anker` covers:
- Basic, Glitter: M5, M5 All-Metal, M5C at 0.4 and 0.6
- Silk: 0.4 on all three, plus M5C 0.6 and M5 All-Metal 0.6 (Studio
excludes M5 0.6)
- Matte, Metallic: 0.4 only (Studio restricts these to 0.4)
- Studio also allows Basic on 0.2 nozzles. That variant is not added,
because the 0.2 Generic profiles use a separate low-flow cap.
# Screenshots/Recordings/Graphs
N/A (profile-only change).
## Tests
- `scripts/check_profile.sh` (local twin of the CI job, run against the
full profile tree with the nightly macOS validator) passes all five
checks: `profile_tool`, `validate_system`, `validate_slice`,
`validate_filament_subtypes`, `validate_custom` (fixtures v1.9.0 to
v2.4.2).
- `python3 -m unittest discover -s scripts/tests -t scripts`: OK.
- The submitter printed with these filament settings in eufyMake Studio
on the M5C's stock brass 0.4 nozzle without issues. This Orca conversion
has been loaded and sliced in Orca 2.4.2 (as user presets inheriting the
Anker generic PLA+) but not yet printed on brass. The printer now has a
different nozzle fitted.
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
wxGTK types every wxDialog as a dialog window, and Mutter offers no
maximize for anything but a normal window, so the maximize button and
shortcut did nothing on GNOME under X11. Modeless plugin windows are now
normal windows, kept off the taskbar as before. Modal ones stay dialogs
so the window manager keeps routing focus from the main window to them.
* Expose App, GL, Printer and Project Facts to Plugins
Plugins that help users file bug reports could only guess these from log lines: the app config is
deny-listed, and the project path, a project export and the device list had no API.
orca.host gains app_info(), gl_info() and selected_printer(), and Plater gains project_path() and
export_3mf_copy(path). The copy export leaves the project's file name, saved state and model
unchanged, does not write the signed-in account as the designer, and raises the audit event of
open(path, "w") so the plugin gets the same permissions as for writing the file itself.
* Remove app_language in Favour of app_info
app_info() returns the same language. app_language() was added after the last release, so no released build has it.
* Read a local plugin's version from its file header
read_install_state() let the sidecar's installed_version override the version
parsed from the entry file's PEP 723 header. That is right for cloud plugins,
whose header may be stale, but a local plugin has no other source of truth, so
the UI kept showing the version recorded at first install after the file was
edited. For local installs, use the header version.
* Keep the cloud installed version of an unsubscribed plugin
A cloud plugin kept as local after unsubscribing is rewritten with installed_from=local and no cloud_uuid, which the local-header rule could not tell apart from a real local install. Its installed_version is still the cloud one while the header may lag, so the next scan showed the stale header version. Record the conversion in the install state and exclude those plugins from the header rule. Add tests for a local plugin following its header and for an unsubscribed cloud plugin keeping its cloud version.
* Add tests for the version of a local plugin
A local plugin's version comes from its entry file header. Cover the case of a plugin edited after install, and a plugin unsubscribed from the cloud, which is now a normal local package.
* Fix decomposition of mirrored transformations
computeRotationScaling() may place the mirror on any axis, giving unstable rotation/scale for mirrored matrices. Decompose them by flipping each local axis and choosing the one needing the least rotation (lowest axis on ties). Merge the extract helpers into a single extract_rotation_scale() and simplify contains_skew() to check that the scale is diagonal.
set_scaling_factor() now preserves the mirror sign on the scaled axis. Selection::scale_and_translate() uses Transformation's rotation and scale instead of calling computeRotationScaling() directly, so scaling a mirrored instance adds no skew.
Add tests for mirrored decomposition, scaling, and skew detection.
* Use Eigen polar decomposition for instance scaling
Selection::scale_and_translate now splits the instance matrix with computeRotationScaling instead of get_rotation_matrix().
Adds a Model test checking that an instance created from another's offset, scale, rotation and mirror reproduces its matrix (as Add instance and Fill bed do).
* tidy
# Description
Restore `"reduce_infill_retraction": "1"` in the shared Bambu Lab
process profile. The BambuStudio profile sync (#15851) replaced it with
`reduce_infill_retraction_mode`, which OrcaSlicer does not support,
causing the setting to fall back to disabled.
This restores the pre-sync behavior and bumps the BBL bundle version to
02.08.00.13.
# Validation
- All 233 selectable BBL process presets inherit the restored setting
with no other resolved settings change.
- Static profile validation, system loading, filament validation, and
all 1,263 slice cases pass.
Fixes#16194
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
# Description
Orbiting or panning the 3D view no longer hitches every five seconds on
Windows and macOS while signed in to Orca Cloud. The stutter is a
regression from #15710.
No change to slicing or to the Orca Cloud connection status shown in the
GUI.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
## Summary
This PR adds first-class parallel printing support to OrcaSlicer for
printers with multiple independent X-axis carriages — the **IDEX** and
**IQEX** hardware families. The feature is named **IMEX** (independent
multi-extruder) internally, which captures the supported topology space
more accurately than either acronym alone, and is exposed in the UI
under the user-facing label **IDEX/IQEX Configuration**. It is designed
to be printer-agnostic and firmware-flexible, with Klipper,
RepRapFirmware, and Marlin all supported for per-tool Pressure Advance
and per-layer temperature management. MMU/AFC setups where multiple
logical filament slots share one physical extruder are supported via a
`physical_extruder_map` profile option. Printers whose firmware handles
copy/mirror placement internally — RepRapFirmware IDEX duplication mode
on Flashforge Creator Pro 2 / Creator 3 Pro is the canonical example —
are supported through a firmware-managed-zones flag that emits a
centered single-half slice for the firmware to fan out.
The implementation spans printer configuration, process settings UI, bed
visualization, per-plate mode selection, ghost-object rendering with
per-plate filament overrides, placement validation, pre-slice conflict
warnings, G-code generation (with per-tool PA, per-layer temperature
management, and optional center-origin slice frame), and layer preview
animation. It supports five distinct topology paradigms:
- **1 gantry × 2–4 tools** — classic IDEX with 2 tools (BCN3D Sigma,
Snapmaker J1, Tenlog Hands-2 style — two independent X carriages on a
shared Y gantry); multi-extruder on a shared rail when extended to 3 or
4.
- **2 gantries × 1 tool each** — two fully-independent XY systems
sharing a bed (Vivedino Xplorer style). Both Y rails independent —
effectively "half an IQEX."
- **2×2 grid, independent quadrants** — IQEX running four separate
prints in parallel.
- **2×2 grid, paired-gantry multicolor** — IQEX where cross-gantry tools
share color responsibilities via the **Span** tile state.
- **Firmware-managed center-origin** — slicer emits a centered
single-half slice; firmware decides where to physically fan it out.
Overlays on any of the above hardware topologies; canonical example is
Flashforge IDEX with RepRapFirmware.
---
## Feature Walkthrough
### Printer Configuration
Printer preset options that declare IMEX capability and geometry:
| Option | Type | Description |
|--------|------|-------------|
| `is_imex` | bool | Marks this printer as IMEX-capable |
| `imex_firmware_managed_zones` | bool | Emit centered single-half
slice; firmware handles copy/mirror placement (default `false`) |
| `imex_gantry_count` | int | Number of independent Y-axis gantries
(rows) |
| `imex_tools_per_gantry` | int (1–4) | Toolheads per gantry along X
(columns) |
| `imex_nozzle_clearance_x` / `imex_nozzle_clearance_y` | float |
Nozzle-to-collision-edge distance in mm (literal, not halved) |
| `imex_tool_layout` | string | Physical orientation: front-left /
front-right / rear-left / rear-right |
| `imex_mode_names` | string[] | Names of user-defined parallel modes;
first entry is the reserved `primary` sentinel |
| `imex_mode_active_tools` | string[] | Tool role assignments per mode
(e.g. `0:P,1:C,2:M,3:M`, or `0:P,1:S,2:M,3:M` with Span) |
| `imex_mode_gcodes` | string[] | Firmware macro to activate per mode |
| `physical_extruder_map` | string[] | 0-indexed map from logical
filament slots to physical extruders (for MMU/AFC — see below) |
Tool roles per mode: **Primary** (P), **Copy** (C), **Mirror** (M),
**Span** (S), **Inactive**.
### Parallel Mode Editor
[Kooha-2026-05-14-13-45-35.webm](https://github.com/user-attachments/assets/af7650e6-d636-478b-9398-3f02ec662f03)
The **IDEX/IQEX Configuration** and **IDEX/IQEX Parallel Modes**
sections live in the printer preset's **Multimaterial** page (visible
only when `is_imex = true`). The mode editor is a visual grid:
- Each row defines one named parallel mode. The first row is a reserved,
non-deletable **Primary** row stored as the sentinel `primary` in
`imex_mode_names`.
- The Primary tool is always T0; **Tool 0 Position** chooses which
physical corner it occupies, and its tile is read-only. The other tool
buttons cycle Inactive → Copy → Mirror → Span → Inactive, color-coded,
with Span offered only where it applies.
- **Span** is only offered when `imex_gantry_count >= 2` AND the tile is
on the primary's gantry row — it declares "this tool is the multi-color
partner of Primary on the same gantry," distinct from a Copy/Mirror role
on a non-primary gantry.
- A G-code field per mode holds the firmware macro to activate that mode
(e.g. `IMEX_COPY` for Klipper). A placeholder-browser button per row
opens the `EditGCodeDialog` for quick insertion.
- Modes can be added and removed (removed via a dedicated
`imex_remove.svg` ScalableButton).
- Deleting the active mode resets affected plates to Primary.
- Tool assignments are preserved across `imex_gantry_count` changes:
going from IQEX (4 carriages) → IDEX (2 carriages) → IQEX restores all
previously assigned roles, and the visible grid anchors to the gantry
row containing the Primary assignment.
<!-- TODO: replace this placeholder with a re-recorded
parallel-mode-editor video
showing the Span tile state, the remove button, and the
placeholder-browser button.
Old URL (out of date):
https://github.com/user-attachments/assets/cabf1f70-206d-4f84-9c2e-594b38b83951
-->
### Span Tile State — Paired-Gantry Multicolor
<img width="2833" height="1300" alt="2026-05-14-135412_grim"
src="https://github.com/user-attachments/assets/52203133-7380-459a-973f-0502a1edc3c2"
/>
For IQEX printers (`imex_gantry_count >= 2`), the **Span (S)** tile role
enables paired-gantry multicolor mirror mode. The same active-tools
string `0:P,1:C,2:M,3:M` is ambiguous between two distinct hardware jobs
— four independent quadrants vs. paired-gantry multicolor mirror — so
the topology has to be declared explicitly rather than inferred. Span on
the primary's gantry row is the declaration.
Behavior driven by Span presence:
- **Multicolor block rule**: requires Span on primary's gantry to allow
multi-color slicing in a parallel mode. Without Span, multi-color
slicing in a parallel mode is blocked with an actionable error message.
- **Ghost aggregation**: one aggregated ghost per non-primary gantry
when Span is present, using the column-paired representative tool.
Mixed-role gantries fall back to per-tool.
- **Zone aggregation**: one row-strip zone per non-primary gantry when
Span is present (cell at primary's column collapses col-sep;
`make_boxes` expands to full-X strip).
- **Aggregated mirror drag**: ghost translates 1:1 with primary in X
(copy-style), with X-flip baked into the mesh-local frame so geometry
still reads as mirrored. Gantries don't share an X rail, so reflecting
motion serves no collision purpose.
- **Carriage collision strip audit**: X-boundary checks require `zr ==
pri_row_k` (matching the existing Y-boundary `c == pri_col` constraint).
Prevents spurious strips on primary's right edge in paired-gantry
mc-mirror, where T3 sits diagonally and can't actually collide with
primary's carriage.
Single source of truth:
`IMEXHelpers::group_imex_active_tools_by_gantry(active_tools_str,
tools_per_gantry)`. Ghost factory and zone calculator both consume it,
so pairing logic lives in one place.
### Per-Plate Mode Selection
Each build plate has an IMEX mode icon in its toolbar (normal, hover,
dark, and dark-hover SVG variants). The mode can be set independently
per plate:
- **Left-click** cycles through all non-sentinel modes in order.
- **Right-click** opens a popup menu listing all modes as radio items
for direct selection.
- Mode changes are recorded in the undo/redo snapshot system.
- The selected mode is persisted in the 3MF project file per plate
(`imex_parallel_mode` key).
https://github.com/user-attachments/assets/5ab497e3-9c2f-476a-ac67-ef34a592395b
When a parallel mode is active alongside multi-material objects on the
same plate, a warning badge (`obj_warning.svg`) overlays the plate icon
— see **Pre-Slice Warning System** below.
### Bed Visualization + Auto-Arrange Constraints
When an IMEX mode is active, the build plate renders the carriage grid:
<img width="2761" height="1447" alt="IMEX bed visualization with active
primary zone and dimmed secondary zones"
src="https://github.com/user-attachments/assets/3ee64b7b-a3ed-40aa-ad07-fa2f1cb74edf"
/>
- The **active (primary) zone** is full brightness.
- **Inactive zones** are dimmed with a color-coded overlay (blue for
copy, orange for mirror).
- **Zone dividers** are rendered as lines across the bed.
- The grid is a full 2D layout: `imex_tools_per_gantry` columns ×
`imex_gantry_count` rows. Zone sizing is based on the **active** tool
count only — inactive tools donate their bed share to active neighbors.
- Span-present configs render as a single row-strip zone per non-primary
gantry rather than per-tool quadrants, reflecting that the firmware will
paint the multi-color across the entire strip.
- Colors are drawn from the Okabe-Ito palette for colorblind
accessibility, with a deuteranopia/protanopia-safe alternate theme.
- **Auto-arrange** is constrained to the primary zone when a parallel
mode is active — `ArrangeJob::process()` replaces full-bed `bedpts` with
the primary zone corners via `PartPlate::imex_primary_zone()`. Collision
strips are additionally registered as hard obstacles through
`m_unselected` (the working NFP placer input, not the dead
`excluded_regions` field), so placement cannot drop parts into the
danger strips.
### Placement Validation
https://github.com/user-attachments/assets/5cb5a94a-981e-4fe9-898e-7e2f891b4f40
Objects placed outside the primary zone block slicing:
- `has_imex_placement_violations()` in `PartPlate` checks each object's
convex hull against the primary zone boundaries and collision strips.
- Violations inject into the existing `update_background_process`
validation pathway — the Slice button is disabled and an error
notification is shown.
- Mirror tools additionally generate X-axis collision strips (copy tools
move in the same direction and cannot collide). Y-direction strips are
scoped to same-column tools to avoid false positives from diagonal
mirror pairs.
- Strip width is taken **literally** from `imex_nozzle_clearance_x` (the
measurement is nozzle-to-collision-edge distance, not carriage
half-width).
- Multi-color block path: a pre-slice rule blocks multi-color slicing in
parallel modes that can't physically support it (e.g. a "fake IMEX" mode
where all tools sit on a single gantry without a Span partner). The rule
is centralized in `imex_multicolor_block_reason()` with unit-test
coverage of every gating case.
### Ghost Object Rendering + Per-Plate Filament Picker
When a parallel mode is active, the slicer renders colored, transparent
**ghost copies** of primary-head instances on the plate — one per
secondary active head, transformed under its Copy/Mirror role (or
aggregated per gantry when Span is present).
- Ghosts track the primary through drag/rotate/scale/mirror and
invalidate on mode, filament-map, or `physical_extruder_map` changes.
- **Left-click** on a ghost opens the `IMEXFilamentPickerPopover` for
that ghost's head — a compact `BitmapComboBox` that writes directly to
the per-plate `imex_head_filament_map` (MMU lane override).
- Mirror ghost geometry is a **true reflection about the zone-boundary
plane** (`x = primary_zone_center.x + gantry_offset.x/2`), so the ghost
stays anchored in the target zone as the primary moves and drag reflects
correctly (primary +X → ghost −X, Y tracks 1:1).
- Hover tooltip: `"Tn → filament N"` with a color swatch. When no
filament resolves to a head, the tooltip surfaces an actionable message
directing the user to extend the extruder count in the Machine tab.
- Ghost rendering + picking are scoped to the active plate — non-active
plates don't draw stale ghosts during arrange/preview transitions, and
click-picking never falls through to a non-current plate's ghost
geometry.
- Per-plate `imex_head_filament_map` round-trips through the 3MF project
file.
### Pre-Slice Warning System
<img width="1786" height="911" alt="2026-05-14-135714_grim"
src="https://github.com/user-attachments/assets/16532d47-2da5-4472-ab9c-4b8b00303bae"
/>
Before a plate slices, IMEX parallel-mode plates are checked for three
classes of conflict:
1. **Multi-material on secondary tools** — when the mode has Copy/Mirror
tools and the plate has multiple filaments active; plate icon gets a
warning badge.
2. **Bed temperature mismatch** — any two carriages configured >5 °C
apart.
3. **Filament type incompatibility** — filaments from different type
families on active carriages.
A dismissible Yes/No `RichMessageDialog` fires once per user action from
both `on_action_slice_plate` and `on_action_slice_all`. The dialog has a
**"Don't show again"** checkbox that persists to `app_config` as
`imex_pre_slice_warnings=false`. A re-enable toggle lives in **Printer
Settings → Multimaterial → IDEX/IQEX Configuration** so suppressed
warnings can be restored.
### Firmware-Managed Zones — Center-Origin Slice
<img width="701" height="197" alt="2026-05-14-135842_grim"
src="https://github.com/user-attachments/assets/3e2dfaab-fb4f-42b1-bf9d-c88dc0c95110"
/>
The `imex_firmware_managed_zones` printer-config option (default
`false`) supports IDEX/IQEX printers whose firmware applies its own
copy/mirror offsets in non-primary modes. Canonical examples:
**RepRapFirmware IDEX duplication mode** on **Flashforge Creator Pro 2 /
Creator 3 Pro**. These printers expect a centered single-half slice at
bed origin and fan toolheads out from there — the slicer-managed
paradigm of placing toolpaths at zone-relative positions produces gcode
the firmware can't reconcile, since it'd double-apply offsets.
When the flag is on and the active mode is non-primary, the slicer
subtracts the primary zone's plate-local center from the gcode emission
frame:
- **Writer offset** is augmented by the primary zone center so emitted
gcode is centered at bed origin.
- **Processor offset** stays at plate_origin only, so the gcode-preview
visualizer renders the centered toolpath at bed center rather than at
the prepare-view zone placement. The user sees what'll physically print
after firmware fan-out.
- **`translate_to_print_space()`** is augmented for frame coherence so
`first_layer_print_min/max` placeholders consumed by user start_gcode
(e.g. Flashforge's M118 "max delta from zero" header) reflect the
centered frame.
Slice-handoff runs per-slice: `PartPlate::refresh_imex_slice_offset()`
is called from `Plater::priv::update_background_process` after
`Print::apply()`, so reslicing with mode toggled but no plate change
picks up the updated offset.
When the flag is off, all related code paths reduce to no-ops
byte-identical to standard slicer-managed behavior. The two stock-code
touches at `Print.cpp:2553-2554` (gcode_offset composition) and
`Print.cpp:2823-2829` (translate_to_print_space) are explicit additive
shifts that collapse to identity when the offset is `Vec2d::Zero()`.
### G-code Injection
The selected parallel mode's G-code is written into the output file
immediately before `machine_start_gcode`:
- Looks up the plate's active mode name in `imex_mode_names` and writes
the corresponding `imex_mode_gcodes` entry.
- Processed through `placeholder_parser_process()` first, so
Klipper-style variable substitution works **and** any `{global}`
declarations flow forward into `machine_start_gcode`.
- Primary mode's G-code field is emitted too.
- Guarded against headless CLI slicing — `ensure_imex_zones()`
short-circuits when `m_plater` is null so the CLI path (used by
upstream's regression-test CI step) doesn't segfault on
`wxGetApp().preset_bundle` dereference.
### Placeholder Parser Integration
<img width="535" height="393" alt="2026-05-14-140227_grim"
src="https://github.com/user-attachments/assets/6d059c77-38f6-4418-bfe9-3bf963e8c836"
/>
Three placeholders register under **Slicing State** and are settable
from anywhere downstream:
| Placeholder | Type | Description |
|---|---|---|
| `imex_mode` | string | The active mode name |
| `imex_mode_index` | int | Index into `imex_mode_names` |
| `imex_mode_gcode` | string | The resolved mode G-code (post-parser) |
### Per-Tool Pressure Advance — Firmware-Agnostic
`set_pressure_advance()` takes an optional tool index (default `-1`,
preserving existing behavior for all non-IMEX call sites). Per firmware:
- **Klipper**: `EXTRUDER=extruder[N]` when `tool >= 0`, bare command
otherwise
- **RepRapFirmware**: `M572 D<N>` when `tool >= 0`, `M572 D0` otherwise
(preserves the pre-IMEX output; a bare `M572` applies to whatever tool
is selected and errors when there is none)
- **Marlin 2**: `M900 K<X> T<N>` when `tool >= 0`, bare `M900` otherwise
- **Marlin Legacy / fallback**: `M900 K<X>` always
- **Repetier**: `M233 X<X> Y<X>` (X is quadratic, Y is linear; same
value applied to both)
`m_imex_parallel_mode` is set once per export from the active plate
mode. PA tool-qualification is gated on this being a **non-primary**
parallel mode — primary-mode prints emit ordinary tool-change PA exactly
like any non-IMEX printer. Secondary active tools in parallel modes
receive explicit per-tool PA at print start since they never go through
a tool-change sequence.
### Per-Layer Temperature Management
In IMEX parallel modes, all active tools (primary + secondaries) get
temperature commands in `layer_change_gcode`:
- Layer 1 temperatures only emit on the first layer; subsequent layers
use normal layer-change temperatures.
- IMEX temperature handling is consolidated into the second-layer
transition.
- Layer-change temperature commands use `M104 T<N>` with
`physical_extruder_map` translation when applicable.
### MMU / AFC Support via `physical_extruder_map`
For printers where multiple logical filament slots share one physical
extruder (MMU, AFC, toolchangers), the `physical_extruder_map` profile
option translates tool slot indices to physical extruder qualifiers
before G-code emission.
- **Fallback**: on IMEX printers, `Print::apply()` uses the profile's
map only when it has one entry per extruder (the `nozzle_diameter`
count). Anything else, including the single-entry default, is replaced
by the identity map `0..n-1`.
- **Used by**: IMEX PA emission, layer-change temperature commands
(`M104 T`, `EXTRUDER=`, `M572 D`), ghost color resolution, ghost cache
key, tooltip lookup, click gate.
- **Profile authoring example** for a 7-slot printer with a 4-lane MMU
on extruder 0 and three independent direct drives on extruders 1/2/3:
```json
"physical_extruder_map": ["0","0","0","0","1","2","3"]
```
Non-MMU printers need no action — the identity fallback handles them.
- A centralized helper in `IMEXHelpers`,
`effective_physical_extruder_map(explicit_pem, nozzle_count)` (with a
`PresetBundle` overload that prefers the project's map over the
printer's), routes all PA/temp/ghost-color lookups through a single code
path.
- **No UI for editing the map** in this PR — non-trivial MMU/AFC layouts
require profile-authoring (hand-edit the printer JSON) or an updated
printer profile shipped by a vendor. A future enhancement would expose a
per-slot extruder picker in the Multimaterial section.
### Layer Preview Multi-Carriage Animation
The sequential preview (scrubber) animates all active carriages
simultaneously:
- One toolhead marker (colored cube) per active carriage, in addition to
the primary.
- Secondary marker colors: cyan (T1), yellow (T2), magenta (T3).
- **Copy** tools: marker placed at the same relative position within
their bed zone as the primary is in the primary zone.
- **Mirror** tools: reflect across the target zone's facing edge
(left-of-copy reflects across copy zone's left edge; right-of-copy
across the right edge). Y position is always zone-relative copy (all
tools on a row share a physical Y rail).
- Carriage footprint boxes use per-carriage `box_offset_x/y` so the
nozzle marker sits at the physically correct edge of the footprint —
zone-based X by default, collision-side edge for Mirror; gantry-behind Y
for back-row primaries, flipped for front-row primaries with a back-row
secondary.
- The filament usage legend notes the active carriage count and mode
name (e.g. `IMEX: ×2 (copy_mode)`).
- **View menu toggle**: `View → Show IDEX/IQEX Toolhead` (visible only
on the Preview tab, only when the active printer is IDEX/IQEX)
hides/shows the per-carriage toolhead representation during preview
playback. State persists to `app_config` as `show_imex_toolhead_boxes`.
Useful when scrubbing through dense toolpaths and the boxes get in the
way of seeing the underlying geometry:
[Kooha-2026-05-14-14-04-09.webm](https://github.com/user-attachments/assets/9869cd4d-3b60-4a6b-8dca-7ee1fc573ac1)
Copy mode:
https://github.com/user-attachments/assets/0774d285-ba36-4b86-a594-fb6572c3aede
Mirror mode:
https://github.com/user-attachments/assets/51528926-94dc-4d90-9b1a-1b42d9044be7
---
## Known Limitations
**Per-layer G-code collision detection**
The placement-time zone check catches gross violations (object placed in
wrong zone) but does not verify that toolpaths on any given layer
maintain adequate X separation between adjacent carriages. A per-layer
check via `ConflictChecker` was designed but deferred. Without it, a
print that passes placement validation could still crash carriages if
the primary object's toolpaths reach too close to a zone boundary.
**Brim avoidance of IMEX zones**
Standard `bed_exclude_area` exclusion zones are already respected by
brim generation. IMEX collision strips are not — the brim generator has
no visibility into them. The correct fix (feeding computed strip
polygons from `PartPlate` through the `Print` object to `Brim.cpp`) was
designed but deferred. In practice, users should leave adequate
clearance between printed objects and zone boundaries to account for
brim width.
**`extruder_printable_area` integration**
IMEX zones are not clipped against per-extruder printable polygons, and
there is no violation check for placing an object outside the
intersection of its active extruder's printable area and the IMEX
primary zone. Deferred pending clarification on the tool→extruder index
mapping.
**Ghost rendering in firmware-managed mode**
When `imex_firmware_managed_zones` is on, ghost rendering is suppressed
entirely. The existing `imex_head_transform` math is slicer-managed
semantics (places ghosts at `primary_zone_center + gantry_offset`) and
produces wrong positions when the toolpath is being emitted in a
centered frame. Proper firmware-managed ghost rendering — showing where
copies/mirrors will physically print after firmware fan-out — needs new
transforms designed around firmware-frame positions rather than a
coordinate-flip of the slicer-managed ones. Deferred to a follow-up.
**Filament-accurate multi-region ghost color**
For paired-gantry multicolor mirror (Span mode), the aggregated ghost
currently uses the representative tool's filament color as a single
solid swatch. A future enhancement would render the ghost split into
per-source-tool regions, each colored by the secondary tool that mirrors
it. Slicer-side correctness is already shipped — the G-code emits the
right T-codes; the ghost is a visual aid only.
**Ghost path overlays**
Secondary carriage toolpaths are not rendered in the layer preview. The
secondary markers animate correctly, but the paths they would trace are
not drawn. Adding ghost path rendering would require duplicating and
offsetting the toolpath geometry per secondary carriage, which is a
significant addition to the libvgcode rendering pipeline.
**No per-mode slicing**
All carriages in a mode execute the same sliced toolpaths (transformed
per zone). There is no support for slicing different objects for each
carriage independently within one mode.
**Global default IMEX mode**
There is no job-level IMEX default mode. Each plate's mode must be set
individually (default is always Primary). A future improvement would add
a global default in the sidebar (following the same pattern as bed type
and nozzle diameter), with per-plate overrides.
**Dynamic GL-rendered mode icons**
The per-plate icon currently uses static SVGs. A GL-rendered
carriage-grid icon that visually represents the mode's tool layout would
be a nicer UX but is deferred.
**Slice-all thumbnail icon refresh**
On the "Slice All" path, plates flagged as IMEX-violated do not refresh
their toolbar icon. Root cause and fix identified, not yet applied.
**Start-G-code filament placeholder**
The resolved per-head filament map is not yet exposed as a
`PlaceholderParser` vector. Exposing it would let MMU firmware macros
pre-load lanes before print start.
**No UI for `physical_extruder_map` authoring**
The MMU/AFC slot-to-physical-extruder map is currently profile-only —
there's no in-app dropdown or editor for it. Users with non-trivial
MMU/AFC setups must hand-edit the printer-preset JSON (or rely on a
vendor-supplied profile). A future enhancement would expose a per-slot
extruder picker in the Multimaterial section so users can declare the
mapping without touching JSON.
**Bundled IQEX printer profile**
A full IQEX printer profile with cover image, bed mesh, and matched
process/filament profiles is not bundled with this PR. Users must
currently author their own printer preset. Deferred to a follow-up
profile-only PR.
**Mode lifecycle gap**
Per-plate mode is stored as a string (the mode name). If a mode is
renamed or deleted from the printer preset after a project is saved, the
plate's saved mode name will not resolve and will silently fall back to
Primary on next load. A warning on load would be a useful addition.
**Firmware-managed prepare→preview frame jump**
In firmware-managed-zones mode, the prepare view shows the part at its
placed-in-zone position while the preview view shows the centered slice.
This is intentional — the views truthfully represent the prepare frame
vs. the post-firmware-fan-out frame — but the visual jump can be
confusing on first use.
---
## Firmware Assumptions
- The implementation assumes the firmware handles all carriage
synchronization and offset math. OrcaSlicer only injects the
mode-activation macro before machine start and is entirely dependent on
user configuration.
- Per-tool Pressure Advance is firmware-aware (Klipper / RRF / Marlin 2
/ Marlin Legacy / Repetier). Layer-change temperature commands use `M104
T<N>` with `physical_extruder_map` translation.
- Tested against Klipper on a real IQEX printer. RepRapFirmware gcode
emission is exercised via profile-driven slice tests (Flashforge Creator
Pro 2 profile); Marlin and RRF paths have not been validated on real
hardware print runs.
- Mid-print mode switching is explicitly not supported. The mode is
locked at print start.
- Firmware-managed-zones mode requires the printer's firmware to
translate centered slice coordinates into physical toolhead positions;
the slicer does not attempt to model the firmware's offset logic.
---
## Files Changed (High-Level)
| File | Change |
|------|--------|
| `src/libslic3r/PrintConfig.{cpp,hpp}` | IMEX config option definitions
+ declarations, `IMEXMode` enum, `physical_extruder_map`,
`imex_head_filament_map` (plate option), `imex_firmware_managed_zones` |
| `src/libslic3r/Preset.cpp` | IMEX keys registered in printer and
process preset option lists |
| `src/libslic3r/PrintApply.cpp` | `physical_extruder_map` identity
fallback on IMEX printers |
| `src/libslic3r/Print.{cpp,hpp}` | IMEX slice-offset field + accessors;
`translate_to_print_space` augmentation for frame coherence |
| `src/libslic3r/GCode.{cpp,hpp}` | Mode G-code injection, per-tool PA
emission (firmware-agnostic), per-layer temperatures for all active
tools, `m_imex_parallel_mode` state, `set_gcode_offset_with_imex_shift`
writer/processor split |
| `src/libslic3r/GCodeWriter.{cpp,hpp}` |
`set_pressure_advance(tool_index = -1)` per-firmware implementation |
| `src/libslic3r/IMEXHelpers.{cpp,hpp}` | `imex_head_transform`
(Primary/Copy/Mirror), `parse_imex_active_tools`,
`imex_primary_tool_for_mode`, `effective_physical_extruder_map`,
per-head filament resolution, `group_imex_active_tools_by_gantry` (Span
pairing), `compute_imex_slice_offset` (firmware-managed) |
| `src/libslic3r/Format/bbs_3mf.cpp` | Per-plate IMEX mode and
`imex_head_filament_map` serialization |
| `src/slic3r/GUI/Tab.{cpp,hpp}` | `IMEXModesCtrl` in Multimaterial
page, tool assignment persistence across gantry count changes, `primary`
sentinel handling, Span tile state, firmware-managed-zones checkbox,
null guard in `clear_pages()` |
| `src/slic3r/GUI/PartPlate.{cpp,hpp}` | Zone visualization (Span
row-strip aggregation), placement violation detection, per-plate mode
icon, ghost volume rebuild on mode/map/object mutation, warning-badge
overlay, pre-slice warning collection, `refresh_imex_slice_offset()`,
ghost suppression in firmware-managed mode, headless-CLI guard |
| `src/slic3r/GUI/Plater.cpp` | Validation pathway injection, per-plate
mode popup, ghost click handling, pre-slice warning dialog with "Don't
show again", IMEX multimaterial conflict routing, per-slice IMEX offset
refresh in `update_background_process` |
| `src/slic3r/GUI/GLCanvas3D.cpp` | Ghost rendering with per-head
filament color and translucent blending, picking via volume composite
id, active-plate scoping |
| `src/slic3r/GUI/GCodeViewer.cpp` | Multi-carriage marker animation,
Mirror math fix, carriage footprint box offsets, legend annotation |
| `src/slic3r/GUI/IMEXFilamentPickerPopover.{cpp,hpp}` | Ghost-click
filament picker |
| `src/slic3r/GUI/Jobs/ArrangeJob.cpp` | Auto-arrange constrained to
IMEX primary zone + collision strip exclusion via `m_unselected` |
| `src/slic3r/GUI/OG_CustomCtrl.cpp` | Empty `option_set` guard for
widget-only lines |
| `resources/images/plate_imex_mode*.svg`, `imex_remove.svg` | Per-plate
mode icons (light/dark/hover variants), remove button |
| `tests/libslic3r/test_imex_helpers.cpp` | Coverage for IMEX head
transforms, `effective_physical_extruder_map`,
`imex_multicolor_block_reason` gating,
`group_imex_active_tools_by_gantry` Span pairing,
`compute_imex_slice_offset` |
---
## Testing Notes
Validated on an IQEX printer (4 carriages, 2×2 grid, Klipper firmware)
and against the Flashforge Creator Pro 2 profile (2-tool IDEX, RRF
flavor, center-origin bed) with the following configurations:
- **Primary only** — baseline, no regression vs. standard
single-extruder workflow
- **Copy mode** (T0 Primary, T1 Copy, same row) — carriage markers
animate in sync offset by strip width; ghost tracks drag/rotate/scale
- **Mirror mode** (T0 Primary, T1 Mirror, same row) — T1 marker reflects
T0 across zone-boundary plane; ghost drag reflects X correctly while Y
tracks 1:1
- **Cross-row copy + mirror** (T0 Primary row 0, T2 Copy row 1, T3
Mirror row 1) — T2 follows T0's zone-relative position; T3 mirrors T2's
X; Y shared per row
- **4-tool copy mode** (T0 Primary, T1/T2/T3 Copy) — zone sizing
correct, all four markers + ghosts render in sync
- **Paired-gantry multicolor (mc-mirror, Span)** — `0:P,1:S,2:M,3:M`:
ghost aggregates to one per non-primary gantry, X-flipped, drag tracks
1:1 with primary; collision strip audit doesn't fire spurious strips on
primary's right edge
- **Firmware-managed copy mode** (Flashforge Creator Pro 2) — emitted
gcode coordinates are centered at bed origin (X = −cube_half_width …
+cube_half_width), `first_layer_print_min/max` placeholders evaluate
symmetrically, M118 header produces correct "max delta from zero" values
for Flashforge's existing template, ghosts suppressed
- **Firmware-managed flag toggle** — flipping the checkbox off restores
slicer-managed iMEX behavior (toolpath at zone position)
byte-identically; flipping on restores center-origin slice
- **Delete active mode** — no crash; plate resets to Primary
- **Placement outside primary zone** — slicing blocked with error
notification
- **Multi-material conflict warning** — plate icon gets warning badge;
Yes/No dialog fires on slice; "Don't show again" checkbox persists
- **Multi-color block rule** — `0:P,1:C` on a single gantry blocks at
slice time with actionable error message
- **Ghost filament picker** — left-click on ghost opens picker;
selection writes to `imex_head_filament_map` and round-trips through 3MF
save/load
- **Per-plate mode selection** — left-click cycles modes, right-click
shows popup, undo/redo correctly reverts mode changes, mode persists
through project save/load
- **Gantry count change round-trip** — reducing from IQEX (4 carriages)
to IDEX (2 carriages) and back restores all previously assigned tool
roles
- **MMU/AFC `physical_extruder_map`** — explicit map routes PA and
temperature commands to correct physical extruder qualifiers;
auto-derive fallback preserves 1:1 behavior for non-MMU printers
- **Headless CLI slicing** — `orca-slicer --slice project.3mf` does not
segfault on IMEX-enabled printers
- **Dark mode** — all four plate-icon variants render correctly
- **Pre-slice warnings suppress + restore** — `app_config` flag flips on
checkbox; re-enable toggle in Multimaterial config restores dialog
- **Stock non-IMEX printer regression check** — slicing a single-head
Voron Trident 350 profile produces byte-identical gcode vs. baseline;
the firmware-managed flag and Span tile state both no-op when not
applicable
- **Full unit test suite** — 247/247 passing, including Layer 1 coverage
for `compute_imex_slice_offset`, `group_imex_active_tools_by_gantry`
Span pairing, and `imex_multicolor_block_reason` gating
Several warning and info dialogs passed wxYES as their only button, so
the only choice was "Yes" even though nothing is asked. Use wxOK.
None of the callers act on wxID_YES. They ignore the result, except
Field.cpp, which only checks that it's nonzero, and every button id is.
wxYES_DEFAULT is 0, so dropping it next to wxOK changes nothing.
Also drop the empty `if (ShowModal() == wxID_YES) {}` in
Plater::priv::load_files.
* Moonraker: pass print=true in upload — fixes Upload & Print race with power-on-upload
Closes#14945.
Upload & Print on the Moonraker (Klipper) host type failed with
HTTP 503 "Klippy Host not connected" on any printer that Moonraker
powers up in response to an upload (the [power] on_when_upload_queued
feature). The file landed on disk, the print never started, and the
user hit an error dialog.
Root cause: after POST /server/files/upload succeeds we immediately
fire POST /printer/print/start. On a cold printer that Moonraker just
powered up, Klippy is still coming up when /printer/print/start
arrives, so Moonraker returns 503.
Fix: add `print=true` to the upload multipart form. Moonraker's own
upload endpoint queues the print inside the upload response — and
when a [power] device with on_when_upload_queued is configured, it
powers the printer on and waits for Klippy READY before starting.
That's the whole point of the power-on-upload feature; our second
POST was defeating it.
Also read `result.print_started` from the upload response — when
true, skip our explicit /printer/print/start (Moonraker handled it);
when false (older Moonraker or buddy-fork that ignores the print
flag), fall back to the explicit call so the existing behaviour is
preserved for those servers.
Reporter and root-cause identification: @RubenOllesch.
(cherry picked from commit bd442155ba)
* Moonraker: treat print_queued as Moonraker owning the print
Reading only result.print_started missed the exact case this PR set out to
fix. Moonraker's upload response carries two flags:
print_started : it began the print immediately
print_queued : it accepted the job but has not started it yet
The power-on path (`[power] on_when_upload_queued`) is the second one:
Moonraker queues the job, powers the printer up and waits for Klippy to
report READY, so it answers print_started=false, print_queued=true.
With only print_started read, moonraker_started_print stayed false, the
fallback fired, and our explicit /printer/print/start hit the same not-ready
Klippy that produced the original 503 — i.e. the fix did not fix#14945 for
the configuration that reported it.
Verified the response schema against Moonraker v0.11.0 (API 1.5.0); an
upload with print=true on a ready printer returns:
{"action": "create_file", "item": {...},
"print_started": true, "print_queued": false}
Both fields are present, so reading print_queued is safe on this version and
the `false` default keeps older hosts on the existing fallback path.
Caught by @raistlin7447 in review of #15032; the fix is their suggestion.
(cherry picked from commit 43e8eff003)
* Moonraker: read the upload reply's fields at top level
Moonraker's FileUploadHandler writes the upload result straight to the
response instead of wrapping it in {"result": ...} like the endpoints
registered through register_endpoint. item.path, print_started and
print_queued are therefore top-level keys. Reading them under result.
silently fell back to the local filename and to "not started", so the
explicit /printer/print/start still ran after every upload.
Also correct the comments on when Moonraker queues a job instead of
starting it, and on what it renames on upload.
Arrange read the zones of the selected plate only and applied them to
every bed it packed. Arranging all plates with a plate in Primary
selected spread a parallel plate's parts across its whole bed, and with
a parallel plate selected, a Primary plate's parts were squeezed into
the zone. Arranging a single plate other than the first also ignored its
collision strips: they were tagged with the plate's index, while the
plate packs into the arranger's bed 0.
The IDEX/IQEX constraints now live in ImexArranger (IMEXArrange), and
ArrangeJob only snapshots each plate's zones on the main thread and
calls it:
- When every bed, plates the arrange may add included, has the same
primary zone, the bed shape is that zone, so parts stay centered in it.
- Otherwise each plate in a parallel mode fences off the rest of its bed
with fixed items on its own bed, and its parts are then arranged again
inside the zone, so they sit centered rather than piled against the
zone edge nearest the bed's center. Plates in Primary keep the whole
bed, and plates the arrange adds take the process preset's mode.
- Zone edges inside the bed get the bed's own edge margin, and strips
also keep room for the brim.
- A part too big for its zone, which libnest2d's first-fit retry places
across the fixed items, is left unarranged.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
# Description
Follow-up to #16242, which turned texture colour mixes into mixed
filament slots but created a slot for every possible mix, even when only
previewing. A new Mixed colors setting (default 8) caps how many mixed
filaments a bake adds. They are picked from the texture's own colours,
and only the ones the bake actually paints with are created. Previews no
longer touch the project's filaments, both previews show a mix in the
colour its slot will have, and the bake paints each mix with the slot it
got, which went wrong whenever the project already held mixed slots.
It also improves the colour preview: the green paint highlight no longer
covers the colours, and the colours a bake writes stay visible in the
gizmo.
Only texture displacement changes. Models without a colour layer behave
as before, and there is no change to project or profile formats.
# Screenshots/Recordings/Graphs
**Before**
https://github.com/user-attachments/assets/554a0d0d-cd9f-4095-840e-ca764dde6d52
**After**
https://github.com/user-attachments/assets/7fe10431-8c3f-4759-bd3f-78b4f7a12fc7
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
* fix: hide symbols of the bundled static openssl
* fix: hide the bundled static OpenSSL symbols on Linux
* fix: relink _ssl/_hashlib when OpenSSL recipe changes
Plugin Pages capabilities (top-level notebook tabs) were missing from the Speed Dial because ActionRegistry only ingested Script capabilities.
Enumerate and subscribe to Pages as well. Launching a page action switches the notebook to that page, swapping it into the visible slot first when it lives behind the overflow dropdown.
* Rebuild the G-code line offsets after post-processing scripts run in place
* Clamp the G-code window reads to the mapped file size
* Add tests for rebuilding the G-code line offsets
* Include <mutex>, <ios> and boost/filesystem/operations.hpp where they are used
* Keep the preview's G-code lines and highlight in step with post-processing scripts
---------
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
The green highlight and tint no longer cover a colour preview, and return while a stroke is
painted. The Fast view also keeps the other parts of a multi-part object.
The new Mixed colors setting caps how many mixed filaments a bake adds. They are picked from the
texture's colours, and only the ones the bake paints with are created. Previewing no longer creates
filament slots, both previews show a mix in its slot's colour, and the bake paints each mix with the
slot it actually got.
* Block undo and redo while a background job runs
A job is queued against the model as it stands and hands its result back
when it finishes, so undoing underneath it leaves that result landing on
geometry it was never computed for. Undo and redo now wait for the job
and say so, and can_undo()/can_redo() report the same, so the toolbar
and the menu items stay in step.
* Say what to do about the running operation, not just that it blocks
Review feedback: "Stop it first" is the only way out the old text offered,
and stopping is rarely what the user wants. Waiting for the operation to
finish works just as well, so the notification now names both.
A `<metadata>` element without a `type` attribute makes
`get_attribute()` return nullptr, which was then assigned to a
`std::string` and read as a C string.
Check it the way the sibling metadata handler already does, and stop the
parse.
Regression test in `tests/libslic3r/test_amf.cpp`, a new file: the suite
had no AMF test at all.
# Description
Belt printer support is now part of `main` (#14394), so the separate
`_belt` nightly builds from the `belt-printer` branch are no longer
needed. Nightly builds are published from `main` only again, under their
usual names. The README drops its section on the parallel belt builds
and lists belt printer support among the main features, crediting its
main author, Joseph Robertson (@HarrierPigeon). The nightly release page
has already been updated to match.
No change to slicing output or to main's nightly asset names — CI and
README only.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The project carried the layer stack but not the user's place in it: a reopened
project landed on slot 0 in Standard, whatever they had been editing. The active
layer and Standard/Pro now live on the volume, next to the rest of its options,
and come back with it. Docked/undocked goes to the application config instead -
it describes the user's window, not the model, so opening someone else's project
must not move their panel.
A layer whose image_data was empty at save time was written with no texture and
came back with none. It now falls back to the file the layer came from, which the
importer put in the user's own texture library, so a project reopened on the same
machine keeps its pixels.
The texture is always stored as .png: the importer converts whatever the user
picked, so the extension was being taken from a name that no longer described the
bytes.
Also adds v2_flip_edges to the cereal field list, which the undo stack was
silently dropping.
# Belt printing for OrcaSlicer
This merges the `belt-printer` branch into `main`. It adds support for
conveyor-belt ("infinite Z") printers like the BabyBelt Pro, the
IdeaFormer IR3 V2 and the CR-30 family, along with the things that had
to grow around it: tilted-plane slicing, supports that terminate on the
belt, belt-aware brims, a purge tower that works without a flat bed,
multi-colour belt prints, a preview that shows the part the way it was
designed, and starter profiles.
It has been a long road (the first attempt was #12733 back in March, the
pipeline has been rebuilt twice since) and a lot of people have put work
into it. Credits are at the bottom; please tell me if I've missed
anyone.
Closes#2628, closes#11344, closes#6885, closes#9004, closes#14188.
---
## How belt printing works
Slicers assume the bed sits on the XY plane and layers stack along Z. A
belt printer breaks that in two ways at once: the belt is tilted
(usually 45°) and the axis that advances the belt is the printer's Z, so
the "bed" lives on the XZ plane and is, in principle, infinitely long.
<img width="6360" height="2702" alt="A flat bed vs a tilted belt"
src="https://github.com/user-attachments/assets/3f5542f2-6dab-42bf-9233-f96d863b40c8"
/>
Rather than teach every part of the slicer about tilted beds, the branch
leaves the slicing engine alone and transforms what goes in and what
comes out. The pipeline has five steps.
<details>
<summary><b>The five steps in detail</b></summary>
### 1. Pre-slice rotation
The mesh is rotated by the belt angle before slicing
(`BeltSliceStrategy`, `BeltTransform`), so that the ordinary horizontal
slicer produces layers that are actually tilted planes relative to the
part. Rotation is about X by default (belt along Y); Y rotation and a Z
lift are supported too. A "global" mode rotates every object about one
common origin, which is what keeps several parts on one belt consistent
with each other.
### 2. Slice
Nothing special. The mesh is rotated, the slicer does what it always
does. This is the reason most of Orca's features (walls, infill, seams,
ironing, painting, …) just work on a belt without belt-specific code.
### 3. Supports
Supports are generated after slicing against a virtual floor: the belt
surface, expressed in the rotated frame (`BeltFloorContext`). Normal,
tree and organic supports all terminate on that plane instead of on Z=0,
and nothing may be generated below it. Painted supports and seams are
transformed with the same `trafo_sliced()` as the layers.
### 4. G-code back-transform
The G-code is rotated back into the model's Cartesian frame
(`BeltBackTransform`). A nice side effect: with step 5 switched off you
can slice at a non-standard angle and print the result on a normal
printer. The first belt-sliced Benchy was printed exactly that way, on a
Sovol SV08.
### 5. Machine frame
The printer doesn't know its bed is tilted. To make straight walls come
out straight, the axes are remapped (the default mapping is X → reversed
X, Y → Z, Z → Y) and the result is sheared and scaled:
```math
\begin{bmatrix} X \\ Y \\ Z \end{bmatrix}
\longrightarrow
\begin{bmatrix} X \\ \dfrac{Y}{\cos\alpha} \\ Z + Y\cdot\tan\alpha \end{bmatrix}
```
This lives in `GCodeWriter` behind a small `MachineKinematics` strategy
(`BeltKinematics` on belts, identity otherwise), so the writer itself
has one code path. The slicing angle and the machine angle can differ if
you want to, e.g. slice at 30° on a 45° machine — mind your nozzle
clearance if you do.
</details>
A short recording of the back-transform from the original PR:
https://github.com/user-attachments/assets/cdb9cc83-711d-48b7-9d86-a014a32c5e8e
---
## What had to change to make it work
Belt mode is gated on the `belt_printer` printer setting; with it off,
every code path below is the old one.
<details>
<summary><b>Slicing and geometry</b></summary>
- `PrintObjectSlice` / `PrintObject`: the rotation, Z lift and
per-object layer-grid shift, plus the invalidation that goes with them.
Modifiers and painted volumes are transformed with the same matrix as
the model.
- `FirstLayerPlane`: "the first layer" on a belt is a band along the
belt, not the first slicing layer. First-layer speed, line width and the
fan band are measured against it.
- `Print::validate`: clearance checks against the gantry instead of the
printable height; skirt, raft, draft shield, the classic prime tower,
arc fitting, spiral vase with brim, and scarf-joint seams are refused or
disabled on belts because each of them either doesn't exist on a belt or
moves the belt the wrong way (a scarf seam starts one layer low, which
on a belt is a 0.28 mm back-step into the previous layer at every seam).
- Z-hop defaults to 0 on belt profiles; a lift on a belt is a belt move.
</details>
<details>
<summary><b>Supports</b></summary>
- `SupportMaterial`, `TreeSupport`, `TreeSupport3D`, `TreeModelVolumes`:
a shared `BeltFloorContext` provides the belt plane; supports clip to
it, extension layers are numbered sequentially, and the first-layer
flange that used to be stamped under everything is gone.
- `build_plate_tilt_x/y` (from #12733) is now derived from the slicing
rotation in `Print::apply`, so GUI and CLI agree; it's capped below 90°.
- Organic supports that reach the belt no longer produce negative flow.
</details>
<details>
<summary><b>Brims (#15155)</b></summary>
A belt first layer is effectively a single line of contact, so a brim
matters more than usual. `BeltBrim` generates per-layer bands along the
belt plus an "apron" ahead of the part, in the object's brim filament.
Two belt-specific settings came with it: **Leading edge brim length**
(more lines on the side printed first) and **Extra brim width** (across
the belt), plus a **Leading edge only** brim type.
<img width="1849" height="1043" alt="belt brim"
src="https://github.com/user-attachments/assets/f963ed8e-53e7-48f8-a495-123cb9ae27f7"
/>
</details>
<details>
<summary><b>Multi-colour: the belt purge tower</b></summary>
The classic wipe tower can't exist on a belt (its G-code bypasses the
transform and it needs a flat bed to stand on). Instead the GUI
generates a long, thin "purge prism" beside the parts, flush with the
far edge of the belt, one per plate (`BeltPurgeTower.cpp`). It is a real
model object, so it is sliced like everything else, and
`Print::_plan_belt_purge` routes every filament change into it via
`flush_into_objects`. It's sized from the flush matrix, split into one
island per simultaneous tool change, snapped onto the parts' layer grid,
cut off after the last colour change and stripped of infill no change
claimed, so what prints is a good deal smaller than the model you see in
Prepare.
</details>
<details>
<summary><b>G-code generation and cooling</b></summary>
- `GCodeWriter` + `MachineKinematics`/`BeltKinematics`: the
back-transform, axis remap, shear and scale, lifts that are belt moves,
and the first-layer travel speed.
- `GCode.cpp` / `BeltGCode`: a belt header (slicing rotation, remaps,
machine tilt) that the preview reads back; it is written outside the
optional header block so printers with a BTT TFT thumbnail still get it.
Exclude-object outlines are emitted in the plate frame.
- `CoolingBuffer`: the "first layers" the fan stays off for are a band
above the belt, marked per extrusion segment by the generator
(`;_BELT_BAND_START/END`) and honoured on every layer.
- `GCodeProcessor`: belt header parsing, start-G-code Z handling, and
height checks that don't compare belt travel against the printable
height.
- `ToolOrdering` / `BeltPurge`: filament changes are detected by
scanning the ordering (an apron layer never carries the first-layer
flag).
</details>
<details>
<summary><b>GUI</b></summary>
- Printer settings tab: the belt group (slicing rotation, angle, global
mode, infinite Y, purge tower, floor settings). The axis remap and
pre-slice remap options are Develop-mode only.
- `ConfigManipulation`: everything that doesn't apply on a belt is
greyed out (skirt, raft, draft shield, the wipe tower group, scarf
seams, …).
- Preview: a "designed view" that back-transforms the toolpaths onto the
model so you see the part upright, with `B` toggling the raw
machine-frame G-code (`GCodeViewer`, `Shortcuts`). The tilt comes from
the belt header, so imported G-code behaves.
- Arrange: parts are packed from the end of the belt that prints first,
colours are grouped into runs so each filament change happens once, the
purge tower's strip and the brim width are reserved, and piles aimed at
an off-centre `best_object_pos` are clamped to the bed (`Arrange.cpp`,
`ArrangeJob.cpp`, one clamp in `libnest2d`).
- `PartPlate`: plate icons stay in the gap between plates on a long,
narrow bed; the plate is open along Y for containment tests on an
infinite-Y belt.
- Calibration: a belt temperature tower (overhang variant) that slices
correctly on a tilt.
- The old tilted-bed rendering in Prepare was dormant and has been
removed; the bed is shown as the slicing pipeline treats it.
</details>
<details>
<summary><b>Config options</b></summary>
Printer: `belt_printer`, `belt_printer_infinite_y`,
`belt_slice_rotation`, `belt_slice_rotation_angle`,
`belt_slice_rotation_global`, `belt_preslice_global`,
`belt_frame_tilt_decouple`, `belt_frame_tilt_angle`,
`belt_support_floor_mode`, `belt_support_floor_offset`,
`belt_support_z_offset_mode`, `enable_belt_purge_tower`,
`first_layer_plane`, `first_layer_plane_offset`,
`first_layer_plane_thickness`, `build_plate_tilt_x/y`,
`gcode_back_transform`, `gcode_remap_x/y/z`, `preslice_remap_x/y/z`,
`preslice_remap_global`.
Process: `belt_purge_tower_width`, `leading_brim_length`,
`extra_brim_width`, brim type `leading_edge_only`. Object:
`belt_purge_tower_object`.
All of them have defaults that leave non-belt printers untouched, and
old `belt_support_floor_mode` values map to `none`.
</details>
<details>
<summary><b>Tests</b></summary>
`tests/libslic3r/test_belt_brim.cpp` and `test_arrange.cpp`, and belt
cases in `fff_print` (`test_print.cpp`, `test_skirt_brim.cpp`,
`test_gcodewriter.cpp`, `test_gcode_processor.cpp`): scarf gate, fan
band, gantry clearance, organic supports on the belt, brim with and
without the purge tower, apron widths, machine mapping at non-45°
angles, first-travel lift, start-G-code Z, arrange clamp and colour
grouping. All three suites pass on Linux, and the tree-wide profile
check passes.
</details>
---
## Starter profiles
Three vendors ship belt profiles. Belt mode needs the printer settings
at **Advanced** or above to show its group.
<img alt="advanced mode"
src="https://github.com/user-attachments/assets/7a519ce5-b3b5-400c-a914-4f208bb577b0"
/>
| Printer | Vendor bundle | Nozzles | Processes | Filaments |
|---|---|---|---|---|
| **Generic Belt Printer** (`MyBeltPrinter`) | Custom | 0.2, 0.4, 0.6,
0.8 | 0.20mm Standard, 0.12mm Fine | library |
| **BabyBelt Pro** (Printcepts) | Printcepts | 0.4 | 0.20mm Standard |
Generic PLA, Generic PETG, eSUN PLA |
| **IdeaFormer IR3 V2** | IdeaFormer | 0.4 | 0.20mm Standard | Generic
PLA, Generic PETG, eSUN PLA |
To set up a printer that isn't listed: pick **Generic Belt Printer**,
set the belt width and length, save the profile, then copy in your
machine's start/end G-code and limits and tune from there.
<img alt="generic belt printer"
src="https://github.com/user-attachments/assets/90134d55-d6fe-4dba-8695-5ea44e78ec2b"
/>
<details>
<summary>BabyBelt Pro</summary>
<img width="2467" height="1392" alt="BabyBelt Pro"
src="https://github.com/user-attachments/assets/2b448cee-339e-43c9-964b-1ee9044044c7"
/>
</details>
---
## Contributions
### @HarrierPigeon
***Majority of design & implementation***
I did most of the work here by myself with AI tools (primarily Claude,
some Codex.)
PRs #12733, #12998, #14385, #15155, #15361, #15156, #15526, #16127
### The @Unlayered3D Team
***Rotation-Mode Pipeline***
The initial version of this sheared the model in the pre-slice pipeline.
Talking with them convinced me to switch to the current
rotate->slice->unrotate-> remap & shear method, which had significant
immediate improvements. Working with them has been a blast.
### @tommasobbianchi
***IdeaFormer IR3V2 Profile, eSUN PLA Tuning, G-Code Render / Preview***
TommyB came in at the perfect time to help keep me motivated and
contributed several things I hadn't had the werewithal to implemement
yet. Without their contributions and encouragement, we wouldn't be here
yet.
### The BabyBelt Community
**BabyBelt Pro** — @rexit1982 for the profile, and @RobMink of
Printcepts for the printer and a lot of patient testing.
**Field reports** — Many members of the BabyBelt community helped,
testing on their machines, providing G-Code and examples of issues, and
encouraging me to keep working on it. Among them:
- @RobMink - creator of the BabyBelt
- @rexit1982 - initial BabyBelt Pro profile, bug hunter
- @shubhracc - found a *lot* of technical bugs
- @NeoDLC - bug hunter
and BabyBelt Discord members who found bugs & gave feedback in no
particular order:
- @horatio42 - also provided a build machine while mine was down
- @HotCubCar - requested belt printer brim support
- Swap_File
- @Nyctelios
- Sup
- @matschi140
- @Rise-Run
- @shooby-dooby
### OrcaSlicer Maintainers
**Generic belt printer** — @SoftFever
**Keeping on top of upstream** @RF47 & @HanifKoh
**Review and fixes** — @HanifKoh (#15685 and the review on this PR).
Among them: plates after the first printed off the bed on belts; painted
supports and seams ignored the belt transform; support generation failed
at a 90° tilt; the object table crashed on the Support column on every
printer because tilt keys were in the per-object tables; apron brims
printed in the wrong filament; `build_plate_tilt` went stale outside the
GUI; the CLI reserved a wipe tower on belts; every G-code file was
treated as belt G-code because the config block carries the angle; tests
didn't compile on Clang/MSVC; plus a long list of smaller clean-ups and
the review questions that led to the clearance check, the header length
fix, the retired floor modes and the removal of the diagnostic logging.
### Additional Thanks
A special thanks to LDO Motors, who provided equipment for validating
multicolor, and my wife, who not only put up with with this obsession,
and the addition of three belt printers to our home, but has encouraged
me to keep going ever since I started this project six months ago.
---
## Known Issues
- The purge prism's first tool is chosen by the shared `ToolOrdering`
logic; on some layouts the print opens on the wrong filament and makes
one extra change at the thin tip of the prism (the "cannot absorb the
full purge volume" warning at a low height).
- Colour grouping in arrange is a soft cost: when the belt is too short
for clean runs, colours overlap rather than spill onto another plate.
- Three OrcaFilamentLibrary filaments still carry `filament_z_hop` 0.4;
belt profiles override it to 0.
Also: slicing at an angle other than the machine's (decoupled frame
tilt) is supported but not something the starter profiles exercise.
## Summary
Fixes#15944.
The plugin audit deny-list matched `secret`, `cert`, and
`conf` as substrings of every path component. This blocked
valid imports during plugin capability execution, for example
`numpy/__config__.py`, because `conf` appeared inside the
module filename.
This PR changes deny keyword matching to use whole path
components instead of substring matches. It keeps the
intended protections for sensitive locations and config
files, while allowing dependency and stdlib modules whose
names merely contain those strings.
## Changes
- Match denied path keywords as whole components instead of
substrings.
- Keep denying sensitive directory names such as:
- `secret`
- `secrets`
- `cert`
- `certs`
- `certificate`
- `certificates`
- `conf`
- `config`
- Keep denying config files by extension:
- `.conf`
- `.ini`
- Allow legitimate Python module/package paths such as:
- `numpy/__config__.py`
- `numpy/_core/_ufunc_config.py`
- `configparser.py`
- `sysconfig.py`
- `logging/config.py`
- `certifi/cacert.pem`
- Include the denied target and reason in `PermissionError`
messages when the audit hook blocks an operation.
- Remove an unused `<memory>` include from
`PluginAuditManager.hpp`.
## Why
The previous substring matching caused false positives for
common dependency and standard-library paths. It also made
failures hard to diagnose because the Python exception did
not include the refused path.
The new behavior is narrower: it blocks sensitive path
components and config file extensions without treating
unrelated names like `__config__.py`, `configparser.py`,
`Conference`, or `Concert` as secrets.
## Testing
- Added/updated unit coverage in
`tests/slic3rutils/test_plugin_audit.cpp` for:
- whole-component keyword matches
- `.conf` / `.ini` blocking
- case-insensitive matching
- false-positive paths from #15944
Plugin used for testing:
[orca_audit_numpy_config_repro.py](https://github.com/user-attachments/files/33143848/orca_audit_numpy_config_repro.py)
The round-trip test named a TextureTileMethod enumerator that does not
exist (MirroredRepeat, not Mirror) and used a bare Approx, which Catch2 v3
keeps in its own namespace. The rest of the file compares floats with
WithinAbs, so these follow it rather than reaching for Catch::Approx.
bbs_3mf.cpp reads and writes the layer JSON but picked up the declarations
through another header, and the same for boost::to_lower.
Make the connected-net layout affordable on a dense patch
Laying a patch out as a connected net cost ~175 ms on a 42k-triangle patch,
against ~21 ms for the unwrap it works from, and the gizmo asks for it on every
preview, overlay and bake. Measured on a real project the grid behind it ran
~19 million triangle-pair tests per net, nearly all of them misses: a cell
holds every triangle whose box touches it, and a candidate really meets a
couple of them.
Keep a bounding box with each stored triangle and answer those misses with four
comparisons instead of a full intersection. The net drops to ~53 ms with
identical output - the seam metrics on the test project did not move by one.
Two further attempts were measured and dropped, and are recorded in the comment
so they are not tried again: a free-space pre-check per chart came out slower,
because a folded chart lands against the net by construction and the cells
under it are occupied anyway, and splitting the boxes into their own array for
locality lost more to growing two vectors per bucket than it gained.
Also pick a pair's fold line from the longest boundary they share rather than
whichever edge came first, and grow the net strongest-adjacency-first rather
than breadth-first by area. Only the fold a chart is reached by comes out
matching, so a chart claimed across a short boundary leaves the long one it
shared with its true neighbour torn.
texture_unwrap_dump reports an unwrap from a saved project - charts, their
topology, folded triangles, where the texture is discontinuous and how long
those seams are. All of the above was found with it, and it is what keeps a
claim about this code honest; reading the 3D view and guessing had produced
three wrong diagnoses in a row.
Its toggle handler did not call Skip(), so CheckBox's own handler, which
redraws the tick, never ran. Each click still flipped and saved the
setting, but the box kept showing it ticked.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Multimaterial page showed an IDEX/IQEX section in Simple mode on
every FFF printer. The pre-slice warnings row's hand-built option
definition left its mode at comSimple; it now uses comAdvanced, like the
IMEX options around it. The parallel modes grid had a group of its own
holding only a full-width widget line, which records no mode, so that
group showed in every mode. The grid now sits in the configuration
group, which follows its Advanced rows. That group shows or hides
everything in it, so the grid keeps itself hidden on non-IMEX printers
(IMEXModesCtrl::set_applicable).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Explain the texture displacement views that cannot be used, and show them
Checker and Distortion are views of a layer's unwrap, so they mean nothing on
any other mapping - but they were offered on all of them and simply drew
nothing when picked. Fade them out on anything but Unwrap (LSCM), with the
reason in the tooltip, and have a click bring the UV editor up on the view it
selected, since that pane is where the unwrap is actually worked on.
Neither view was visible even when it did apply. The overlay is built from the
base patch and drawn with a polygon offset, which biases depth values rather
than moving geometry, so it cannot win against the displaced preview standing
in front of it. Leave that preview out while a UV-check view is on and draw the
undisplaced surface instead, which is what the offset assumes and what the
mapping being inspected belongs to.
In the UV editor, a tool that cannot be used right now is faded rather than
disabled. A disabled window gets no mouse events on GTK or MSW, so every one of
those tools - Cut, Join, Unjoin, Clear seams, Clear UV edits, the select modes,
Snap, Frame - silently had no tooltip in the state where the user most needs to
know what is missing. Each now says what to do instead. The tile size, the
island statistics, the status line and the three select modes gained tooltips
of their own; the select modes now name the gestures they enable, which were
documented nowhere.
* UV editor: keep the mouse capture balanced
The canvas captured the mouse on every button press without checking whether
it already held one, released it in a single place, and handled no capture
loss at all. Two sequences leaked a capture: pressing a second button during a
drag nested a second one that the single release on button-up could not undo,
and a modal R/S skipped that release entirely while waiting for a confirming
click that may never come.
A leaked capture is not a local problem on macOS, where wxEVT_MOUSE_CAPTURE_LOST
is never sent and nothing recovers it. While any wx window holds a capture,
wxOSX routes every mouse event to that window and never calls through to
NSWindow, so the application stops seeing motion and enter/leave, and native
tooltips stop appearing anywhere in it.
Capture through grab_mouse()/drop_mouse() so there is at most one, give it back
on any button-up including a modal gesture (which tracks the pointer and needs
no capture), and cancel on wxEVT_MOUSE_CAPTURE_LOST: commit nothing, put back
what a modal rotate or scale already applied, and do not release a capture that
is already gone.
---------
Co-authored-by: ExPikaPaka <mrfsfyt@gmail.com>
* Fix a crash on loading a 3MF with empty project settings
opt_float() dereferences what option<>() returns without checking it, and
option<>() is called with create = false. Three CLI sites read printable_height
that way, so a 3mf whose Metadata/project_settings.config holds an empty object
takes the CLI down with a null dereference. Both models shipped in
resources/handy_models are such files, so `--info` on either of them segfaults.
Guard the three reads the way the neighbouring reads of
extruder_clearance_height_to_rod and friends already are. All three target
variables are initialised to 0 and the consumer tests for > 0, so an absent
setting already had a defined meaning and nothing changes for a project that
carries the setting.
* Add a CLI regression test for a project with empty settings
Runs --info over a copy of a shipped model whose Metadata/project_settings.config
has been rewritten to an empty object, so the test keeps covering the crash no
matter what settings the shipped models carry later.
Verified both ways: the test passes against this branch and fails with a
segmentation fault against a build without the guards.
---------
Co-authored-by: Rodrigo Faselli <162915171+RF47@users.noreply.github.com>
libpng reports a bad file by longjmp()ing back to the buffer set with
setjmp(), and the frame it lands in must own nothing that needs
destroying: with exceptions enabled MSVC unwinds the stack as part of
longjmp, and returning from a frame unwound that way crashes. It did on
Windows while working everywhere else.
The read callback also returned quietly on a short read, leaving libpng
to decode whatever happened to be in the output buffer.
The calls that can fail now sit in two helpers that own nothing but
pointers, so every C++ object the decoders need stays in their own
frames, and a short read is reported through png_error().
Every BambuStudio project opened with a "BambuStudio Project" info dialog
(or, from BambuStudio 2.8.2, one saying the file is newer than the
compatible version and to update the software), followed by the
configuration-substitution dialogs for the project and its embedded
presets. None of them asks anything and all of them fire for every
BambuStudio file.
For BambuStudio projects (untagged files newer than 2.3.2, the existing
test) log the version with the unrecognized settings, and each replaced
value, instead. The geometry-only, invalid-values and G-code safety
dialogs stay, and other 3MFs are unchanged.
The Slice-plate hover popup (FilamentGroupPopup, a wxPopupTransientWindow)
takes the mouse capture while it is shown, and on macOS its OnIdle handler
reacquires that capture whenever the cursor sits outside the popup. If the
popup is still shown when the modal filament grouping dialog opens, wx routes
every dialog mouse event to the now-hidden popup, because WX_filterSendEvent
short-circuits to the capture window while GetCapture() is non-null. The
dialog's filament blocks never receive a mouse-down, so they can't be dragged
and the whole app looks frozen even though its modal loop is healthy and the
keyboard still works.
Dismiss the popup synchronously before the dialog opens: Dismiss() hides it
and releases the capture, and hiding it stops OnIdle from reacquiring. This is
a no-op where the popup is never shown (Linux, where the hover popup is
disabled, and any non-dual-nozzle printer).
MoonrakerPrinterAgent filled physical_extruder_map from AFC's per-lane
extruder_index whenever the printer's filaments were synced. IMEX does
not need it: printers carry their map in the printer profile, and IMEX
reads it from there. The sync also wrote device state into the edited
printer preset, so a sync could mark the preset modified or replace a
hand-tuned map. The two Moonraker agent files go back to upstream's
version. The sync can return separately, writing the map somewhere other
than the printer preset.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
## Description
Addresses the three items in @raistlin7447's review of 2026-10-08 on
#14394
(https://github.com/OrcaSlicer/OrcaSlicer/pull/14394#pullrequestreview-5459955028),
one commit each.
**Leading-edge brim with a leading overhang (BeltBrim.cpp).** The
leading-edge cut was taken at the first layer with geometry. With an
overhang on the leading side that layer is the overhang's tip, which is
sliced before the part reaches the belt and does not touch it, so the
cut lay ahead of the part. Reproduced on the BabyBelt profile with a 20
mm cube and a fin over its leading end, leading brim length 10 mm, width
5 mm:
| Part | Leading-edge brim before | After |
|---|---|---|
| Plain cube | 53 brim lines | 53 (unchanged) |
| Cube + 30 mm fin | 9, a sliver well ahead of the part | 53 |
| Cube + 40 mm fin | none | 53 |
| Cube + 30 mm fin, leading length 0 | none | 18 |
As suggested, the cut now uses the first layer whose contact band in the
footprint loop is non-empty: the loop records it while it builds the
footprint.
**Dead empty-layer drop (GCode.cpp).** The by-layer
`collect_layers_to_print()` built its groups only from the per-object
entries, and the per-object overload already drops every belt entry that
prints nothing, so the group-level drop could never remove anything.
Removed; its explanation moved to the drop that does the work. No output
change.
**First-layer point test comment (GCode.cpp).** Reworded to say what the
lambda undoes (what `point_to_gcode()` added and the writer took off),
since on a belt `m_origin` is rotated by `on_set_origin()` and is not
"the instance part". Comment only.
## Tests
- New test "Leading-edge-only brim ignores an overhang ahead of the
part" (30 and 40 mm fins): the leading-edge brim of the cube with the
fin must match the plain cube's. Fails without the fix (plain 53 brim
layers vs 10 and 0 with the fins), passes with it (53 and 53).
- `fff_print_tests` 364 cases and `libslic3r_tests` pass.
- Before/after G-code on the current `belt-printer` head (baseline built
from it, both binaries run from the build tree with the same resources):
byte-identical for a multi-color belt project, a two-filament belt
project with the belt purge tower, two cubes printed by object, a
flat-bed organic-support project, and two plain-cube leading-edge brims.
Only the three overhang cases change, as in the table.
- `OrcaSlicer_profile_validator -s` on Printcepts, IdeaFormer, Custom
(belt) and Prusa (control): clean.
- `scripts/clang_tidy_diff.py` against `belt-printer`: clean.
Unrelated, noticed while comparing outputs: `PrintObject::m_id`
(Print.hpp) has no initializer, so on the CLI path the `; printing
object ... id:` labels can carry an arbitrary value that differs between
builds. Pre-existing; not touched here.
OS: Linux (Ubuntu), GCC, local build. Written with AI assistance (Claude
Code); every change reviewed and tested locally as listed.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Adds AnkerMake PLA+ Basic, Matte, Metallic, Silk and Glitter, converted from eufyMake Studio's bundled AnkerMake profiles (AGPL-3.0). Bumps the Anker vendor version to 02.04.00.06.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The comment said the lambda takes off "the instance part" of m_origin.
On a belt m_origin has been rotated by on_set_origin() by then, so
m_origin minus the plate offset is not the instance shift. Say what the
code does: undo what point_to_gcode() added and what the writer took
off. Comment only.
Reported by raistlin7447 in the review of #14394.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The by-layer collect_layers_to_print() dropped every merged layer group
in which nothing prints, and the per-object overload drops every belt
entry that prints nothing. The merged groups are built only from the
per-object entries, so after the second drop every group holds at least
one entry that prints and the first never removed anything. Remove it
and keep its explanation at the drop that does the work. No output
changes.
Reported by raistlin7447 in the review of #14394.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The leading-edge-only brim is the outer brim cut down to the part's first
contact with the belt. The cut was taken at the first layer with
geometry, but with an overhang on the leading side that layer is the
overhang's tip, which is sliced before the part reaches the belt and
does not touch it. The cut then lay ahead of the part: a 30 mm fin
left a sliver of brim well ahead of a 20 mm cube, and a 40 mm fin, or a
leading brim length of 0, left none at all.
The footprint loop already finds the layers that touch the belt (their
contact band is not empty); record the first of them and cut there.
The new test slices a cube with and without a 30 or 40 mm fin over its
leading end and checks that the leading-edge brim is the same. It fails
without the fix.
Reported by raistlin7447 in the review of #14394.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* build_linux.sh: add -J to build several dependencies at once
The top-level deps build is fixed at -j1, so one dependency compiles at a
time while the small ones leave most cores idle. -J N raises that level.
-j still applies in full to each dependency, so the worst case is -J times
-j compile jobs: ninja has no job server to share a pool across the nested
builds. Without -J nothing changes.
* Quote the job count for shellcheck (SC2086)
---------
Co-authored-by: ExPikaPaka <mrfsfyt@gmail.com>
A plugin enabled at startup loads before the main frame exists, so a
dock panel it opens from on_load was dropped by the one-shot CallAfter
that found no plater. Opened a moment later, before the frame was laid
out, the pane was sized against the unsized frame and track_docked_size
kept that width. Poll until the plater is shown on screen, then build
the pane; release the reserved id instead when the app is closing.
The bed-axes toggle added for the Design tab's reference planes reads
m_design_sketch_tool, which only exists under SLIC3R_CAD. Compute the
flag once and read the sketch tool inside the same guard as its other
uses.
The last headers with a using or namespace alias at namespace scope:
- TCPConsole.hpp imported boost::asio::ip::tcp into Slic3r::Utils for
two member declarations. The alias is now a private member of the
class.
- WebSocketClient.hpp declared four namespace aliases and a tcp alias
at global scope, each used only by the header. The names are spelled
out.
- Repair.hpp aliased CGAL::Polygon_mesh_processing as PMP in
Slic3r::tex2color. The three functions that use it declare the alias
themselves.
- PreciseSeam.hpp, Thumbnails.hpp and MarchingSquares.hpp used a
using-declaration or directive for one or two spots each; those spots
are qualified. Thumbnails.hpp's "PNG"sv default argument becomes
"PNG", which converts to the std::string_view parameter the same way.
- tests/sla_print/sla_test_utils.hpp had "using namespace Slic3r;" and
tests/filament_group/fg_test_serialization.hpp "using json =
nlohmann::json;" at global scope. The headers qualify their own names;
the two SLA test sources get the directive themselves.
Also removed: twelve type aliases in headers that nothing references
(ConflictObjName, CircleSqf, CircleSqd, TRawBuffer, DistanceFunction,
SamePair, ExtruderNozzleInfos, Vec2dEvent, Vec2dsEvent, Vec3dEvent,
t_option, t_optgroups, Plater::fs_path) and a duplicate
fn_ft_job_msg_destroy alias in FileTransferUtils.hpp.
46 files include the same header twice at file scope, outside any #if,
66 times in all:
Model.cpp included Model.hpp twice, Utils.hpp <algorithm> and
<string_view> twice, seven GUI headers <wx/dataview.h> and
<wx/artprov.h> twice. The second include of each is removed.
GCodeSender.cpp and GCodeSender.hpp have been commented out of
libslic3r/CMakeLists.txt since 2022 and their only two includes are
commented out as well. Both files go, with the commented lines, and
the CMake entry for SLA/SupportTreeIGL.cpp, a file that no longer
exists.
The mixed-colour metadata options are parallel per-slot arrays in the project
config. A project saved before they were sized per slot stores a single value
for the gradient ones, and one saved before they existed stores none. The GUI
sizes all seven to the filament count when it opens a project; the CLI kept
the stored arrays and exported one-element defaults for absent ones, so a
project it exported carried one-element arrays where the GUI writes one entry
per filament. Slicing is unaffected, every reader treats a missing entry as
not mixed / no gradient, but the GUI-vs-CLI comparison reported the four
gradient keys on every mixed-filament project.
The resize helper moves from PresetBundle.cpp, where it was file-local, to
PrintConfig.cpp next to set_filament_dev_options(). It creates an option the
config lacks before sizing it, a no-op for the bundle's project config where
all seven always exist. The CLI calls it with its filament count once the
project and loaded filaments are merged, after the check that every mixed
slot has a filament of its own.
## Description
Addresses every item of @raistlin7447's review of 2026-10-07 on #14394
(https://github.com/OrcaSlicer/OrcaSlicer/pull/14394#pullrequestreview-5447529307),
one commit per item, plus a follow-up commit from a second adversarial
pass over the result.
**Organic supports (the one non-belt difference raistlin's export
fixtures found).** The debug-strip commit dda58b07cd had deleted the
loop in `organic_draw_branches()` that trims every branch slice against
the collision volume, the bed and the belt plane. It is restored exactly
as on `main` (plus the belt-floor clip). New test: a cube carrying a 60
mm plate, organic supports, flat-bed printer; on every support layer no
support extrusion may come within 0.2 mm of the part's slice. To be
clear about what it proves: it guards that invariant, but on this
fixture the loop's own effect is a sub-millimetre reshaping of one
branch (checked by running the test with the loop compiled out), so the
test does not by itself fail without the loop. The loop's effect is
shown separately by slicing six organic fixtures with the stripped and
the restored binary (CLI): on a plate-over-cube fixture the stripped
build brings a branch to 0.02 mm from the part's slice at the cube's
corner where the restored build keeps 0.39 mm; the Bulbasaur project
differs in ~2000 support lines; a fixture with no wall near the branches
is byte-identical.
**G-code (belt only).**
- First-layer speed test: the writer passes points with the plate origin
already removed, so only the instance part of `m_origin` is subtracted
now.
- The mixed-filament sub-layer pass calls `on_set_origin()` like the
main instance loop.
- `m_belt_in_band` is reset per object in by-object printing, with the
cooling buffer.
- `m_layer_count` counts only the layers that are written, through the
same predicate `collect_layers_to_print()` uses
(`belt_object_layer_prints_something()`); the by-object overload drops
the empty belt layers as well, so both print sequences write the same
layer changes. The empty-layers test now runs for both sequences and
checks `; total layers count` too. Side effect worth knowing: with the
empty entries dropped per object, a multi-filament belt layer no longer
selects a filament it then prints nothing with. On belt_project.3mf (two
filaments, belt purge tower) the T commands go from 472 to 106 with the
extruded length per filament unchanged; every removed tool change was
followed by no extrusion.
**Invalidation / ordering.**
- `posSlice` now also invalidates `posDetectOverhangsForLift` (not
belt-gated: a re-slice starts the layers over with empty overhang
regions while the step stayed done; this makes an incremental re-slice
match a fresh slice).
- `btLeadingEdgeOnly` takes part in the layer-0 outer-wall-first rule
and the matching `brim_type` → `posPerimeters` rule (not belt-gated:
`Print.cpp` already prints it as an outer brim on a flat bed).
- Adding or removing an object invalidates the support step of the other
belt-brim owners, so their brims are clipped against what is on the
plate now.
**Belt brim (found during the GUI pass, pre-existing since #16236).**
"Leading edge only" produced no brim at all: the cut that narrows the
outer brim to the first contact was taken at `layers().front()`, which
since the lead-in change is an empty layer whose contact lies ahead of
the part, so the whole region was clipped away. The cut is now taken at
the first layer with geometry; `leading_edge_only` joins the
all-brim-types test and a new test checks the brim starts no later than
the part and covers fewer layers than the outer brim.
**UI.** Build plate tilt X/Y are read-only on a belt printer (they are
derived from the belt tilt). The belt temperature tower refuses a range
without an embossed model, before the project is replaced, instead of
falling back to the 230–190 model.
**Strings, dead code, comments.** Tooltip and comment say cot and
1/|sin| (what `MachineFrameTransform.cpp` does); `gcode_remap_*` labels
and tooltips are `L("literal")` so they are extracted; removed
`belt_remapped_bbox()`, `belt_min_z()`, `m_belt_global_xy_correction`,
`LayerTools::has_belt_brim`, the `belt_surface_z` constant, and (second
pass) the unused kinematics inverse (`to_logical`,
`apply_axis_remap_inverse`, `to_build_volume` and their state), the
`world_coordinates()`, `is_active()` and `belt_brim_areas_by_layer()`
accessors and two unused overloads; rewrote the comments that still
described removed code (BeltBrim.cpp SEQUENCING, GCodeWriter.hpp,
calib.cpp/hpp, GCode.hpp, BeltSliceStrategy, PrintObjectSlice.cpp,
PrintApply.cpp).
Not changed, noted for a follow-up: the outer-wall-first rule keys on
numeric layer 0, which on a belt is usually an empty lead-in layer, so
the part's first contact layer does not get the rule; and a
leading-length-only brim (zero base width) is excluded by the
`brim_width > 0` test. Both need a geometry-based rule rather than a
one-line change.
## Screenshots/Recordings/Graphs
Build plate tilt fields greyed out on a belt printer, the temperature
tower error dialog, and the brim before/after deleting a neighbouring
object are attached below (from the Xvfb GUI pass).
## Tests
- `fff_print_tests`: all cases pass (includes the new organic test and
the extended empty-layers test in both print sequences);
`libslic3r_tests` pass.
- Organic test run with the loop compiled out (temporary local switch):
passes either way on this fixture, see above; the CLI comparison on six
fixtures is where the loop's effect is visible.
- `OrcaSlicer_profile_validator -s` on the belt vendors and Prusa as
control; `scripts/orca_profile_tool.py check`; profile tool unit tests
(281).
- `scripts/clang_tidy_diff.py` against `belt-printer`: clean.
- GUI pass on Xvfb (Linux): tilt fields greyed/editable with belt
on/off; temperature tower error for 250–200 leaves the project
untouched, 230–190 loads the tower; multi-colour demo by layer 595
slider layers = 595 layer changes with matching labels and no greying
while dragging; two cubes by object 314 = 314; outer brim complete after
deleting the neighbouring cube; organic supports clear of the part on
the belt preset and on a flat-bed variant; raw G-code toggle via menu
and `B` keeps the slider index; no crash or assert in the logs. The
leading-edge brim finding from this pass is fixed above.
OS: Linux (Ubuntu), GCC, local build. Written with AI assistance (Claude
Code), every change reviewed and tested locally as listed.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
The leading-edge-only brim is the outer brim cut down to the part's
first contact with the belt, and the cut was taken at layers().front().
Since the slicing frame starts at the belt below the footprint (#16236)
that is an empty lead-in layer whose contact lies ahead of the part, so
the cut removed the whole region and the brim type produced no brim at
all. Take the cut at the first layer with geometry.
The all-brim-types test now includes leading_edge_only, and a new test
checks that the brim starts no later than the part and covers fewer
layers than the outer brim.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Guard the layer count and the per-object layer collection against an
object that is left without a layer to print on a belt (the counting
loop stepped before begin() and front() was taken of an empty vector).
Check the belt temperature tower's embossed model before the current
project is replaced, not after. Only invalidate the support step of
objects that own a belt brim when an object is added or removed. The
empty-layers test now counts an extrusion only where material is laid
down along a move. The BeltBrim.cpp SEQUENCING note says exactly which
layers are read, and the machine-frame scale is 1/|sin|.
Remove more code that nothing calls: the kinematics inverse
(to_logical, apply_axis_remap_inverse, to_build_volume and the state
kept for them), the world_coordinates(), is_active() and
belt_brim_areas_by_layer() accessors, the PrintConfig overload of
physical_tilt() and the DynamicPrintConfig overload of
compute_belt_height_and_floor(). Comments in GCode.hpp,
BeltSliceStrategy.hpp/.cpp and PrintObjectSlice.cpp that described the
retired pre-slice remap and plane-evaluator still did; the purge-tower
width tooltip named the wrong switch.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
BeltBrim.cpp still described the brim as running inside the parallel
support step; it runs sequentially after it (generate_belt_brim). The
GCodeWriter, calib.cpp and calib.hpp comments referred to an inheritance
layout and a dynamic_cast that no longer exist.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
belt_remapped_bbox() had no callers; belt_min_z() and
m_belt_global_xy_correction were written but never read;
LayerTools::has_belt_brim was set but never read; belt_surface_z was a
named zero.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The machine-frame transform is a shear of cot(tilt) and a scale of
1/sin(tilt), not tan and 1/cos; fix the tooltip and the matching comment
in BeltGCode.cpp. The gcode_remap_* labels and tooltips were passed
through L() as variables inside a lambda, which the string extraction
does not see; pass L("literal") at the call sites.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The calibration fell back to the 230-190 tower when no embossed model
existed for the requested range, so the printed numbers did not match
the temperatures. Show an error naming the range and stop instead.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
update_fff() derives build_plate_tilt_x/y from the belt tilt on a belt
printer, so a value typed into those fields was silently overwritten.
Disable the two fields while belt_printer is on.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
A belt brim is clipped against the other objects on the plate and is
built with its object's support step. When an object was added or
deleted only the print-level skirt/brim and export steps were
invalidated, so the remaining objects kept brims clipped against objects
that were no longer there, or overlapping ones that had arrived.
Invalidate posSupportMaterial on every object in that case on a belt
printer. Also reword the comments that still described the global Z
offset as a minimum across all objects.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The first-layer rule that prints the outer wall first when a brim is
attached to it, and the brim_type change rule that regenerates the
perimeters for it, only knew btOuterOnly. btLeadingEdgeOnly, the belt
brim at the part's first contact, is an outer brim too.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
posSlice's invalidation list did not include posDetectOverhangsForLift.
A re-slice starts the layers over with empty overhang regions while the
step stayed done, so GCode::needs_retraction() had no overhangs to test
against until something else invalidated it.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
collect_layers_to_print() drops the belt layers that print nothing (an
object's empty lead-in), but m_layer_count still counted every object
and support layer, so "total layers count", the total_layer_count
placeholder and the M73 progress disagreed with the layer changes in the
file. Count with the same predicate, shared through
belt_object_layer_prints_something(). The by-object overload of
collect_layers_to_print() now drops those layers as well, so both print
sequences write the same layer changes.
Dropping the empty entries per object has one more effect on multi-
filament belt prints: a layer no longer selects a filament that it then
prints nothing with. On belt_project.3mf (two filaments, belt purge
tower) the T commands go from 472 to 106 while the extruded length per
filament is unchanged; each of the removed tool changes was followed by
no extrusion.
The empty-layers test now runs for both print sequences and also checks
"total layers count" against the layer changes.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
In by-object printing the cooling buffer is reset for every object, but
m_belt_in_band, which tracks whether the extrusion is inside the band
along the belt where the part fan stays off, kept the previous object's
value. If the previous object ended inside the band the next one never
emitted its band start marker. Reset it with the cooling buffer.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
process_layer()'s sub-layer pass (several filaments in one layer without
a purge tower) calls set_origin() per instance like the main instance
loop, but not on_set_origin(), which on a belt printer runs the origin
through the belt transform. Add the call so both passes place the
instance the same way.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The writer hands set_first_layer_point_test() a point with the plate
origin (its own XY offset) already removed, but the test subtracted the
whole of m_origin, which carries the plate origin as well as the
instance shift. On a plate other than the first the point was moved by
the plate origin a second time and the band test looked at the wrong
spot. Subtract only the part of m_origin that is not the writer's
offset.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
dda58b07cd stripped debug instrumentation from TreeSupport3D.cpp with a
script, and that script also deleted the loop in organic_draw_branches()
that trims every branch slice against the collision volume, the bed and,
on a belt, the belt plane. This is the generator every printer uses, not
a belt code path, and it is the one place where raistlin7447's export
fixtures differed from main with belt printing off. Restore the loop as
it was on main, with the belt-floor clip.
The new test prints a cube carrying a 60 mm plate with organic supports
on a flat-bed printer and checks on every support layer that no support
extrusion comes within 0.2 mm of the part's slice. It guards that
invariant; on this fixture the loop's own effect is a sub-millimetre
reshaping of one branch (verified by slicing the fixture with and
without the loop), below the asserted gap, so the test does not by
itself fail without the loop.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Classic tree supports (tree hybrid / slim / strong) on a belt slid down
the belt plane ahead of the part instead of landing on it.
`TreeSupportData` added the belt surface to every layer's outlines, so
the belt fed the collision and avoidance maps, and a node that descends
onto an obstacle is pushed out of it; on a tilted surface that walks the
branch down the belt. This takes the belt out of the outlines. The belt
is where a branch ends, and that is already handled: `drop_nodes()`
stops a node once its whole circle is in the belt
(`belt_node_landed()`), and `draw_circles()` clips every layer's circles
to the belt plane, so the branch tapers to a tip on it. Organic got the
same treatment in #16236 (the belt is no longer a support blocker
there).
One file, +7/−12. Non-belt printers are untouched: the removed block
only ran when the belt floor context was active.
## Before / after
Cube with a fin whose underside is parallel to the layers, 20 mm ahead
of the cube, tree hybrid, Left view:
| | support footprint along the belt | filament for support |
|---|---|---|
| before | belt Z 43–139 (sweeps 72 mm ahead of the part) | 2403 mm |
| after | belt Z 60–139, columns parallel to the up direction | 1606 mm
|
Organic on the same model: belt Z 74–139 (unchanged). Before/after
screenshots follow in a comment.
## Tests
- *Belt supports reach the belt under a leading overhang* passes for
normal, organic and tree_hybrid; all `[belt]` tests pass;
`fff_print_tests` 355 and `libslic3r_tests` 1116 pass on the branch.
- `scripts/clang_tidy_diff.py --base upstream/belt-printer`: no
findings.
- Fork CI (Build all) on this change: unit tests green on Linux x86_64,
Linux aarch64 and macOS arm64
(https://github.com/HarrierPigeon/OrcaSlicer/actions/runs/37601644023;
its Windows and slice-check failures are the ones #16262 fixes).
- Scripted GUI pass on belt-printer + this change: tree hybrid, organic
and normal supports at Y≈120 all reach the belt (lowest 0.17–0.19 mm);
with the part within its height of Y = 0 all three generators now behave
the same (support before the belt start, plate-boundary error shown),
where tree hybrid used to be the odd one out (clipped, hanging 9.5 mm
above the belt).
- Written with Claude Code; reviewed and run by me.
TreeSupportData added the belt surface to every layer's outlines, so the
belt fed the classic tree's collision and avoidance maps. A node that
descends onto an obstacle is pushed out of it, and on a belt that walked
the branch down the tilted surface, ahead of the part, before it could
end: tree hybrid/slim/strong supports swept far along the belt where
organic supports dropped straight down. Take the belt out of the
outlines. The belt is where a branch ends, and that is already handled:
drop_nodes() stops a node once its whole circle is in the belt
(belt_node_landed()) and draw_circles() clips every layer's circles to
the belt plane, so the branch tapers to a tip on it.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Two one-file fixes that get `belt-printer`'s CI green again after
#16236; both failures are mine.
## Changes
1. **Tests: qualify `Polyline` in the belt overhang test for Windows.**
Both Windows builds fail at `tests/fff_print/test_print.cpp:1398`
("reference to 'Polyline' is ambiguous"): the GDI function of the same
name, like the `Polygon` fix in #16196. `Slic3r::Polyline`.
2. **Profile validator: slice belt printers with two cubes along the
belt.** The slice check (`-s`) prints one cube per printer with a height
range 4–10 on filament 2 and, on belt printers, expects a plain `T1`.
Since #16236 a belt object's slicing Z starts at the belt below its
leading end, well below the part's first printed layer, so that range
falls into the empty lead-in and filament 2 is never used; all six belt
printers reported "the filament change never fired". Belt printers are
now sliced with two cubes one behind the other along the belt, the
second on filament 2. Other printers are unchanged.
## Tests
- Root cause for both confirmed in the upstream logs (run 37583336264
and the push run on 0b11311d40) and reproduced locally with the rebuilt
validator.
- `OrcaSlicer_profile_validator -s -l 2`: Printcepts 8/8, IdeaFormer
8/8, Custom 20/20 (the four MyBeltPrinter nozzles included), Prusa 95/95
as a non-belt control.
- `fff_print_tests` and `libslic3r_tests` pass;
`scripts/clang_tidy_diff.py --base upstream/belt-printer`: no findings.
- A fork run of Build all with these two commits on top of belt-printer
(plus a pending belt change) was green on every job: Windows x64 and
arm64 builds, Slice check, unit tests on Linux x86_64, Linux aarch64,
macOS arm64, Windows x64, Windows arm64 and both Flatpaks:
https://github.com/HarrierPigeon/OrcaSlicer/actions/runs/37607869527
- Written with Claude Code; reviewed and run by me.
The slice check (-s) prints one 10 mm cube per printer with a height
range on filament 2 and expects the filament change to fire. Since
#16236 a belt object's slicing Z starts at the belt below its leading
end, well below the part's first printed layer, so the range 4-10 falls
into the empty lead-in and filament 2 is never used: every belt printer
reported "the filament change never fired" and the Slice check job on
belt-printer went red. A height range in slicing Z does not map onto a
part on a belt in any case. Slice belt printers with two cubes one
behind the other along the belt, the second on filament 2, which gives
the one plain T1 the check looks for. Other printers are unchanged.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Windows headers declare a global Polyline, so the unqualified name in
test_print.cpp is ambiguous there (both Windows builds of belt-printer
fail at tests/fff_print/test_print.cpp:1336), as Polygon was in #16196.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Fixes the preview layer bar on belt prints with several parts along the
belt (reported with a cube on filament 1 and a 3DBenchy on filament 2,
no purge tower): the top slider layer held nearly the whole print, the
slider jumped every other layer through the single-colour stretch before
the second part, and with the belt purge tower the whole print greyed
out while dragging.
## Cause
Since #16236 the slicing frame of a belt object starts at the belt below
its leading end, so its first layers are empty. On a single part they
carry the brim bands. With several parts along the belt the later parts'
empty layers fall between the earlier parts' printing layers and were
written to the G-code as layer changes with no moves at all. The preview
numbers its layers (`libvgcode::Layers`) from the vertices it is given
and expects consecutive ids, so at the first such gap it stopped
creating layers and folded everything after it into the last one.
## Fix
- `GCode::collect_layers_to_print` drops the belt layers that print
nothing (no object, support or brim content): no layer change without
moves in the file.
- `libvgcode::convert` renumbers the layers consecutively over the moves
that exist, so a file with empty layers from any source still previews
correctly.
- The layer slider labels each belt layer with its print Z (the slicer's
layer Z, which increases along the belt) instead of libvgcode's toolpath
height, which on a tilted layer is wherever its last extrusion happened
to end; the slider assumes the list increases, so the labels showed "0 /
max" on alternate layers. The processor reads that print Z from the
`;Z:` tag non-BBL printers write (it only knew `; Z_HEIGHT:`), on belt
printers only, so nothing changes for other printers.
## Verification
- New regression test *Belt G-code has no layer that prints nothing*
(two cubes 60 mm apart along the belt): fails on the previous code with
one empty layer, passes now.
- `fff_print_tests` 356 passed, `libslic3r_tests` 1116 passed;
`scripts/clang_tidy_diff.py --base upstream/belt-printer`: no findings.
- The reported project sliced through the CLI: 595 layers, none without
an extrusion, Z strictly increasing.
- Scripted GUI pass on the reported project with and without the purge
tower: the slider has one entry per G-code layer, each step shows a thin
tilted strip advancing along the belt, the top layer alone is a thin
strip, nothing greys out while dragging, the slider opens at the top
after slicing, and every label reads the layer number and the print Z
matching the G-code's `;Z:` (checked at the top, mid-print and through
the two-part stretch); raw-view toggle and slider retention unchanged.
Left as is: the lower handle at the bottom still reads `1 / 0.00` rather
than the first layer's Z (index correct); pre-existing.
Remove three update_non_diff_values_to_base_config scenarios that
tested the truncation guard from upstream #13316. IMEX does not change
that code, and upstream's own tests cover it: removing the guard fails
#13316's test, and loosening it to `>=` fails #16107's.
The equal-size scenario broke when #16107 changed how variants are
matched, and the scalar-key scenario could not fail at all.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
## Speeds up OrcaSlicer incremental rebuild on Linux
Profiled `build_linux_image.sh`: 96 s, of which 51 s in the dependency
audit.
**`appimage_is_elf_file()`** ran `file` and `grep` per candidate. An
AppDir holds ~9.6k of them, 4.8k being the bundled Python runtime and
none of them ELF: ~19k processes, 14 s. Reads the four-byte magic
instead. Checked against the old result on 4000 files, no disagreement.
**The dependency walk** popped its queue with `"${queue[@]:1}"`, which
rebuilds the whole array each time. At ~4.8k entries that was 22 s of
copying an array around. Uses a read index.
Audit still passes. `shellcheck` v0.11.0, the version CI uses, is clean.
## Notes
The 96 s -> 12.7 s. Measured on a 32-core / 48 GB machine, but the audit
is a serial bash loop, so cores and RAM is not the bottleneck. On slower
hardware the saving should be larger
## Images
<img width="1987" height="782" alt="Screenshot_20261007_092516"
src="https://github.com/user-attachments/assets/96fcb917-38e5-49e7-8cbe-b37be1a2f23a"
/>
<img width="1807" height="742" alt="Screenshot_20261007_092621"
src="https://github.com/user-attachments/assets/4a65ffae-4d59-41f0-a1c7-ee5b49f3c4be"
/>
# CHANGES / TESTS
• Uses "Extruders" as tab name if it has multiple or it uses "Extruder" for single ones
• Keeps selected extruder while switching between "Extruders" and "Motion ability" tab
• Revert functions are working
• New extruders generated with values so they will shown in "Unchanged values" dialog if you try to change preset while its edited. revert functions not works properly without this
• BBL printers visible as Left / Right while normal printers visible as T1 / T2. i think there should be a separate option for how many toolheads and how many extruders it has. we might see 4 nozzles on same toolhead if one brand is brave enough :)
• Sidebar and other sections updates itself properly
<img width="751" height="173" alt="Screenshot-20261005173918" src="https://github.com/user-attachments/assets/3be0a6bd-84bd-4d15-822c-ed34acd98a9f" />
<img width="768" height="184" alt="Screenshot-20261005173905" src="https://github.com/user-attachments/assets/46164926-e385-4482-9fd1-41325eb9f61d" />
<img width="755" height="289" alt="Screenshot-20261005175318" src="https://github.com/user-attachments/assets/4550fa51-4c71-4a0a-b15e-9ca82dd0f1ad" />
# FIXES
• Extruders count on parameters section not updated when extruder count changed on printer settings. fixed on this PR
<img width="800" height="478" alt="Screenshot-20261005174702" src="https://github.com/user-attachments/assets/734e53df-f23f-4f83-8f87-2ecfeb6c162c" />
• New extruders gets randomly modifed parameters. fixed on this PR
<img width="871" height="87" alt="Screenshot-20261005174840" src="https://github.com/user-attachments/assets/254dae2a-fcfa-44ed-b3b0-038faf019abb" />
• Changed parameters not triggers revert / modified on extruder tabs. fixed on this PR
• Multi switch on motion ability tab not updated on extruder count change. fixed on this PR
<img width="813" height="184" alt="Screenshot-20261005174929" src="https://github.com/user-attachments/assets/f4f33f8f-b1bd-4a30-b91b-6ff632d08c2a" />
Since the slicing frame of a belt object starts at the belt below its
leading end, its first layers are empty. On a single part they carry
the brim bands; with several parts along the belt the later parts'
empty layers fall between the earlier parts' printing layers and were
written to the G-code as layer changes with no moves at all. The
preview numbers its layers (libvgcode::Layers) from the vertices it is
given and expects consecutive ids, so at the first such gap it stopped
creating layers and folded everything after it into the last one: the
top slider layer held nearly the whole print, the slider jumped every
other layer through the single-colour stretch before a second part on
another filament, and with the belt purge tower the whole print greyed
out while dragging.
Drop the belt layers that print nothing (no object, support or brim
content) in GCode::collect_layers_to_print, and renumber the layers
consecutively over the moves that exist when converting a result for
libvgcode, so a file with empty layers from any source still previews
correctly. The layer slider labels a belt layer with its print Z (the
slicer's layer Z, which increases along the belt) instead of libvgcode's
toolpath height, which on a tilted layer is wherever its last extrusion
ended; the slider assumes that list increases and showed "0 / max" on
alternate layers. The processor reads that print Z from the ";Z:" tag
non-BBL printers write (it only knew "; Z_HEIGHT:"), on belt printers
only, so nothing changes elsewhere. Regression test: two cubes 60 mm apart along the
belt produce no layer without an extrusion and the header's layer count
matches.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The device drying options hold several values per filament, as many as
the filament preset gives, and a project stores them as the filaments'
values one after another. The CLI filament merge wrote them like an
option with one value per filament, putting each preset's first value at
the filament's own index, so a project with three filaments whose preset
gives "1", "0" was exported with 1;1;1;0;1;0 where the GUI writes
1;0;1;0;1;0.
The merge now leaves these options out of the per-filament pass and
rebuilds them afterwards from every filament's values in slot order.
Without a fixed number of values per filament one slot cannot be
replaced in place, so the stored values are kept when any slot has no
config to rebuild from.
The Bambu network plug-in's code protector rewrites one page of its own signed __TEXT after loading. The hardened runtime tolerates that until the page is evicted; the next read of it then kills OrcaSlicer with CODESIGNING Invalid Page. Bambu Studio signs with allow-unsigned-executable-memory for this reason; with it added, the same build survives critical memory pressure that killed it in 30 s without.
A project's listed settings are carried onto its base preset by update_non_diff_values_to_base_config, which matched variants by exact name and id. A variant the base gained after the project was saved got the base's value, while the same value in a user preset now falls back to the preset's first variant of that extruder. So an old project opened with its printer preset already modified, and saving it wrote the base's values into the 3MF.
The function now maps variants with map_variant_indices, as update_diff_values_to_child_config does: a base variant the project does not list takes the project's first variant of the same extruder. The variant lists themselves stay the base's, so a fallback never writes one variant's name over another's.
The 02.08.02 series appended queue_plate_id to PrintParams and nothing
after it changed the ABI OrcaSlicer calls, so adding the field brings
the current layout up to 02.08.04. Make 02.08.04 the latest series and
drop 02.08.01 from the whitelist: its PrintParams no longer matches, and
its malformed bind table is refused by dyld on macOS 27, so it cannot
load there. A stored 02.08.01 falls back to the latest series through
the existing unsupported-version path.
Profiling build_linux_image.sh: 96 s, of which 51 s in the dependency audit.
appimage_is_elf_file() ran file(1) and grep per candidate. An AppDir holds ~9.6k
of them, 4.8k being the bundled Python runtime, none of them ELF: ~19k processes
for 14 s. Read the four-byte magic instead; checked against the old result on
4000 files, no disagreement.
The dependency walk popped its queue with "${queue[@]:1}", which rebuilds the
whole array each time. At ~4.8k entries that was 22 s of copying. Use a read
index.
96 s -> 12.7 s. The audit still passes.
The directive sat at global scope in a header that DeviceManager.hpp
includes, so most of the GUI compiled with all of std in the global
namespace. 42 files had come to rely on it, mostly for string, vector
and unordered_map, four of them for the ""sv and ""ms literals.
Those sites are qualified. GCodeViewer.cpp spelled the type as
std::vector<::string>, which only resolved through the directive. The
files that use the ""sv and ""ms literals get a file-scope
"using namespace std::string_view_literals;" or
"using namespace std::chrono_literals;", as other sources already do.
The *Show raw G-code (belt only)* toggle retired in #16236 returns, as
an item of the Preview canvas view menu (the eye-icon popup, after
*Labels*) with its `B` shortcut, and only there: no legend checkbox.
Unlit, the preview shows the designed (upright) view; lit, the raw
machine-frame G-code, which is what to look at when checking the machine
frame transforms. The item only appears on a belt printer in Preview.
The toggle is view only (exported G-code is byte-identical either way),
and the layer slider now keeps its layer index across the reload (the
layer Z values differ between the two views, so the old keep-by-Z lost
the position).
## Verification
- Scripted GUI pass: item present only in the belt Preview menu (absent
in Prepare and on a non-belt printer), toggles from the menu and from
`B` with the eye following the state, legend has no belt entry, exported
G-code identical with the view on and off, slider stays at its layer
through toggles, 3MF reopen and printer switch unaffected.
- `fff_print_tests` 355 passed, `libslic3r_tests` 1116 passed;
`scripts/clang_tidy_diff.py --base upstream/belt-printer`: no findings.
Wiki: OrcaSlicer/OrcaSlicer_WIKI#374 documents the menu item and
shortcut with screenshots.
Gets the `belt-printer` CI green again after #16236:
- **Check profiles**: the profile tool's unit test
`test_obsolete_keys_match_the_loader_ignore_set` compares
`OBSOLETE_KEYS` with the loader's ignore set in
`PrintConfigDef::handle_legacy()`, which gained the twelve retired belt
keys. Adds them to the tool's list.
- **clang-tidy**: `tests/fff_print/test_print.cpp` used `std::sqrt`
without `<cmath>` (misc-include-cleaner).
Verification:
- `python3 -m unittest discover -s scripts/tests -t scripts`: 281 tests
pass.
- `scripts/orca_profile_tool.py check`: no errors.
- `scripts/clang_tidy_diff.py -p build-tidy --base upstream/main` on
this head, i.e. every line the belt branch changes against `main` (100
files, the same check the *Merge Belt Printing Into Upstream* PR runs):
no findings.
The "Show raw G-code (belt only)" toggle, retired in #16236, returns as
an item of the Preview canvas view menu (with its B shortcut), and only
there: no legend checkbox. Unlit, the preview shows the designed,
upright view; lit, the raw machine-frame G-code, which is what to look
at when checking the machine frame transforms. The toggle is view
only; exported G-code is the same either way.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
# Description
The default Orca Cloud API URL omits its scheme, so libcurl interprets
it as HTTP and follows the server redirect to HTTPS. Recent libcurl
versions intentionally do not forward the `Authorization` header across
protocol/port-changing redirects, causing Orca Cloud profile sync to
receive HTTP 401 `missing_authorization` responses and eventually log
the user out.
Use the HTTPS API URL directly. Besides restoring sync with current
libcurl versions, this improves security by preventing the bearer access
token from being sent in the initial unencrypted HTTP request.
# Screenshots/Recordings/Graphs
N/A — no UI changes.
## Tests
- `git diff --check`
- Confirmed with current libcurl that the scheme-less URL redirects and
loses the authorization header, while the direct HTTPS URL retains it
A using-directive or using-declaration in the global namespace of a
header reaches every file that includes it, and a using-declaration also
makes the include checker treat that header as the one to include for
the name. google-global-names-in-headers reports both, on changed lines
like the existing check, so headers that still have one are not held to
it until the line is touched.
The check does not see a using inside a namespace.
clang_tidy_diff.py's closing message assumed every finding was a missing
include; it now says other findings need a manual fix.
- ClipperUtils.hpp imported jtMiter, jtRound and jtSquare into the
global namespace for every includer. No code names them there.
- BBLStatusBar.hpp, BBLStatusBarBind.hpp, BBLStatusBarPrint.hpp,
BBLStatusBarSend.hpp and ProgressStatusBar.hpp re-exported their class
into Slic3r::GUI. Nothing refers to the class through that namespace.
- Jobs/SendJob.hpp, Jobs/BindJob.hpp, Jobs/UpgradeNetworkJob.hpp and
AuxiliaryDataViewModel.hpp declared "namespace fs = boost::filesystem;"
at global scope without using it.
Each of these headers put a using or namespace alias at global or
namespace scope, which every includer inherited:
- BBLTopbar.hpp: "using namespace Slic3r::GUI;" at global scope, reached
through MainFrame.hpp. Seven source files used GUI names unqualified
outside the namespace because of it, one of them as "::RadioBox".
- IMSlider.hpp and TickCode.hpp: "using namespace CustomGCode;" inside
Slic3r.
- ProjectTask.hpp, Jobs/PrintJob.hpp and ConfigWizard_private.hpp:
"namespace fs = boost::filesystem;". PresetBundle.cpp and GUI_App.cpp
had no alias of their own.
- VoronoiUtils.hpp: "using VD = Slic3r::Geometry::VoronoiDiagram;" at
global scope.
The headers now spell the names out. Source files that used them get
the qualifier, or a using of their own where there are many uses.
151 using-directives, using-declarations, type aliases and namespace
aliases in source and test files that nothing refers to: the name is
never used, it duplicates a using already in scope, or the code sits
inside the namespace it names. Each one was removed on its own and the
file still compiled, both as it is and with every header-level using
taken away, so none of them was only redundant because a header leaks
the same name.
With the using gone, 28 #include lines and one forward declaration had
no other reference left in their file (boost/optional.hpp without any
optional, property_tree headers without any ptree) and go with it.
No header is touched.
* Stop Leaking json Through Headers and Drop Includes Kept Only for the Name
AppConfig.hpp, DeviceManager.hpp and UserManager.hpp carried a global
"using namespace nlohmann;", json_diff.hpp a global "using json =
nlohmann::json;" and PrinterFileSystem.h a global "using nlohmann::json;".
Every file that included one of them, directly or not, could write a
bare json, and 63 did without declaring it.
The last two also made the include checker treat json_diff.hpp and
PrinterFileSystem.h as the headers that provide json, so they were
included from files that use nothing else from them: 57 of the 59
includers of json_diff.hpp never name json_diff.
The five statements are removed. Headers that use the type now spell
nlohmann::json, source files declare their own "using json =
nlohmann::json;", and the includes that only supplied the name are
dropped or replaced by <nlohmann/json.hpp>.
Eight files reached json_diff.hpp only through an include that is now
gone and with it lost that header's "using namespace std;". The std
names they used unqualified are qualified.
* Declare json in OrcaSlicer.cpp on Every Platform
OrcaSlicer.cpp had its "using namespace nlohmann;" and the json include
inside the Linux-only include block, so on Windows and macOS it took
json from AppConfig.hpp's global directive, which is gone. The include
and a "using json = nlohmann::json;" now sit outside the block.
The loader's ignore set in PrintConfigDef::handle_legacy() gained the
belt options retired in #16236, and the profile tool's unit test checks
that its OBSOLETE_KEYS matches that set, so the Check profiles job
failed on belt-printer. Add the twelve keys.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Follow-up to #16195 and the review discussion on #14394 (yw4z's note
about the third column on the *Belt tilt* row). Removes the belt options
that are redundant or unused before the branch ships, so they never need
compatibility handling after a release, and fixes supports under a
leading overhang. Every removed key is on `handle_legacy()`'s ignore
list, so existing profiles and 3MFs load silently.
## Removed
- **`belt_slice_rotation_global`**, **`preslice_remap_global`**,
**`belt_preslice_global`** (*Global mesh transforms*) and
**`gcode_back_transform`** — the global mode and the back-transform are
what belt printing is; they are presumed on wherever the flags were
consulted (`PrintObjectSlice`, `BeltBackTransform`, `BeltGCode`,
`Print::process`, `PrintApply`, `GCodeViewer`). The *Belt tilt* row is
axis + angle only; the three `fdm_belt_common.json` drop the keys.
- **`preslice_remap_x/y/z`** — no profile used the pre-slice axis remap;
the belt tilt axis plus the G-code axis remap cover the machines that
exist, and its implementation only agreed with itself for a plain swap.
The forward transform is the rotation.
- **`belt_support_z_offset_mode`** and **`belt_support_floor_mode`** —
the first was never read by a generator; the second's only shipped value
(*Generator only*) is now the behaviour.
- **`first_layer_plane`**, **`first_layer_plane_offset`**,
**`first_layer_plane_thickness`** and `FirstLayerPlane.{cpp,hpp}` — the
first-layer band is measured from the belt surface and is one first
layer height thick.
- `belt_brim_instances_compatible()` and its validation warning:
instances along the belt get their brim.
## Supports under a leading overhang (the clipping at the object's local
Z = 0)
The slicing frame of a belt object started at its lowest vertex, but the
belt under the leading end of an overhang lies below that, so no
generator could reach it: normal supports stopped at the object's lowest
layer, and both tree generators carried extension hacks sized from the
pre-rotation bbox and capped at global Z = 0 (right only for the
trailing half of the belt). The frame now starts at the lowest
belt-floor point under the footprint, less a 10 mm margin along the belt
for the base of a support column, and the extensions are gone:
- **Normal supports** run in the object frame and get the global belt Z
offset shifted onto the result (as organic already did). With the offset
on the object layers, a top contact at negative Z turned the
intermediate-layer count negative and the generator allocated layers
until the kernel killed it — any overhang in the leading half of the
belt did this. The first-layer flange expansion is skipped on a belt
(the first support layer is the leading tip, not a flange).
- **Classic tree** nodes keep dropping until their whole circle is in
the belt, so a branch tapers to a tip on the belt instead of stopping a
radius above it.
- **Organic**: the belt is no longer a support blocker. A blocker is a
collision, and a branch descending onto one slides off it, down the belt
and ahead of the part; the belt is where branches end, which the
per-layer floor clipping already does.
Regression test *Belt supports reach the belt under a leading overhang*:
a cube with a fin whose underside is parallel to the layers, 20 mm ahead
of the cube and up to 41 mm of slicing Z above the belt, for normal,
organic and classic tree supports; the lowest support layer must sit on
the belt beneath its own lines.
The belt object height (the layer range) is now estimated from the box
of the mesh as placed on the bed. `raw_bounding_box()` has the
instance's Z offset removed, which was harmless for the old
rotated-extent estimate but not for one anchored at the belt floor (a
point's rotated z and the floor under it move in opposite directions
under a Z shift): with the first version of this change every part came
out as a wedge, sliced only up to its diagonal, in the GUI and CLI
alike. Caught by a GUI test pass; the leading-overhang test now also
checks that the whole part is sliced.
## Belt brim after the parallel support step
`belt_brim_obstacles()` reads every object's layers and support layers,
which another object's support step rebuilds (and now shifts) at the
same time. The brim is generated sequentially once the parallel step is
over (`PrintObject::generate_belt_brim()`). This is the race behind the
Windows arm64 segfault in *Belt brim of each object precedes its
perimeters on its own filament*.
## UI
- *Belt tilt* is two rows: the angle (Advanced) and the axis (Developer;
a profile-level kinematics choice). A shared line is shown by its first
option's mode, so they cannot share one.
- *Machine frame transforms* is five single-option rows (G-code remap X
/ Y / Z, Decouple machine-frame tilt, Machine-frame tilt angle — the
angle row only appears when decoupled) instead of two multi-column
lines; the remap fields got full labels since they stand alone now.
- The gravity indicator on the bed is a plain line along the up
direction (no cone, 60 % of the axes' length), per yw4z.
- The *Show raw G-code (belt only)* legend/canvas toggle and its `B`
shortcut are gone; the preview is the designed view.
Also carries the two-line `phong.fs` fix from #16226 (merges as a
no-op).
## Verification
- `libslic3r_tests` 1116 passed (92 648 assertions); `fff_print_tests`
351 passed (561 696 assertions).
- `scripts/clang_tidy_diff.py --base upstream/belt-printer`: no
findings.
- `scripts/orca_profile_tool.py check`: no profile references a removed
key.
- GUI target builds; a scripted GUI pass (xdotool) checked the settings
groups in every mode, slicing, export, instances, the purge tower,
calibration dialogs, the wizard, printer switching and 3MF round-trip.
The wiki pages (OrcaSlicer/OrcaSlicer_WIKI#374) get a follow-up dropping
the removed sections once this is in.
The belt object height is estimated from a bounding box swept through
the tilt rotation. raw_bounding_box() has the instance's Z offset
removed, which did not matter while the estimate was the box's rotated
Z extent (a Z shift moves every corner alike), but the frame now starts
at the lowest belt-floor point under the footprint, and a point's
rotated z and the floor under it move in opposite directions under a Z
shift: the offset box under-estimated the height by twice the object's
height above the bed, so the layers stopped at the part's diagonal and
every part came out as a wedge (GUI and CLI alike; the unit tests never
checked the top). Use the box of the mesh in the frame it is sliced in
(trafo_centered(), Z as placed on the bed), and have the leading
overhang test check that the whole part is sliced.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The slicing frame of a belt object started at its lowest vertex, but the
belt under the leading end of an overhang lies below that, by the
overhang's length times the tilt's shear. Every support generator works
in layers at z >= 0, so none of them could reach it: normal supports
stopped at the object's own lowest layer, and the two tree generators
each carried a stack of hacks to extend themselves below it (a post-hoc
copy of the lowest base area in TreeSupport, "virtual belt raft layers"
in TreeSupport3D/TreeModelVolumes), sized from the pre-rotation bbox
and capped at global z = 0, which is only right for the trailing half
of the belt.
Start the frame at the lowest belt-floor point under the footprint
instead, less a 10 mm margin along the belt for the base of a support
column (BeltSliceStrategy::apply_preslice_transforms and
BeltTransformPipeline::compute_belt_height_and_floor agree on it). The
layers between it and the first vertex come out empty, which belt
slicing already tolerates, and the generators need no extension at all:
- normal supports: the generator anchors its layer grid at the frame
origin, so run it in the object frame and shift the global belt Z
offset onto the result afterwards, as organic supports already did.
With the offset on the object layers a top contact at negative z
turned the intermediate-layer count negative and the generator
allocated layers until the kernel killed it (any overhang in the
leading half of the belt). Drop the first-layer flange expansion on a
belt: the first support layer is the leading tip of the support, not
a flange, and inflating it put lines in the air ahead of the belt.
- classic tree: a node now keeps dropping until its whole circle is in
the belt, so the branch tapers to a tip on the belt instead of
stopping, a radius above it, when its centre crosses.
- organic: the belt is no longer a support blocker. A blocker is a
collision, and a branch descending onto one slides off it, down the
tilted belt and ahead of the part; the belt is where branches end,
which the per-layer m_belt_floor clipping already does.
The belt brim is generated after the parallel support step instead of
inside it: belt_brim_obstacles() reads every object's layers and support
layers, which another object's support step rebuilds (and, now, shifts)
at the same time. This is the race behind the Windows arm64 segfault
in "Belt brim of each object precedes its perimeters on its own
filament".
Also: the belt tilt axis moves to Developer mode as its own row (a
shared line is shown by its first option's mode), first_layer_plane
band thickness, belt_support_floor_mode, belt_preslice_global and
gcode_back_transform are retired and presumed on, the gravity arrow is
a plain line along the up direction, and the "Show raw G-code (belt
only)" preview toggle is gone.
Regression test: "Belt supports reach the belt under a leading
overhang" slices a cube with a fin whose underside is parallel to the
layers, 20 mm ahead of the cube and up to 41 mm of slicing Z above the
belt, for normal, organic and classic tree supports, and checks that
the lowest support layer sits on the belt beneath its own lines.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Removed, with the keys added to handle_legacy()'s ignore list so saved
profiles and 3MFs keep loading:
- belt_slice_rotation_global and preslice_remap_global. Both were only
consulted when belt_preslice_global ("Global mesh transforms") was off,
which no profile does; belt_preslice_global is now the single global
mode and is presumed on everywhere the old flags were ORed in
(PrintObjectSlice, BeltBackTransform, BeltGCode, Print::process,
PrintApply). The Belt tilt row is axis + angle only.
- preslice_remap_x/y/z. No profile used the pre-slice axis remap; the belt
tilt axis plus the G-code axis remap cover the machines that exist, and
its implementation only agreed with itself for a plain swap (matrix
columns vs remap_bbox rows). BeltTransformPipeline::build_preslice_remap,
remap_bbox and has_preslice_remap are gone, the forward transform is the
rotation, and the G-code header no longer carries the remap.
- belt_support_z_offset_mode. Saved and invalidated steps, but no support
generator read it.
- first_layer_plane and first_layer_plane_offset, with FirstLayerPlane.cpp.
On every shipped configuration the band is measured from the belt
surface (GCode::belt_height_above_floor) and the evaluator was only
reached for an explicit XY/YZ/XZ choice or a non-zero offset, which
nobody set. first_layer_plane_thickness stays as the band unit,
relabelled "First layer band thickness".
UI: the Machine frame transforms group is five single-option rows (G-code
remap X / Y / Z, Decouple machine-frame tilt, Machine-frame tilt angle;
the angle row is shown only when decoupled) instead of two multi-column
lines, and the remap fields carry full labels.
Also carries the phong.fs struct fix from #16226 so the worktree build
links its shaders.
libslic3r_tests and fff_print_tests pass; clang-tidy diff check clean;
orca_profile_tool.py check clean.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Fixes the `unable to load shaders: phong` error at startup on
`belt-printer` after #16195. That PR added `vec3 up_direction` to the
`SlopeDetection` uniform struct in `phong.vs` (110 and 140) but not in
`phong.fs`, so the vertex and fragment stages declared the `slope`
uniform with different struct types and the program failed to link.
`gouraud.fs` already carried the member; `phong.fs` now does too.
Shader-only change.
#16195 added slope.up_direction to the SlopeDetection uniform struct of
phong.vs (110 and 140) but not to phong.fs, so the two stages declared the
uniform with different types and the program failed to link: "unable to
load shaders: phong" at startup, and studio lighting / realistic phong
rendering fell back. gouraud.fs already carries the member.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Gives every option in the *Belt printer* and *Machine frame transforms*
groups, the build plate tilt, the belt purge tower enable and the belt
purge tower width a wiki link (the *Wiki* button next to the option),
pointing at the pages and anchors added in
OrcaSlicer/OrcaSlicer_WIKI#374. The two purge tower links that pointed
at a whole page now point at their section. String arguments and
`label_path` assignments only; the wiki's Tab-link validator passes
against this `Tab.cpp` with that branch.
# Description
Filament changes on the Prusa CORE One MMU3, MK4 MMU3 and CORE One INDX
now use the settings PrusaSlicer 3.0 ships for them. On the MMU3
printers these are Prusa's per-material ramming, load and unload speeds,
cooling moves and stamping. On the INDX they are Prusa's multi-tool
ramming and 10 mm³ minimal purge, plus a filament start G-code that
restores pressure advance after the purge station disables it and then
sends `M573 R`, as PrusaSlicer does.
This supersedes #16008. Thanks to @nuclearmistake for that work, which
identified what these printers need for reliable filament changes. This
PR reaches the same goal with the existing `include` mechanism rather
than new printer-level overrides. The values live in shared templates in
the Prusa bundle (PLA and PETG families with High Flow ramming variants,
and one for the INDX), and the filament presets include them, so each
material keeps its own values as in PrusaSlicer. For the MK4 MMU3, ten
small presets inherit the MK4 filament tunes and include the templates;
they cover the materials PrusaSlicer offers with the MMU3 and replace
the library generics as that printer's defaults. The CORE One MMU3 0.6
nozzle now uses the 0.6 Generic PLA and Prusament rPLA profiles, as in
PrusaSlicer. Unlike #16008, filaments from Orca's shared library keep
Orca's defaults on these printers.
No engine change, and printers without an MMU3 or INDX are unaffected.
MK4 MMU3 users also get the MK4's material tuning, such as temperatures
and cooling, from the new presets.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
# Description
This PR follows up on #16019 with fixes and workflow improvements to the
Design tab. Sketching now waits for you to pick a plane, the reference
planes stay out of the way until they are needed, and the camera follows
the mouse controls set in Preferences, as suggested by @Felix14-v2 in
https://github.com/OrcaSlicer/OrcaSlicer/pull/16019#issuecomment-5932747510.
It also fixes rendering, camera and Move bugs, which are listed below.
https://github.com/user-attachments/assets/6aac0a03-8405-403a-93b2-1fbf95da9fe7
**Improvements:**
- Pan and orbit follow Preferences > Control while sketching too. A
right-drag no longer ends the polyline chain or drops the point already
placed, and in the Touchpad camera style Alt+move and Shift+move orbit
and pan while a draw tool is selected.
- The XY, XZ and YZ planes are separate labelled squares around the
origin, in their axis colours, with dash-dot axes. They no longer cross
through the bed as one grey smear. Hovering over a plane greys it out,
and selecting one makes it solid.
- The reference planes stay hidden until a sketch needs them. A new
Origin row in the Feature tree keeps them on screen, and Ctrl+Shift+O
toggles it.
- A sketch opens only once its plane is chosen. If a flat face or a
reference plane is already picked, the sketch opens on it at once.
Otherwise the planes appear, and the next plane or face you click opens
the sketch. Esc or Cancel leaves without one.
- The Bed checkbox is now a row under Origin, and its state is
remembered across sessions. Both rows sit above the feature list as
fixed view switches.
- Zoom to selection is available on Feature tree and Bodies rows and in
the right-click menu.
- A Sketch button replaces the FEATURES label, the import icons now have
an arrow, and the document icons are sized consistently.
**Fixes:**
- Feature previews no longer z-fight with the bodies. The Hole preview
now hides the bodies and shows only the result, as Fillet/Chamfer and
Draft already do.
- Body edge lines stay in sync with the bodies after undo, hide and
delete.
- The bed, grid and reference planes stay aligned on every plate, not
only the first.
- The Fit camera button frames the selection and sketches, or everything
on show, instead of always framing the whole bed.
- Grabbing a move arrow no longer makes the body jump on the first drag.
- Move no longer gets stuck after Esc or a click off the gizmo. Esc and
Cancel put the body back, and undo waits until the Move is confirmed or
cancelled.
- The camera stays still when the first body appears.
- The view buttons no longer cover the status line at display scales
above 100%.
- The import buttons no longer stay highlighted after a click.
- The Bodies right-click menu no longer closes at once on Linux.
# Screenshots/Recordings/Graphs
- **Z fighting issue**
Before fix:
https://github.com/user-attachments/assets/0da30897-8d81-494a-a206-9b4dfd8fa9c0
After fix:
https://github.com/user-attachments/assets/04ccad39-dbbc-4bcf-8d5c-3ba4894cb755
- **Sketch plane rendering**
Before:
<img width="374" alt="{CCFCEB16-20FA-4202-A5F8-5F2219E44C0B}"
src="https://github.com/user-attachments/assets/2dfd19b2-9991-45c7-91e3-e139d111229c"
/>
After:
<img width="1706" height="1258" alt="image"
src="https://github.com/user-attachments/assets/3ec89038-4614-4c04-9f20-37bfc50ee5ef"
/>
- **Wrong body frame lines**
Before:
https://github.com/user-attachments/assets/81ccae26-2a9d-4ecb-9928-678dbe9fb94c
After
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
* fix: crash in LAN mode when the printer type is not known yet
InputIpAddressDialog::set_machine_obj() built the help image name from
the printer config with no fallback. If the printer type is empty or
unknown, for example before the first push_all arrives on a flaky LAN
link, the lookup returns "" and create_scaled_bitmap("_en") throws.
The dialog is opened by the "LAN Connection Failed" handlers in
MediaPlayCtrl and MediaFilePanel, where nothing catches the exception,
so the app crashes.
Use input_access_code_x1 when there is no image, and the _cn image for
zh_CN, the same as ConnectPrinterDialog::init_bitmap().
Ported from Bambu Studio 52ca2ec5d1.
* fix: return an empty bitmap for an empty icon name
create_scaled_bitmap() threw when a caller passed an empty name. That
happens when a printer config lookup has no entry, for example in
AMSSetting::update_ams_img() for a printer type with no AMS image.
Log an error and return wxNullBitmap instead.
Ported from Bambu Studio 52ca2ec5d1.
They now sit above the features' framed list as fixed view switches:
a click no longer selects them, they stay put while the features scroll,
their eyes line up with the features' eyes, and their labels dim when
hidden. Ctrl+Shift+O toggles the Origin from the keyboard, as
Ctrl+Shift+B does the Bed. The test script's position for the first
feature row is calculated, not measured, and needs re-measuring on the
test setup.
The bed's show/hide state is now remembered across sessions like the Origin
row's, shown by default. The GUI ladder's ribbon x-coordinates are shifted by
the removed checkbox's derived width and still need re-measuring on the rig.
Sketch on a picked flat face or reference plane opens the sketch on it at once. With nothing
picked it no longer enters sketch mode: the reference planes and axes come up, and the plane or
flat face clicked next opens the sketch and puts them away. Esc or Cancel leaves without one.
A picked plane is used up by the sketch on it and dropped by Esc or a click on nothing, the plane
prompt is no longer replaced by a stale tool hint, and clicking the face a sketch was just
cancelled on picks that face again rather than the whole body.
* perf: write post-processed G-code without a per-line copy
* perf: size the post-process line map from the first pass
* test: line ends of the exported G-code
* test: include the headers the line-ends test and gcode() helper use
Ramming, load/unload, cooling-move and stamping values follow the
PrusaSlicer 3.0 presets, which retune several of them relative to
2.9.6. The CORE One MMU3 0.6 nozzle now uses the 0.6 Generic PLA and
Prusament rPLA profiles, as in PrusaSlicer.
* Use a System clang-tidy When Available and Make --fix Converge in One Pass
scripts/run_clang_tidy.sh only looked at CLANG_TIDY and the venv it creates,
so a clang-tidy already on the system was never used. It is now the first
choice: the pinned version outright, another version after a prompt that
says results may differ slightly from CI, which -y and an existing pinned
venv skip.
Two problems in clang_tidy_diff.py made --fix need several runs and still
leave the plain check failing:
- A deleted #include orphans uses on unchanged lines. The plain check runs
such a file whole and reports them, but --fix kept the line filter to the
changed lines, so they were never fixed. Fix mode now runs the file whole
first and then fixes exactly the changed lines plus the lines that run
found wanting, so unrelated lines are still never rewritten.
- clang-tidy exits non-zero for the findings it just fixed, so every fixed
file was reported as failed and the user ran --fix again to see what was
left. A file --fix changed is now checked again and the fixed files are
listed separately from what --fix could not add.
CI runs the script without --fix and is unchanged.
* Keep the a/ b/ Diff Prefixes Whatever the User's Git Config Says
parse_diff recognises a changed file by its +++ b/ header. With
diff.noprefix or diff.mnemonicPrefix set, git prints +++ src/x.cpp or
+++ w/src/x.cpp instead, every file was dropped, and the local check
reported no changed C++ lines. The diff is now asked for the a/ and b/
prefixes outright, which overrides both settings.
* Warn When No Remote Points at OrcaSlicer/OrcaSlicer
Without one, run_clang_tidy.sh compares against origin/main. When origin
is a fork whose main already holds the commits, the check finds nothing
and says so, without hinting at why. The script now names the base it
fell back to and how to point it at the upstream repository.
Esc and Cancel now put the body back, and Confirm keeps the new position. A click off the
gizmo only moves the camera and no longer leaves the Move card and its buttons dead. Starting
another edit or a rebuild keeps the position, and undo is refused until the Move is confirmed
or cancelled.
The XY, XZ and YZ planes no longer cross through the bed. Each is a small square in its axis
colour, set off from the axes into the corner that faces the default front view, with its name
written in the plane. Dash-dot axes run between the squares and replace the bed's axis triad
while they show. Hovering a plane greys it and selecting one makes it solid, and picking a
solid face now clears a previously picked plane.
Grabbing an arrow anywhere along its length snapped the body's centre to that point as soon
as the mouse moved. The body now moves by how far the cursor travels from where the arrow
was grabbed, including when it is grabbed while looking straight down the axis.
The XY/XZ/YZ planes are cut along each other and drawn back to front, with
lines along every crossing, and their fills are strong enough for the order
to show. Before, they blended into one grey smear and were too pale to work with.
Every option in the Belt printer and Machine frame transforms groups, the
build plate tilt, the belt purge tower enable and the belt purge tower
width get a wiki link, matching the pages added in
OrcaSlicer/OrcaSlicer_WIKI#374. The two purge tower links that pointed at a
page now point at their section.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Fixes the Windows x64 and arm64 build failures on `belt-printer` after
#16195: `tests/fff_print/test_print.cpp` includes `<Windows.h>`, so the
unqualified `Polygon` in the new TreeModelVolumes blocker test is
ambiguous with GDI's `Polygon()` (`error: reference to 'Polygon' is
ambiguous`). It is the only error in both logs. The type is now written
`Slic3r::Polygon`.
tests/fff_print/test_print.cpp includes <Windows.h>, so an unqualified Polygon
in the new TreeModelVolumes test is ambiguous with GDI's Polygon() and fails
the Windows x64 and arm64 builds on belt-printer.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
In the Design tab the button always swung the camera to the plate view and framed the whole
bed, whatever was selected: the tab's picks and sketches are neither the canvas's selection
nor its volumes. It now frames the selected faces, body, edges, vertex, sketch region or
sketch entities, and with nothing selected everything on show — the visible bodies, the
feature preview and the sketches — keeping the current view direction. An empty tab still
frames the bed as before.
Follow-up to #14394, addressing @raistlin7447's review (review
5421968464) item by item, plus the tests it asked for.
## Review items
1. **Stale belt offsets after switching printers** —
`PrintObject::slice()` now zeroes `m_belt_min_z`,
`m_belt_global_z_offset` and `m_belt_global_xy_correction` before
slicing. They were only written in belt mode, so a project switched to a
normal printer (or whose tilt axis was set to None) kept the old
offsets, which shifted the adaptive infill octree and the organic
support layers.
2. **Blocker indexing in `TreeModelVolumes`** — a test now pins the
index the support blockers land on with a raft (object layer + raft
layers), including the layers just below and just above where an
unshifted blocker would sit.
3. **Arrange clamp** — the final-alignment clamp in libnest2d is opt-in
(`NfpPConfig::clamp_to_bin`) and arrange sets it for belt printers only.
Printers with an off-centre `best_object_pos` (A1 mini, H2 family) keep
their alignment; a flat-bed test pins that and the existing clamp test
is now a belt test.
4. **Belt view from the file, not the preset** —
`GCodeProcessor::apply_config(DynamicPrintConfig)` carries the file's
belt keys (and, for a belt file, its
`printable_area`/`printable_height`, which the Rev remaps need) into
`export_config_for_render()`; `GCodeViewer` enables the belt view from
the header tilt. A normal `.gcode` opened with a belt printer selected
is no longer back-transformed, and a belt file opened on another printer
brings its own tilt and remaps.
5. **Purge-prism snap vs. support-only changes** —
`belt_shift_layer_grid()` also shifts `m_belt_floor_z_shift_cached` and
`m_belt_global_z_offset`, so the restored floor and the organic support
layers follow the snapped grid.
6. **Raft / draft shield on a belt** — `update_print_fff_config()`
resets `raft_layers` and `draft_shield` with the usual warning dialog
instead of only greying out the fields `Print::validate()` rejects.
7. **First-layer travel speed and second-layer temperature** —
`GCodeWriter` takes a first-layer point test instead of the
`FirstLayerPlane`; `GCode` installs one that goes through
`on_first_layer(point)` (the belt surface, as the extrusions use),
converting the writer's logical point back to the object frame.
`past_first_layer_band` uses a new `belt_layer_past_first_layer_band()`
on the same basis. The `FirstLayerPlane` path is kept for an explicit
XY/YZ/XZ choice or a non-zero plane offset, as before.
8. **Leading-edge brim test** — `belt_brim_clip_leading_edge()` is
exported and called by both the generator and the test (which also
checks the kept area and the cut-beyond-region cases).
9. **phong.vs** — both `110/phong.vs` and `140/phong.vs` get
`up_direction` and the `dot()` slope test, so studio lighting and
realistic phong highlight overhangs with the tilt.
## Remap gating
`preslice_remap_*` and `gcode_remap_*` are gated on `belt_printer`
through one helper, `BeltTransformPipeline::axis_remap_enabled()`. The
fields are only offered in the belt group, so a value left in a profile
must not change a non-belt print. That helper is the one place to widen
if a non-belt use ever needs them.
## Tests (as requested)
- Belt-only keys at non-default values leave non-belt G-code unchanged.
- Switching a sliced project from belt to non-belt (and to tilt axis
None) matches a fresh slice.
- A support-only change on a belt purge print matches a fresh slice.
- Non-belt start G-code moves keep the first-layer Z in the processor.
- The belt brim's segment count (not pass count) catches a band emitted
twice back to back.
## One fix outside belt code
The belt-to-non-belt test exposed a gap that `main` shares:
`PrintObject::invalidate_step(posSlice)` re-invalidates
`posSupportMaterial` but not `posSimplifySupportPath`
(`invalidate_steps()` does not propagate), so after any re-slice the
regenerated support paths were exported unsimplified — extra vertices
and tiny `E.00001` moves. `posSimplifySupportPath` is now in that list;
with it the re-sliced and fresh outputs match byte for byte (comments
aside).
## Verification
- `libslic3r_tests`: 1116 passed, 2 skipped. `fff_print_tests`: 351
passed (561 427 assertions). Built on Linux with GCC against OCCT 8.0.1
deps.
- `scripts/clang_tidy_diff.py -p build-tidy --base eb5b9a77b9`: no
findings.
- The GUI files (`ConfigManipulation.cpp`, `GCodeViewer.cpp`) compile;
the preview change was not exercised interactively.
Code review items (raistlin7447):
1. PrintObject::slice() zeroes m_belt_min_z, m_belt_global_z_offset and
m_belt_global_xy_correction before slicing. They were only written in belt
mode, so a project switched to a normal printer, or whose tilt axis was set
to None, kept the old offsets and shifted the adaptive infill octree and the
organic support layers by them.
2. TreeModelVolumes shifts the support blockers into the raft-offset index
space; a test now pins the index the blocker lands on.
3. The final-alignment clamp in libnest2d is opt-in (NfpPConfig::clamp_to_bin)
and arrange sets it for belt printers only. Printers with an off-centre
best_object_pos keep their alignment; a flat-bed test pins that.
4. The preview's belt view follows the loaded G-code, not the selected printer:
GCodeProcessor carries the file's belt keys (and, for a belt file, its bed)
into export_config_for_render(), and GCodeViewer enables the belt view from
the header tilt.
5. belt_shift_layer_grid() also shifts the cached belt floor and the global Z
offset, so a support-only or brim-only change after the purge-prism snap
matches a fresh slice.
6. update_print_fff_config() resets raft_layers and draft_shield on a belt
printer instead of only greying out the fields Print::validate() rejects.
7. GCodeWriter takes a first-layer point test instead of the FirstLayerPlane;
GCode installs one that measures from the belt surface, like its
extrusions, so the first-layer travel speed and the second-layer
temperature change no longer depend on the gcode_remap_* convention.
8. belt_brim_clip_leading_edge() is exported and called by both the generator
and the test.
9. Both phong.vs shaders use slope.up_direction for the overhang highlight.
The pre-slice and G-code axis remaps are gated on belt_printer through
BeltTransformPipeline::axis_remap_enabled(), so belt keys left in a profile
cannot change a non-belt print.
Tests requested in the review: belt-only keys at non-default values leave
non-belt G-code unchanged; switching a sliced project from belt to non-belt
(and tilt axis None) matches a fresh slice; a support-only change on a belt
purge print matches a fresh slice; non-belt start G-code moves keep the
first-layer Z in the processor; the belt brim's segment count catches a band
emitted twice.
The belt-to-non-belt test exposed an unrelated gap: invalidate_step(posSlice)
re-invalidated posSupportMaterial but not posSimplifySupportPath, so after
any re-slice the regenerated support paths were exported unsimplified.
posSimplifySupportPath is now in that list.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The line started at a fixed inset scaled like the ImGui style, while the navigator and the
round canvas buttons scale with the monitor's DPI on Windows, so at 150% the buttons covered
its first words. It now starts past the edge the canvas reports for that corner.
Follow-up to #14394. The `clang-tidy` job on that PR fails on 129
`misc-include-cleaner` findings: the belt sources and tests use `std::`,
Eigen, `Point`/`PrintConfig` and `BeltBrim` symbols without including
the header that provides them, which only compiled because the
precompiled header supplied it.
This adds every include the job names, in each file's existing include
style (`"../"` in the `GCode/` and `Support/` subdirectories, quoted
`libslic3r/` paths in the GUI and tests). 41 files, includes only, no
code changes.
Verified locally with the job's own command, `scripts/clang_tidy_diff.py
-p build-tidy --base eb5b9a77b9` (compile database configured with
`SLIC3R_PCH=OFF`, clang-tidy 22.1.8): no findings left on the 105
changed files.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_01L6Kg5igmmMU2YLoK6HrsWV
The clang-tidy job on #14394 fails on 129 misc-include-cleaner findings:
the belt sources and tests use std::, Eigen, Point/PrintConfig and
BeltBrim symbols without including the header that provides them, which
only compiled because the precompiled header supplied it. Every include
the job names is added, in each file's existing include style ("../" in
the GCode/ and Support/ subdirectories, quoted libslic3r/ paths in the
GUI and tests). No code changes.
Verified with scripts/clang_tidy_diff.py -p build-tidy --base eb5b9a77b9
(SLIC3R_PCH=OFF compile database, clang-tidy 22.1.8): no findings left.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01L6Kg5igmmMU2YLoK6HrsWV
raistlin7447 pointed this out on #14394. The preview skips reconverting
a G-code result it already shows, but belt printers were exempt from
that check, so every time you came back to the Preview tab on a belt
print, it reconverted every toolpath and uploaded it to the GPU again,
even though nothing had changed.
The exemption existed for one reason: switching between the designed and
raw views changes the toolpaths for the same result, so the B toggle
needs a fresh conversion. This change keeps that, but narrows it. The
viewer now remembers which view the result was converted for and reuses
it as long as the view hasn't changed. B, the legend checkbox and the
toolbar menu still trigger a new conversion. The print settings the
back-transform reads can't change without producing a new G-code result,
so the result id covers those the same way it does for every other
printer.
I checked it with the same scripted GUI run on both builds: a BabyBelt
Pro benchy, sliced once, then three Prepare → Preview round trips and
two presses of B, counting the viewer's own log messages.
- Before: 10 full conversions, 0 reuses.
- After: 3 full conversions (the slice and the two B presses) and 7
reuses for the tab switches. B still switches views and comes back
exactly.
The merged tree builds cleanly, and the fff_print (including all the
belt tests) and libslic3r suites pass.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_01AJzy1xeQV3FePh5HfahDyn
load_as_gcode() skips the conversion and GPU upload when it is handed the
result it already shows, but belt printers were exempt from that cache,
because the designed/raw view changes the toolpath geometry for the same
result. So every preview reload of a belt print, switching back to the
Preview tab for one, converted and uploaded every toolpath again.
Remember the view the result was converted for and reuse it while both
match. Toggling the view (B, the legend checkbox, the toolbar menu) still
converts again. The print config the back-transform reads cannot change
without a new G-code result, so the result id covers it as it does on any
other printer.
Suggested by raistlin7447 on #14394.
The slice sweep here runs the nightly validator, built from main, so it
cannot expand custom G-code that uses a setting the PR adds to the engine
and reports it as an undefined placeholder. The belt printer PR fails on
exactly that: the BabyBelt Pro start G-code passes
[belt_slice_rotation_angle] to its firmware, and main has no such setting.
A PR that changes src/ also runs Build all, whose Slice check runs the same
sweep with the validator built from the PR (it passes on that PR). So the
sweep here now runs only for PRs that leave src/ alone, which are the
profile-only PRs it exists for and the ones pr-merge-bot gates on. If the
base commit cannot be fetched, the sweep runs as before.
Merges `main` (eb5b9a77b9) into `belt-printer` so #14394 is mergeable
again. It had gone CONFLICTING after main moved 55 commits past this
morning's merge.
The only conflict is in `src/slic3r/GUI/Plater.cpp`: main translates the
pressure-advance test name (#16142) on the line right after the belt
guard that keeps PA Line and PA Pattern off belt printers. Both are
kept:
```cpp
// ORCA-Belt: PA Line / PA Pattern have the belt plumbing in place ...
{ ... belt guard unchanged ... }
const auto calib_pa_name = _L("Pressure Advance Test");
```
Main's other changes since the last merge that touch belt-modified files
were checked by hand, and none of them reach belt code:
- `GLCanvas3D.cpp`: popup flag and comments around the canvas-toolbar
menu; the belt "Show raw G-code" item is untouched.
- `GCodeViewer.cpp`: position-window scrollbar colours.
- `Tab.cpp`, `calib_dlg.cpp`, `PrintConfig.cpp`, `ArrangeJob.cpp`,
`bbs_3mf.cpp`, `GUI_App.cpp`, `GUI_ObjectList.cpp`, `Plater.hpp`: small
edits away from belt code.
A follow-up PR fixes the red **Check profiles** on #14394; merge it
right after this one.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_01AJzy1xeQV3FePh5HfahDyn
Brings belt-printer up to main eb5b9a77b9. One conflict: main translates
the pressure advance test name (#16142) on the line after belt's guard that
keeps PA Line and PA Pattern off belt printers; both are kept.
run_clang_tidy.ps1 had not been run on Windows before.
- Run native commands through Invoke-Quiet. Under
$ErrorActionPreference = "Stop", Windows PowerShell made CMake's
first stderr line fatal, so configure always failed.
- Pass the --line-filter name with native separators. clang-tidy
matches it against the end of the file's native path, so on Windows
every misc-include-cleaner finding was dropped.
- Decode subprocess output as UTF-8 and let stdout replace characters
it cannot encode. A changed line with text such as 打印 crashed the
script under cp1252.
- Check VCToolsInstallDir and WindowsSdkDir in VsDevCmd's output
before applying it, so a failure names the command to run and leaves
the calling shell untouched.
- Log the git_commit_hash_header build, use -LiteralPath for logs, and
hide VsDevCmd's stderr as build_win.bat does.
On every platform, git quotes non-ASCII paths and appends a tab to a
+++ header whose path contains a space, and parse_diff dropped both.
changed_files and the workflow's changed-files step now pass
core.quotePath=false, and parse_diff strips the tab.
* plates-toolbar-scrollbar-size
* update
* filament grouping dialog
* mixed filament list
* Update StepMeshDialog.cpp
* moves plot scrollbar
* fix position of popups
* printer agent combo box width
* match multiline text control background
* add dots to configure button
* also correct label color for multiline input
* side tools connecting text color
* recenter dialog text color
* preferences experimental features
* upgrade panel hyperlink color + right margin
* fix position of + sign on printer selector
* speed control popup
* bbl fan control window
* AMS materials setting
* monitor > ams section
* match background color of multiline text editor on project page
* Fix SwitchButton colors
* bbl camera popup
* BBL > Send print dialog
* Revert "BBL > Send print dialog"
This reverts commit 128e145897.
* Revert "bbl fan control window"
This reverts commit 4a0db62790.
* Revert "bbl camera popup"
This reverts commit e2301e3237.
* Revert "monitor > ams section"
This reverts commit 66e6894eb1.
* Revert "fix position of + sign on printer selector"
This reverts commit 1523ba24c0.
* Revert "upgrade panel hyperlink color + right margin"
This reverts commit 541f2514c5.
* Revert "side tools connecting text color"
This reverts commit a9ab074247.
* Revert "recenter dialog text color"
This reverts commit f4670d32b2.
* Revert "AMS materials setting"
This reverts commit b12827f9b5.
* Revert "Fix SwitchButton colors"
This reverts commit 2d4f1d1fe9.
* match object list background color
* match compare dialog wxDataViewCtrl background color
* fix centering of iconized buttons on linux
* fix compare dialog background color not applied on linux
* edit gcode dialog components background color
* Update Plater.cpp
* fix dev button font size
* Fix scaling issue on SwitchButton while using 150%
* transfer or discard changes dialog wiki label
* "Transfer or Discard changes" / "Unsaved Changes" dialog header color
* object table colors & header spacing
* progress dialog
* fix progressbar look on linux
* fix build
* fix build
* fix centering of iconized icon again
* fix header background color
* revert bbl fan control
* add includes
---------
Co-authored-by: Noisyfox <timemanager.rick@gmail.com>
Upstream's clang-tidy gate now checks that the lines a pull request changes
include the header for every symbol they use. The IMEX sources, their tests,
and the lines this PR adds to shared files relied on the precompiled header and
transitive includes. This adds the includes clang-tidy names; no code changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Panning or orbiting with a right-drag no longer ends the polyline chain or drops
the point already placed. Only a right-click that doesn't move does. In the
Touchpad camera style, Alt+move and Shift+move now orbit and pan even while a
draw tool is armed.
The bodies now always draw over the preview on every face a feature leaves unchanged, for
all features. Since a hole's cut sits inside the body, the Hole preview now hides the
bodies and shows the result alone, as Fillet/Chamfer and Draft do.
# Description
Follow-up to #15238. A review of the Design tab, and then testing its
Linux AppImage on desktops, turned up UX inconsistencies, kernel bugs
and a crash. This PR fixes them and fills the gaps found along the way.
Everything stays behind `SLIC3R_CAD` and the `enable_cad_feature`
preference. With the preference off, Prepare and Preview behave and
render exactly as before.
The first four commits are Design-tab fixes that landed on the fork
after #15238 and were never sent upstream: the value-field label, the
`GUI.hpp` include, sketch tool batches 9–14 and honest tool messages.
The later commits build on them.
**Interaction: one rule for mouse, Enter and Esc** (charter
`docs/CAD/cad_ux_guidelines.md` §4.2)
- Enter does what ✓ does; Esc does what ✗ does. Esc steps back one level
(value field → gesture → tool → selection) and never deletes anything.
- A click on empty space clears the selection and never applies a
pending operation.
- A right-click with nothing pending opens the offer. The click is
recognised by drift alone; the 200 ms timing rule is gone, since §6.2
forbids timing-dependent gestures.
- The panel's CHAR_HOOK owns Delete, Esc, Ctrl+Z/Y and F, so the
duplicate handlers in `GLCanvas3D` are removed.
- Ctrl+Z/Y work inside a sketch, and Edit ▸ Undo/Redo drive the Design
history while that tab is shown.
- A value the geometry cannot take (zero length, sweep > 360°, …) is
refused inside the field, and the field says why.
**Kernel**
- Solver: tangency and point-on-line pick the correct side;
circle–circle tangency is handled; the partitioned solve maps sentinel
references; constraints that could not be applied are reported instead
of silently dropped.
- Trim, extend, offset and mirror are fixed for arcs and ellipses.
Negative scale is handled, and zero-radius circles no longer produce
geometry.
- Feature → body references resolve by identity, so hiding, reordering
or deleting a feature no longer re-targets later features.
- Hole standards table corrected: inch countersinks are 82°. New threads
use the nominal diameter; new patterns use inclusive spacing. Both are
gated by flags, so existing recipes rebuild unchanged.
- **A closed loop that crosses or folds back is not a region.** Such a
loop passed the closed-loop check and MakeFace, and extruded into an
invalid solid with no caps. `SketchEngine::wires_to_face` now checks the
face (BRepCheck) and fails with a reason; `sketch_loop_defect()` finds
the crossing or cusp exactly, and the sketch tints the loop red and
marks the point.
- Recipe: saves origin, body colours and the loop auto-close setting in
a trailing block; the v4 reader is frozen as `load_flat_v4`; the 3MF
reader caps the entry at 1 GiB and the legacy entry name never overrides
the current one. New fields (`dressup_edges`, the Text parameters,
`revolve_axis_entity`) are appended at the end of the framed recipe, so
existing projects load and rebuild unchanged.
- OCCT failures are caught before `std::exception` in the new code (OCCT
≥ 8).
**Modelling**
- **Several solid edges in one Fillet/Chamfer.** Shift/Ctrl+click adds
or removes edges of the picked body; one feature dresses them all at one
size, every id resolved against the same body. The card and the offer
name the count.
- **Sketch Offset takes the whole outline** the picked curve belongs to,
with a live preview, instead of one segment (on a text outline a segment
is a fraction of a millimetre, so nothing seemed to happen).
- **Revolve about a line of the sketch.** The axis was only the sketch
plane's X or Y axis. It can now be any line of the profile sketch: a
construction centerline (preselected when the sketch has exactly one) or
an edge of the profile. The gizmo draws the axis dashed. A profile on
both sides of the axis is refused with the reason (MakeRevol failed
there with none). Surface Revolve takes the same axes.
**Panel, offer, text**
- The offer menu: its strings are translatable (the generator emits
`L()` markers, and the files are added to `list.txt`); it opens with a
title naming the selection; blocked verbs stay in place, greyed, with
their reason; each verb has one refusal wording.
- Extrude defaults to Join when the profile touches a solid, otherwise
New body. All result menus use one vocabulary: New body / Join / Cut /
Intersect.
- Interference and volume/area reports go to the status line, in
mm³/mm². Delete Body no longer asks for confirmation, since it is
undoable.
- **Text is its own feature** ("Text N" in the tree). Its dialog offers
any installed font (bold, italic) and the height in mm; it is modeless
and opens at the top right of the window, and the outline is drawn in
the view where it will go while typing. Editing the feature reopens the
dialog with its string, font and height. The outlines are saved too, so
the project opens the same on a machine without that font.
- New feature names match the card header (`Extrude 3`) and are
translated. Terminology and units are unified ("Coordinate system",
"Angle (°)").
- The Prepare Sketch/Primitive gizmos can be selected only when the CAD
feature is on.
- MCP: write methods are refused while the tab is busy (including while
the Text dialog is open); a timed-out command never runs; only a socket
is ever unlinked; `fillet`/`chamfer` accept an array of edges.
**Rendering in the Design view**
- Once extruded, a part was hard to read: both lights of the object
shaders sit near the camera, so the sides of a part came out in nearly
the same tone, and nothing marked where one face ends.
- The phong shader gains a studio lighting model, selected by a new
`lighting_model` uniform: a sky/ground hemisphere in world space, a key
light from the upper left and a weak fill, a plastic-like highlight, and
a darker base with a faint sheen toward the silhouette.
`GLCanvas3D::set_studio_lighting()` enables it per canvas; only the
Design canvas does. Every canvas sets the uniform on each use (0 for the
slicer's canvases), so Prepare and Preview render exactly as before.
- Every B-rep edge of a body is drawn as a 2 px dark line, depth tested
and pulled a few pixels toward the eye, so it hides behind the faces in
front. Seams of closed surfaces and degenerate edges are left out
(`GeometryEngine::display_edges`).
**Crash fix (Linux AppImage)**
- Drawing anything in a sketch crashed the AppImage: the constraint list
labels were formatted from narrow literals holding `—`, `·` and `°`;
`AppRun` sets `LC_ALL=C`, so wx's conversion returned NULL and
`wxString::Format` dereferenced it. They now go through
`wxString::FromUTF8`, as the panel's other non-ASCII literals already
do.
Docs updated to describe the design as it now stands: `design_tab.md`,
`interaction-model.md` and the portability note.
No change to slicing, profiles or presets. Existing 3MF recipes load and
rebuild as before: the v3/v4/v5 fixture tests pass.
# Screenshots/Recordings/Graphs
None attached. The behaviour was checked on the fork's Linux tester
AppImages, driven under Xvfb (see Tests).
## Tests
- `libslic3r_tests` on this branch, rebased on `main` (dc0e26918): 943
of 944 cases pass (1 skipped), 79,733 assertions. Linux, system OCCT.
- New or updated cases cover: the solver side and tangency fixes;
trim/offset/mirror on arcs and ellipses; body identity across history
edits; datum remapping, expressions, bindable fields; circular pattern,
hole standards, threads; the round trip of colours, origin and
auto-close; `body_touching_sketch`; multi-edge fillet/chamfer and its
save/load; the crossing/cusp loop analysis and the extrude refusal; the
Text parameters' save/load; `display_edges` on box, cylinder and cone;
revolve about a centerline, about the profile's own edge, refused
through the profile, stale axis index, axis save/load. The
truncated-recipe test accounts for every new tail field.
- All touched GUI translation units compile with GCC, and with clang
`-fsyntax-only` under the macOS job's warning flags. The modified
shaders pass `glslangValidator`.
- `gen_offer_table.py --check` passes, and `xgettext` + `msgfmt
--check-format` are clean on the CAD sources.
- On the fork's Linux tester AppImages (same code on `cad-mainline`),
driven headless:
- drawing lines and rounded rectangles no longer crashes;
- Offset on one side of a rectangle selects and previews the whole
outline; Enter adds it;
- Text: the dialog opens at the top right, "Text N" appears in the tree
and the outline in the view while typing; double-clicking the feature
reopens it and a new string redraws in place;
- a block with a through hole and fillets: faces separate by
orientation, fillets shade round, edges are drawn and hidden behind the
part;
- a half-profile beside a construction centerline: Revolve preselects
the centerline, shows the axis dashed and builds the expected tube
(4000π mm³).
- **Not verified by hand:** multi-edge selection and the red loop
marking; Windows and macOS were not run.
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The only use of Slic3r::create_thread in this file sits behind
!BBL_RELEASE_TO_PUBLIC && __WINDOWS__, so the include looked unused in
every configuration CI builds. Windows Debug and RelWithDebInfo builds
failed without it.
A Linux job configures without the precompiled header and runs clang-tidy over the C++ lines a pull request changes. The only check for now is misc-include-cleaner for missing includes; .clang-tidy is where further checks get enabled.
scripts/run_clang_tidy.sh (Linux, macOS) and scripts/run_clang_tidy.ps1 (Windows) run the same check locally: the same configure, the clang-tidy version pinned in scripts/clang_tidy_requirements.txt, and the same comparison against OrcaSlicer's main. They offer to install what is missing, or print the command to do it by hand.
* Remove Unused Project Includes and Forward-Declare Where a Type Is Only Referenced
Generated with include-what-you-use and applied conservatively. Only OrcaSlicer's own headers, the ones under src/ and tests/, are removed or forward-declared; standard-library and third-party includes are left alone. An include is removed only when both the Release and the Debug configuration leave it unused, never from inside a conditional block, and never from a file with platform-specific blocks, which only gain includes. Files whose only use of a header sits behind a feature or debug macro (libvgcode's OpenGL ES and marker code, the ARACHNE/TESTS_EXPORT_SVGS debug output) keep their includes.
clonable_ptr.hpp gains #pragma once; it had no include guard and was only safe while Config.hpp was its sole includer.
* Remove Unused Project Includes From Files With Platform-Specific Code
A Linux include-what-you-use run cannot see the code inside _WIN32, __APPLE__ or __linux__ blocks, so its verdict is only taken where nothing the removed header declares, directly or through what it includes, is named inside those blocks. Removals also have to hold in both the Release and Debug configuration and never touch a line inside a conditional block.
* Restore the libslic3r Precompiled Header and Direct Includes Lost in the Platform Pass
The platform-file pass treated pchheader.hpp as an ordinary header and
emptied it, and left GUI_Preview.hpp and 14 other files relying on
headers they no longer reached directly.
* Restore MainFrame.hpp in ParamsDialog.cpp for the Windows-Only Reparent Call
* Include Headers That Files Reached Through Ones the Cleanup Removed
* Drop Includes Duplicated by the Cleanup or by Main's Own Additions
* Leave PreciseSeam.cpp as Main Has It After the Precise Seam Rework
The Design view drew sinking outlines by looking up its bodies in the
plate's model, reading past the end of an object's volumes once a body
was committed. The Design canvas no longer draws sinking outlines.
Merged by /bot merge on behalf of @peachismomo (id 52488812).
Grants: resources/profiles/OrcaFilamentLibrary/filament/Elegoo, resources/profiles/OrcaFilamentLibrary.json, resources/profiles/Elegoo, resources/profiles/Elegoo.json
Head: 00276bf8ee
Commit to Plate now sends all visible bodies to Prepare as one object with a
part per body, so their relative placement survives. A dropdown beside the
button switches to Commit to Plate (as bodies), the previous one-object-per-body
behaviour, and the choice is remembered.
A painted relief did not survive closing the project: nothing about the feature
was written to the .3mf, so the layers, their textures and the paint mask were
all lost on save. The layer struct's cereal save()/load() serve the undo/redo
stack only, which is one session and one binary stream.
Write the layer stack and the per-volume options as JSON in the archive, and
each layer's texture as the image file it was loaded from, both referenced by
path - the same split EmbossShape already makes for its SVG. The image stays
out of the XML deliberately: it is binary and routinely megabytes, and base64
in an attribute would bloat the one file every reader parses just to list the
objects. JSON rather than the cereal stream next to it because that one is
positional and unversioned, which would make every future field a
project-breaking change; an unknown key is ignored and a missing one keeps its
default, so a project written by either side keeps loading.
The paint masks go in as one attribute per layer slot, mirroring paint_color.
Older readers ignore attributes they do not know, so a project written here
still opens in a build without the feature - it loses the relief, which is all
it could have done with it anyway.
The layers are restored while the volume is being built rather than after the
archive is walked: volumes are constructed only once the whole archive has been
read, so a hook placed after the walk ran before any volume existed and
restored nothing.
GLGizmoSlaSupports, GLGizmoHollow, GLGizmoFaceDetector, GLGizmoText and GLGizmoAdvancedCut were already left out of the build, and GLGizmos.hpp, the only header including some of them, had no includers. VoxelizeCSGMesh.hpp uses types that no longer exist, SLA/bicubic.h does not compile, and Utils/ProfileDescription.hpp is included nowhere. Their CMake and gettext source-list entries go with them.
Brings `belt-printer` up to date with `main` (4b4a261787) so that #14394
merges cleanly again, and adds the follow-up fixes the merge needs. This
PR targets `belt-printer`, not `main`.
## Commits
1. **GCode: hold the writer by value again.**
- Belt printing had turned `GCode::m_writer` into a `unique_ptr`, so
that `BeltGCode` could swap in a new writer carrying the belt
kinematics.
- Nothing subclasses `GCodeWriter`, and `set_kinematics()` can install
the belt mapping on the existing writer. This commit removes the swap.
- About 190 `m_writer->` edits revert, which takes `GCode.cpp` from 29
conflict hunks with main down to 3.
- The G-code is identical to the current `belt-printer` head on two
BabyBelt projects (see Verification).
2. **Merge upstream/main.** The resolutions are listed in the merge
commit. The ones that needed a decision:
- `write_belt_header()` follows main's relocated header block (#15897,
#15915).
- First-layer acceleration keeps the per-path first-layer plane test,
now with main's cached nozzle index (#16028).
- The arc-to-polyline fallback moves into the out-param
`extrude_arc_to_xy`, which is the overload `GCode` now calls (#16108).
- The belt fields join `GCodeProcessorResult`'s forwarding assign.
- The Clipper2 renames (#15969).
- Belt printers still reserve no CLI wipe tower (#15837).
- The new sparse-layer tower options are hidden for belt printers
(#15841).
3. **Belt: register the raw G-code toggle as a Preview shortcut.**
- Main's assignable shortcuts (#15706) replaced the key switch that
carried **B**.
- The toggle is now `ToggleBeltRawGcode`, bound to B in the Preview,
where B was free. It is listed in the shortcuts dialog and can be
rebound, and the legend shows whichever key is bound.
4. **Precise Seam: slice modifiers in the belt slicing frame.**
- The new `slice_single_volume_regions()` (#16072) sliced modifiers with
`trafo_centered()`, so on belt prints the modifier regions landed in the
unrotated frame.
- It now uses `trafo_sliced()`, as the seam enforcers and support
volumes already do. On non-belt printers the two transforms are the
same.
- A regression test is included.
5. **Belt profiles: inherit what they repeat and pass main's profile
checks.**
- The BabyBelt Pro and IR3 V2 filaments name their single extruder
variant, as the library-based filaments of other vendors do, and drop
overrides that repeat the library value.
- The Custom belt base inherits `printer_extruder_id` from its parent.
- `normalize` drops the obsolete keys. `fix-variant` gives the machine
limits their silent-mode entry, which they previously read from the
single value.
- All three vendor versions are bumped.
- Printcepts and IdeaFormer keep their own machine and process bases,
because only filaments can inherit across vendor bundles.
6. **Profile validator: accept a belt printer's tool change without a
tower.**
- The slice sweep (185cfe4323) forces a prime tower and requires its `CP
TOOLCHANGE START` block.
- Belt printers have no wipe tower: they purge into a prism object.
Their filament change is the plain `T` command, which they emit
throughout the slice.
- On belt printers the validator now looks for that `T1` line instead.
## Verification
All on Linux.
- **`m_writer` equivalence.** The current `belt-printer` head
(b22384a559) and commit 1 were each built and used to CLI-slice two
BabyBelt benchy projects (389k and 804k lines of G-code). The output is
identical apart from the per-run object ids in the `printing object …
id:` comments.
- **Merged branch, belt-specific checks.**
- The belt header is written, no `;_BELT_BAND` markers leak, and the
belt axis never steps back.
- Against the Oct 3 belt + main merge, the only G-code differences are
fill ordering on a few layers and time estimates. Both come from main's
changes since then.
- Against the pre-merge output the differences are much larger. That is
expected: main's CLI now refreshes a project's settings from its system
presets (#15953, #16038), so for example `z_hop` follows the belt
profiles' 0.
- **Tests.**
- `fff_print_tests`: 345/345 test cases pass, including 30 `[belt]`
cases.
- `libslic3r_tests`: 1081 passed, 2 skipped.
- The new Precise Seam test fails 266 of its 284 assertions with the fix
reverted.
- **Profiles.**
- `scripts/orca_profile_tool.py check` passes for all 69 vendors.
- `OrcaSlicer_profile_validator -s -l 2`: all 1271 slices succeed.
Before commit 6, the six belt printers failed.
- A flattened before/after snapshot of every belt preset shows no value
a belt printer reads has changed.
- **GUI** (BabyBelt Pro, clean datadir):
- B switches the Preview between the designed and the raw machine-frame
G-code, and pressing it again restores the view exactly.
- The legend reads "Show raw G-code (belt only) [B]".
- The shortcuts dialog lists the toggle under Preview → Display as
rebindable.
- The sparse-layer tower options stay hidden in Advanced and Expert
modes.
Main has since gained one CI-only commit (f3d0b8a553), which merges
cleanly on top.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_01AJzy1xeQV3FePh5HfahDyn
The validator slices every printer with two filaments and the prime tower
forced on, then requires the tower's CP TOOLCHANGE START block as proof
that change_filament_gcode ran. A belt printer has no wipe tower: it purges
into a prism object on the belt, so Print::has_wipe_tower() is false and
the change is the plain T command set_extruder() emits. All six belt
printers failed the sweep on that alone, although each one changes
filament throughout the slice. Look for the T1 line on belt printers.
main's profile checks reject the belt bundles: their filaments override
variant keys with one value under the library's six-variant presets, the
copied vendor commons carry keys the slicer no longer reads, the IR3 V2
and BabyBelt Pro machine limits miss the silent-mode entry, and the Custom
belt base pins printer_extruder_id to one entry under a three-variant list.
- The BabyBelt Pro and IR3 V2 filaments name the one extruder variant
their printers have, as the other vendors' filaments built on the
library do, and drop every override that only repeats the library value
(diameter, density, temperature range, most of the fan settings...).
The eSUN filaments inherit the narrowed list.
- The Custom belt base inherits printer_extruder_id from its base.
- normalize drops silent_mode, adaptive_layer_height and
tree_support_with_infill; fix-variant gives the machine limits their
silent-mode entry, which they read from the single value before.
- The IR3 V2 drops two limits equal to its base in both modes.
- Custom's index is regenerated and all three vendor versions bumped.
Flattening every belt preset before and after, nothing a belt printer
reads changes: printer_extruder_id is one id per variant, all 1.
Printcepts and IdeaFormer keep their own machine and process bases: only
filaments can inherit across vendor bundles (from OrcaFilamentLibrary), so
they cannot build on the Custom belt printer.
slice_single_volume_regions() sliced Precise Seam modifiers with
trafo_centered(), but a belt printer slices its layers with
trafo_sliced(): the belt rotation, any pre-slice remap and the lift off
the plate on top. On a belt print the modifier regions landed in the
unrotated frame, away from the walls they were meant to place the seam
on. Slice them with trafo_sliced(), as the support volumes and the seam
enforcers already are. It equals trafo_centered() off a belt printer.
Main's assignable shortcuts replaced the canvas key switch that carried
the belt "show designed / show raw G-code" toggle on B. Register it as
ToggleBeltRawGcode, bound to B in the Preview (B is only taken on the
Plater, by the mesh boolean gizmo), so it can be rebound and is listed in
the shortcuts dialog. The legend checkbox shows whatever key is bound.
Brings belt-printer up to main 4b4a261787. Resolutions:
- G-code header (#15897, #15915): main moved the header, config and
thumbnail block later in _do_export; write_belt_header() moves with it,
still after the thumbnails and outside the BTT_TFT gate.
- _extrude: first-layer acceleration keeps the per-path first-layer plane
test with main's cached nozzle index (#16028); main's set_speed out-param
form (#16108) everywhere else.
- GCodeWriter (#16108): the arc-to-polyline fallback for machine mappings
that cannot express G2/G3 now runs in the out-param extrude_arc_to_xy,
which is the overload GCode calls, and appends to the caller's string.
- GCodeProcessorResult: the belt fields join main's forwarding assign.
- Clipper2 (#15969): belt arrange helpers take Slic3r::Point; the tree
support join types lose their ClipperLib qualifier.
- CLI arrange (#15837): belt printers still reserve no wipe tower.
- Wipe tower options (#15841): the two new sparse-layer toggles are hidden
for belt printers like the rest of the tower options.
- Keyboard shortcuts (#15706): main's registry replaces the old key switch;
the belt view toggle is re-registered in the next commit.
- Print::process: the belt purge-plan undo runs before main's SliceStarted
event.
- scripts/filament_id_snapshot.json: deleted on main (a77209af8f).
- Includes and appended tests: union of both sides.
Belt printing turned GCode::m_writer into a unique_ptr so BeltGCode could
swap in a freshly built writer carrying the belt kinematics. Nothing
subclasses GCodeWriter: the machine mapping lives in its MachineKinematics,
which set_kinematics() installs on an existing writer. A GCode is built for
every export and the only state on the writer when init_belt_writer() runs
is the plate offset, which the swap had to copy across by hand.
Install the belt kinematics on the writer in place, drop the copied offset,
and drop the virtual markers on GCodeWriter that the old subclass needed.
Every m_writer-> in GCode.cpp goes back to m_writer., which is most of the
belt diff in that file and most of its conflicts with main.
The pressure-advance pattern keeps its shared_ptr writer: the unique_ptr
kinematics make GCodeWriter move-only and that class must stay copyable.
Selections are drawn as opaque faces in the selection colour with a cased outline instead of a
translucent tint over the body, so they read on a body of any colour. Selecting a Feature tree
row lights the faces that feature made rather than its whole body, which also makes fillet and
chamfer rows highlight again.
# Belt Printing Bug Fixes & Feature Updates
This should be the majority of substantive work keeping ``belt-printer``
from being ready to merge into ``main``. It includes Hanif Koh's review
fixes from #15685 and the answers to his review on #14394, findings from
running the branch on a BabyBelt Pro and an IR3 V2, crash fixes
contributed by Unlayered3D, and arrange and purge-tower changes for
multi-colour belt prints.
The merge of current `main` into this branch is prepared and tested
locally. The conflicts are in the acceleration refactor of
`GCode::_extrude`, the ClipperLib namespace clean-up and a few test
files.
Tested with `libslic3r_tests`, `fff_print_tests` and `libnest2d_tests`
on Linux, validated on a stock Klipper BabyBelt Pro.
## New features
**Belt arrangement.** Parts of the same colour are grouped along the
belt into a single print run. Packing starts at the end that prints
first, following the slicing rotation and the sign of the angle. Arrange
reserves the purge prism's strip and the brim width along the bed edges,
then regenerates the prism from the result instead of moving it as a
part. Piles aimed at an off-centre `best_object_pos` are clamped to the
bed. Grouping uses a soft cost: if the belt is too short for separate
runs, colours overlap rather than move to another plate.
**Purge tower sizing.** The prism stops at the plate end. The purge
planner's existing warning reports what a shortened bar can't absorb. A
brim is accepted next to the purge tower again because the purge plan's
layer-grid shift now also moves the brim's apron bands.
**First-layer fan band.** On a belt, "the first layers" are a band along
the belt rather than the first slicing layers. The generator marks where
each extrusion enters and leaves the band. The cooling buffer keeps the
fan off inside it on every layer, taking precedence over overhang and
bridge fan requests.
Thanks to:
@Unlayered3D, @shubhracc, @dlc60, @Rexit
**Profiles.** Z-hop defaults to 0 on belt printer bases and belt
filaments; it can be turned back on. Axis remap options are shown only
in Develop mode. IdeaFormer, Printcepts and Custom bundle versions are
bumped.
Thanks to: @RobMink, @Rexit
## Bug fixes
- Scarf joint seams no longer start below the layer on a belt.
Previously, each seam caused a 0.28 mm belt back-step into the previous
layer ("the belt jumped backwards and the head hit the part"). — credit:
@dlc60
- The CLI no longer rejects every belt print with -102. The
printable-height check compared machine Z, which is belt travel on a
belt printer.
- The belt header is written outside the optional file header block, so
printers with a BTT TFT thumbnail still get belt view in the preview. —
credit: BabyBelt Discord
- The dormant tilted-bed rendering is removed from Prepare view; the bed
is shown as the slicing pipeline treats it. — credit: HanifKoh
- Plate icons, number and name no longer run across the neighbouring
plate on a long, narrow bed; their scale is bounded by the gap between
plates.
- Modifiers and support blockers no longer extend a belt object's sliced
range. The `is_model_part` filter had gone missing with some debug
logging. — credit: HanifKoh
- Crossing-perimeter avoidance no longer dereferences a null layer in
either pass while travelling on a brim apron layer. — credit:
Unlayered3D
- 3MF files with non-finite vertex coordinates are rejected instead of
crashing qhull during load. — credit: Unlayered3D
- The CLI no longer crashes on a project without `printable_height` or
with fewer filaments than were loaded. — credit: Unlayered3D
- The island tour cache is keyed on the island layout, preventing
out-of-bounds reads on later layers with fewer islands. — credit:
Unlayered3D
- The top/bottom painting projection no longer erases from an empty
vector when no shell layers are requested. — credit: Unlayered3D
- Belt purge planning detects filament changes by scanning the tool
ordering instead of checking the first layer's flag, which a brim apron
layer never carries. — credit: Unlayered3D
- The purge prism never gets a brim, regardless of its config. — credit:
Unlayered3D
- Containment tests treat the plate as open along Y on an infinite-Y
belt printer. — credit: Unlayered3D
- Belt brim lattice lines close to the belt move uphill; narrow bands no
longer get near-duplicate lines.
- Organic supports that reach the belt slice without negative flow.
- Hanif Koh's review items: restored the gantry clearance check in
`Print::validate`, read the pre-slice remap header at its real length,
removed unused `clip_support_fills()` and the two unimplemented support
floor modes (legacy values map to `none`), dropped the per-extrusion
transform determinant, indexed apron layers into the first layer's
nozzle map, read the brim axis from the config, removed tagged
diagnostic logging and planning-doc references, and documented the
exclude-object frame. — credit: Hanif Koh
- Hanif Koh's fixes from #15685 include the plate offset in the belt
writer, painted supports and seams under the belt transform, shared
build-plate tilt helpers, the belt header as the source of the tilt,
brim band loop and filament, and G-code export invalidation. — credit:
hanifkoh
At this time there are no known issues with belt printing nor any known
regressions in non-belt-printing execution paths. I have been using
these builds for all of my printing for several months now and have had
no issues.
* Add opt-in printer overrides for filament tool-change settings
Allow printer presets to define uniform ramming, loading, unloading,
cooling, purge, filament scripts and pressure-advance enable settings
without duplicating material presets. Apply overrides during preset
composition and FDM normalization, and expose the switch in Multimaterial.
Keep the feature disabled by default. Omitted or empty override vectors
preserve material settings; a single value applies to every filament,
including an explicitly empty script. Reject multi-value overrides.
Preserve empty float vectors across project serialization and initialize
empty nullable filament overrides before resizing them, preventing preset
cache generation from accessing an empty vector.
Include focused override tests and document the configuration semantics,
Prusa MMU3 integration and INDX tool-change behavior.
Co-authored-by: Codex <codex@openai.com>
* Add Prusa MMU3 and CORE One INDX profiles with shared material tuning
Add MK4 MMU3 and four-tool/eight-tool CORE One INDX printer definitions,
process presets and printer resources. Reuse ordinary MK4 and CORE One
printer/process inheritance while retaining device-specific startup,
shutdown, tool-change and wipe-tower behavior.
Move uniform MMU3 tip forming and INDX handling into machine filament
overrides. Keep MMU3 pressure advance and purge material-specific, retain
INDX material tuning, and share surviving materials with migration aliases
for retired MMU3 and XL tool-change copies.
Preserve unrelated Prusa filament identities, scalar value formats and
inheritance rather than applying broad profile cleanup.
Validation: Prusa profile checks and all 69 printer smoke slices passed.
The final cleanup preserved emitted commands with identical filament
selections.
Co-authored-by: Codex <codex@openai.com>
* Refresh filament controls after loading printer presets
Synchronize the plater filament controls after preset loading, even when the internal filament list already matches the nozzle count.
Co-authored-by: Codex <codex@openai.com>
* Fix nullable Z-hop overrides in Prusa filament variants
Represent empty overrides as nil for each inherited extruder variant so
the native profile loader preserves machine Z-hop settings.
Validation: full profile checks, native loading, and 1,115-printer slicing
sweep passed.
Co-authored-by: Codex <codex@openai.com>
* Separate Prusa profiles from machine-owned filament overrides
Retain profile tuning without unsupported machine override keys. Move supporting code, tests and override documentation into a separate feature change.
Co-authored-by: Codex <codex@openai.com>
* Default INDX tools to hardened high-flow nozzles
Use the High Flow variant for every INDX tool and its dedicated filament presets. Raise Generic PLA throughput to 28 mm3/s.
Co-authored-by: codex <codex@openai.com>
* Use normal filament-change lifts for INDX and MMU3
Avoid duplicate INDX retraction and account for its 12.5-second dock swap in print estimates.
Co-authored-by: codex <codex@openai.com>
* Set z_hop_types for the machine too
* Fix profile check failures in the Prusa CORE One filament presets
---------
Co-authored-by: Codex <codex@openai.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
* Add Pragostroj KINARB profile set
This adds the new Pragostroj vendor profile with KINARB 1HB and 2HB machine models, nozzle variants, common machine/process settings, and default material mappings. It also includes the corresponding filament and print presets for PLA, PETG, PP, and HIPS, covering the printer family’s standard profiles and tuning.
Each Feature tree row now carries Edit, Show/hide and Delete, and each
Bodies row carries Move, Show/hide and Delete. These act on that row
instead of on the selection. The eye shows a closed eye when the item
is hidden. Every clickable icon in the sidebar now highlights on hover.
Rename and Color now come first in the right-click menu, as items of
their own instead of inside Modify. This applies wherever they appear,
whether opened from the viewport or from a Bodies row.
Clicking an already-selected row no longer starts a rename; use F2,
the row's right-click menu, or the Rename command.
Restore slicing progress after notification reset
Ensure the cleared slicing-progress controller is recreated before its initial state transition, and calculate the Daily Tips size before positioning the popup.
perf: skip estimating curled walls when nothing reads them
The curled extrusion estimate ran whenever a region had overhang speed on,
which is the default, but only the slowdown for curled perimeters reads the
curled lines it produces, and that slowdown is off by default. The step now
also requires a region with the slowdown on, and clears the curled lines
when it skips the estimate, so none are left from an earlier slice.
Also fixes stale fan commands due to the stale curled lines on the reused layers.
Every file that included STEP.hpp, directly or not, got namespace fs = boost::filesystem at global scope, and 29 sources and three headers relied on it without saying so. Headers now spell out boost::filesystem, and each source that uses fs declares the alias itself.
all_paths_inside() accepts the path bounding box only within 3*EPSILON of the
bed and otherwise tests every move; a belt print's moves are machine-frame
coordinates whose Z is belt travel, so that test can never pass, and the
designed view's min-corner anchor leaves the box a fraction of a millimetre
below zero. Every multi-object belt plate therefore reported a path beyond the
plate. The belt preview now judges the back-transformed box with a millimetre
of room.
ensure_belt_purge_tower only ever looked at the current plate and kept a single
prism, so the other plates had no tower and switching plates moved the one
prism around. Every plate is now planned on its own: a prism that lies on no
plate is stale, a plate whose prism matches its recorded inputs is left alone,
the rest are deleted and recreated, highest index first.
The plate's icons, number and name scale with the plate's depth, but they sit in
the gap to the next plate, which scales with its width. On a long, narrow bed
(a 95 x 500 mm belt) they came out 40 mm wide and ran across the neighbouring
plate. The scale is now also bounded by the gap, which leaves ordinary beds
unchanged.
* Fix CLI Crashes on Malformed Project, Assemble List and No-Input Runs
Four CLI paths indexed vectors without checking their size and crashed
with SIGSEGV on malformed input:
- A project inherits_group whose length is not the filament count plus
the process and printer entries was split by position. It is now
ignored with a warning, as if the project had none.
- An assemble list object with an empty filaments list passed validation
and was then read at index 0. It is now rejected as a config error, as
is a negative filament id.
- --slice N --arrange 1 on a project without plate metadata read the
missing plate data. It now falls back to the plate's own filaments,
like the other plate data reads.
- --assemble with no input model built an object with no volumes. It is
now rejected as invalid parameters.
A tests/cli script covers each case through the binary, since all four
live inline in CLI::run().
* Move the Assemble List Parser into libslic3r
Behaviour-preserving move of the --load-assemble-list JSON parser and
its plate/object structs from the CLI into libslic3r/Format/AssembleList,
so the format can be unit tested. The parser returns its own
AssembleListResult and takes the plate limit as a parameter; CLI::run
maps the result to the same exit codes as before. Every validation rule
and log message is unchanged.
Adds Catch2 coverage of the valid layout and each validation rule.
* Keep the Process and Printer of an inherits_group of the Wrong Length
A project whose inherits_group did not have one entry per filament plus
the process and printer entries was loaded as if it had none. The CLI
then looked for system presets under the names of the user presets,
found none and refused to slice a project that slices on main.
The group is now read as before: the process first, the printer last
and the filaments in between, up to the filament count. A filament
without an entry counts as a system preset. A group with fewer than two
entries is still ignored. The warning stays.
Plater::set_bed_shape read the belt keys from the plater's own config, which
never carries them, so the branch that tilted the bed model, drew the slicing
arrow and plane and switched the build volume to belt mode never ran. The
Prepare view shows the bed as the slicing pipeline treats it, flat; the
gravity arrow from build_plate_tilt stays, as does the preview's belt view,
which takes its angle from the G-code header.
The G-code viewer takes the belt tilt only from the belt header comments, but
they were written inside the header block that is left out when a BTT TFT
thumbnail is configured, so such a printer never got belt view. The comments
are not part of the header block; they go after it, and after the thumbnails
that firmware needs first.
The purge plan snaps every object onto one layer grid after the brim is built;
the per-layer brim bands follow their layers but the apron bands below the
first layer carry their own print_z and were left behind, which is why a brim
was refused next to a purge tower object. The shift now moves them too and the
combination is accepted again.
On a belt the parts print in belt order, so every colour change between parts
is a filament change. Arrange packs items in extruder order already, but it
grew the pile around its centre, so the colours ended up interleaved. A belt
print now packs from the leading end of the bed, each row filling across the
belt before the pile advances, and the objective charges an item for every
packed part of another colour it does not fully follow along the belt,
counting the tilted layers that reach cot(angle) * height past a part, so
each colour prints as one run. The direction follows the slicing rotation: a
rotation about X prints toward +Y, one about Y toward -X, and a negative angle
flips it. Packing from the edge also means the brim has to be kept on the
bed: a belt brim is printed brim_width wide for every brim type, so that much
is reserved along every edge (between parts the brims may overlap, as on any
printer).
The purge prism is regenerated from the arranged parts, flush with the far
edge of the bed, yet arrange moved it about like a part and packed parts into
the strip it comes back to. Arrange now skips the prism and reserves its strip
with a fixed virtual item, like a bed exclusion area, whenever the parts use
more than one filament.
The prism's length follows the parts plus a ramp per unit of height; with the
height at its cap that ran 100 mm past the end of a 500 mm belt and the project
could not print. The bar now stops at the plate end, and the purge planner's
existing warning reports what the shortened bar cannot absorb.
Release builds install each vendor as its preset cache alone. The
read-only preset load the CLI uses to resolve an inheriting user preset
passed allow_cache = false to keep caches from being written, which
also stopped them from being read, so every vendor fell back to JSONs
that are not installed and the CLI failed.
The flag now only gates writing: a read-only load reads caches and
writes none. The filament library is also read from its cache whenever
that is all that is installed, so a vendor updated over the air still
resolves against it.
* Refresh a CLI Project's Filament Settings From Their System Presets
The CLI loads a project's printer and process settings as the GUI does,
taking every key the project does not list as changed from the current
system preset, but it kept the stored filament values. A project saved
before a profile update then sliced with old filament values on the
command line and with the current ones in the GUI.
Every project filament that no loaded filament replaces is now resolved
by its system preset name and fed to the filament merge the up-to-date
path already uses, which keeps the keys listed in
different_settings_to_system and maps per-variant values onto the
preset's variants. This covers a plain run, --uptodate without
--uptodate-filaments, and the slots --load-filaments leaves empty. The
merge tells refreshed entries from loaded ones per entry instead of by
the global loaded-filament count, and the entries are kept in slot
order. A project filament saved under a name the presets have since
split per nozzle is resolved through the name conversion the GUI uses,
which PresetBundle now exposes.
* Check the Project Refresh Test's Result Directly
Shellcheck SC2181: test the checker's exit status in the if instead of
reading $? afterwards.
After undo, redo, New Design, a project load, a delete or a reorder, a sketch
profile picked beforehand stayed selected even though it was gone. The
right-click menu then offered "Sketch profile" over an empty design with Create
greyed out. Those operations now clear the selection.
The Confirm button no longer stays greyed in a sketch session after a card whose
preview was invalid has been closed.
Delete and reorder now wait while a sketch or constrain session, the Text dialog,
an Insert placement or the move gizmo is open, instead of editing whichever
feature took its place in the list.
* Save device url in all cases and load printer url after hot-reload finishes
* Recreate web view from scratch as only URL fix seems not robust enough
* Add the same robust browser recreation for WebViewDialog
It should eliminate possible issue with blank Home and other pages
in the same way as Printer page
* Remove redundant fallback leftover
* Fix webview reset state and replay Project info on page reload
The first-show webview reset now runs only on Windows, reloads the last
printer URL and resets the Project page's ready state. The Project tab
replays its 3MF info whenever the page reloads, so it no longer goes
blank after a theme switch or a slow first load. The Device tab no longer
loads an extra time on first open, and the Home tab no longer navigates
twice. NeedsRecreateOnShow() logs is_recreating_gui so one language
switch shows whether the reset ever fires.
* Build plugin pages on first show
A language switch rebuilt every plugin page's browser while the main
window was being recreated, which left plugin tabs blank on Windows.
Plugin pages are now lazy pages, never prebuilt, and are removed left
to right so removing pages never builds one only to destroy it.
A plugin page's script now starts when its tab is first opened;
messages posted before that are dropped.
---------
Co-authored-by: SoftFever <softfeverever@gmail.com>
Co-authored-by: Noisyfox <timemanager.rick@gmail.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
The sidebar can move to either side, float, be resized, and collapse with
the canvas button or Shift+Tab. Its layout is remembered separately from
Prepare's, starts where Prepare's sidebar is, and View > Reset Window
Layout resets both tabs.
* Add Missing Includes Across src/libslic3r
Every libslic3r source and header now directly includes the headers declaring what it uses, rather than relying on the precompiled header or transitive includes. Generated with clang-tidy misc-include-cleaner, with libslic3r headers spelled libslic3r/... so they resolve outside the library's private include paths. MultiMaterialSegmentation.hpp, Support/SupportParameters.hpp and Format/STEP.hpp are made self-contained by hand.
* Make the libslic3r Headers Compile on Their Own
Each now includes, or forward-declares, what it uses instead of relying on what its includers happened to include first. Left out: I18N.hpp, which errors on purpose when included from GUI code, and VoxelizeCSGMesh.hpp and SLA/bicubic.h, which nothing includes and which no longer compile at all.
* Add the Includes Missing From the Hand-Fixed libslic3r Headers
clang-tidy would not edit these headers while they failed to compile on their own, so the first pass skipped them. With the headers now self-contained, a second pass adds the rest.
* Keep Windows Setup Ahead of the Added libslic3r Includes
Print.cpp and Thread.cpp open with a _WIN32 block that has to come first; without the precompiled header, Print.cpp otherwise reaches windows.h through OCCT with NONLS defined and boost/regex fails. OpenVDBUtils.cpp and SLA/SupportTreeBuilder.cpp had includes inside #ifndef NOMINMAX, which libslic3r defines on Windows, so those were skipped there. .clang-tidy also ignores the MSVC STL and UCRT internals, Boost.Multiprecision's fwd.hpp and CPython's Windows include directory.
* Re-Add libslic3r Includes After the Clipper2 2.0.1 Migration
Rebasing onto main took main's version of the files the Clipper2 migration rewrote, so their added includes are restored here, along with includes for main's new code. Clipper2's individual headers are now ignored by clang-tidy: they only build the Z variant through clipper2_z.hpp, which defines USINGZ first, so including clipper.core.h and the like directly broke ClipperZUtils.cpp.
* Ignore Clipper, libpng, mcut and Boost.Polygon Internals in clang-tidy
Each only works through a wrapper or umbrella header: libslic3r/clipper.hpp or clipper_z.hpp configure Clipper before including it, png.h pulls in libpng's config headers, and Boost.Polygon's headers only compile through polygon.hpp or voronoi.hpp.
* Ignore minilzo's Config Headers in clang-tidy
lzoconf.h and lzodefs.h are internal to minilzo.h, which is what the code includes.
* Add Missing Includes Across the Remaining Sources and Tests
Covers src/slic3r/Utils, src/slic3r/plugin, src/slic3r/Config, src/libvgcode, src/dev-utils, src/OrcaSlicer.cpp and tests/, the directories left after src/slic3r/GUI and src/libslic3r. Generated with clang-tidy misc-include-cleaner. libvgcode's own headers are included by relative path as in the rest of that library, and Catch2 and pybind11 with angle brackets as elsewhere in the repo.
* Make the GUI and Test Headers Compile on Their Own
Each now includes, or forward-declares, what it uses instead of relying on what its includers happened to include first. Headers that only compile on one platform, or that nothing built includes, are left alone.
* Keep Windows and nanosvg Setup Ahead of the Added Includes
OrcaSlicer.cpp and several tests set _WIN32_WINNT, WIN32_LEAN_AND_MEAN or NOMINMAX before including Windows.h, and the profile validator defines NANOSVG_IMPLEMENTATION before any libslic3r header. The added includes had landed above those blocks, which broke the Windows build.
* Add the GUI Includes the First Pass Missed
Covers headers that only became editable once they compiled on their own, and wx symbols whose suggested header changed as the clang-tidy ignore list grew after the src/slic3r/GUI pass.
* Keep the Added Test Includes Below the NOMINMAX Guard
test_marchingsquares.cpp and test_texture_displacement.cpp had includes inside #ifndef NOMINMAX, which the tests inherit as defined on Windows from libslic3r, so those were skipped there. .clang-tidy also ignores the MSVC STL and UCRT internals, Boost.Multiprecision's fwd.hpp and CPython's Windows include directory, as in #16068.
A page built on first click is now built before it is shown, as the idle
prebuild already does, so its controls are not created inside a visible
window. This takes the Design tab's first open on Windows from ~9 s to ~1 s.
A tool change added from the layer slider switches heads mid-print like a
painted color, but the parallel-mode checks only looked at the filaments of the
plate's objects, support and prime tower. A one-filament plate with a slider
change to a copying head's filament passed them. They now include those tool
changes, as the plate's warning badge already did.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Only the primary and its Span tools print a Span plate's colors; the other
active tools replay them. The multicolor rule never checked where each color
went, so a filament on a copying or mirroring tool, an unused one, or past the
end of the physical extruder map was accepted. It is now refused, and the
message names the tools that print colors and where the map sends the
offending filament.
The plate's warning badge also read the raw physical_extruder_map rather than
the effective one slicing uses, so on printers that set no map it disagreed
with the slicer.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Three commits: G-code export caches its filament config slot and repeated
option lookups, per-plate bed type overrides follow the printer's multi-bed
support, and the gizmo checkboxes and texture displacement panel get styling
and refresh fixes. No conflicts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The name field rebuilt its suggestions in wxEVT_COMBOBOX_DROPDOWN, which
ComboBox sends after it has sized and shown the popup from the items it
already held. With every suggested name taken, the popup opened around an
empty list: a small empty box on GTK, a black one on Windows. The list is now
rebuilt from the field's own mouse-down and double-click, which run ahead of
ComboBox's handlers. With nothing to offer the popup stays closed and the
click focuses the field instead.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The editor only replaced an empty name, so a row could be given a name another
row already had. A plate stores its mode by name and find_imex_mode() takes the
first row with it, so the second row was unreachable, and the plate's mode
list repeated the name: left-click stuck on it, or looped without getting back
to Primary.
An edited name that another row already carries, or the reserved Primary name
in any case, is now replaced when the edit is committed: "copy" becomes
"copy 2". The edited row yields, so plates keep resolving to the row they
meant, and tabbing through a field without changing it checks nothing.
Resetting a row to a saved name that another row has since taken does the
same. The plate's mode list comes from imex_plate_mode_choices(), which lists
each name once, so a profile that already has duplicates still cycles.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The editor took its background from the app's window default, which in dark
mode is a gray the dark-mode walk has no entry for. In dark mode it never
matched the page's palette color, and an editor built in dark mode kept that
gray after a switch to light while the controls inside it changed. It now
takes the page's color, which the walk remaps both ways.
Its labels and G-code text had no color of their own, so on Windows they took
the system's text color, which follows Windows' theme rather than Orca's. They
now start from the page's label and input text colors, which the walk maps
with the page.
The walk also ran the tool tiles' role colors through the palette map, which
turned their labels gray after a theme change. The tiles carry wxBU_AUTODRAW,
the flag it skips, so the role colors stay as set.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The remove button drew imex_remove.svg, a redrawn delete.svg, and the help
button drew the mascot question icon. They now use delete and icon_qusetion,
the tip icon the send-print dialog uses, and imex_remove.svg is gone.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Fix gizmo checkbox styling and Texture Displacement resize
Use shared BBL checkboxes in Texture Displacement and restore white toolbar checkmarks so other gizmos keep proper checkbox contrast. Also fix Texture Displacement resizing only after mouse movement by requesting additional frames while its layout is still changing.
* Fix gizmo checkbox styling and Texture Displacement resize
Use shared BBL checkboxes in Texture Displacement and restore white toolbar checkmarks so other gizmos keep proper checkbox contrast. Also fix Texture Displacement resizing only after mouse movement by requesting additional frames while its layout is still changing.
* Update GLGizmoTextureDisplacement.cpp
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
The update merged the same upstream commits this branch already carries. Its
GLCanvas3D resolution kept the assembly view painter block after the ghost
render, which section view removed along with the members it calls, so that
tree does not build. This merge keeps the removal; nothing else differs.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Five commits: a section view for the 3D canvas, a faster G-code preview, user
preset values kept on extruder variants they don't list, debug build CMake
fixes, and a 30 minute timeout on macOS notarization.
Two conflicts, both in GLCanvas3D. The header's were neighboring
declarations, kept from both sides. In the transparent pass, the section view
removed the assemble view's painter block that followed our ghost render;
the ghost render stays where it was, at the end of that pass.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The band was a second pass over the finished layer that fought the fan commands
the layer pass had already written (overhang, bridge and resume requests). The
generator now marks where each segment enters and leaves the band and the layer
pass treats the band as the strongest fan request, so there is one place that
decides the fan.
The purge plan moves objects onto a common layer grid after the brim bands are
built, so the two cannot share a print. The prime tower setting alone still does
not block a brim. The missing-prism warning now counts the filaments the objects
use, as the GUI does.
notarytool submit --wait has no timeout. On 2026-10-02 it hung for
over 5 hours in a main build. Since #16044 a new push no longer
cancels a running main build, so nothing stopped it and six waiting
main runs were replaced without starting.
Over the last 30 days the step succeeded 206 times, with a median of
4.3 minutes and a maximum of 14.8.
With best_object_pos away from the bed centre the placer packs the pile
inside the bin and then translates it so its centre lands on that point,
without checking that it still fits there. A belt printer aims at the
leading end of the belt (BabyBelt Pro: 0.5, 0.05), so any pile longer than
the 50 mm around that point was pushed past the edge: four 90 mm parts on
the 95 x 500 mm belt ended with one across the edge and one outside while
290 mm of belt stayed free.
The final alignment now stops the pile at the edge of the bin; the items'
inflated boxes leave the object spacing as the margin. A pile that does not
fit along an axis is centred on it, as before.
The plate's mode button showed one icon whatever the mode. It now shows the
mode the plate slices as, the one its ghosts follow. imex_mode_kind() reads it
from the heads beside the mode's primary: none is Normal, any Span head is
Custom (the multicolor modes), otherwise any Mirror head is Mirror, and heads
that all copy are Copy. One Mirror head is enough because an IQEX mirror mode
copies within the primary's gantry.
The icons are Felix14-v2's: the four kinds, each in light and dark with a hover
state, replacing the single mode icon. Two fixes to them: the light Normal
border used the dark theme's gray, and an opacity="undefined" attribute hid one
of its strokes in nanosvg. The knight outlines are drawn at 1.0 rather than
0.8, matching the other plate icons and keeping Copy and Mirror apart when
zoomed out.
Co-authored-by: Felix14_v2 <75726196+Felix14-v2@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Fix debug build after qhull upgrade
We upgraded qhull from 8.0.1 to 8.0.2 in 504a5d3b70, which contains a commit qhull/qhull@16159c648c `use same CMake target name for Debug and non-Debug`, so this target name check is no longer required
* Fix issue like `IMPORTED_LOCATION not set for imported target "opencv_world" configuration "RelWithDebInfo".` when build Debug config
Seven commits: painted multi-material segmentation made deterministic, missing
includes added across src/slic3r/GUI, U1 high-flow nozzle variants, Windows
ARM64 build and HTTPS fixes, and a rule added to the wxWidgets agent skill.
Three conflicts, all include lists: upstream's include pass and ours each added
to the same lists in GCodeViewer.hpp, PartPlate.hpp and PartPlate.cpp. Resolved
as the union of both. Three includes both sides had added at different places,
which git kept twice, are kept once: Color.hpp and <set> in PartPlate.cpp,
<sstream> in Plater.cpp.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The IMEX fixtures set seven nozzles but kept multifilament_config's single
filaments x filaments flush block. get_flush_volumes_matrix splits that block
across the nozzles, leaving each with 7 values, and
ToolOrdering::reorder_extruders_for_minimum_flush_volume then reads them as a
7 x 7 matrix, past the end of the buffer. One of the affected tests segfaulted
on Windows x64; ASan reproduces the overflow in that test on Linux, where it
passed only by luck.
The fixtures now repeat the block once per nozzle, as the GUI does, and size
flush_multiplier to match, since append_full_config takes the nozzle count from
it. The helper is shared in test_helpers.
Co-authored-by: HanifKoh <76276251+HanifKoh@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Keep User Preset Values on Extruder Variants They Don't List
A user preset stores the variant list its parent had when it was saved.
When the parent later gains variants, update_diff_values_to_child_config
matched variants by name only and left the new ones at the parent's
value, so the user's settings were silently replaced there, and a
re-save wrote the system values into the user's file.
An unmatched parent variant now takes the child's first variant of the
same extruder, the rule slicing already uses in get_config_index_base.
A child without a variant list covers the parent's first extruder. The
name match also no longer indexes the child's extruder ids when it has
none.
* Share One Variant Column Rule Between Slicing, User Presets and Projects
Three places chose which variant column a value comes from, each with
its own copy of "the same variant and owner, else the owner's first
column": get_config_index_base when slicing, the user preset merge in
update_diff_values_to_child_config, and normalize_filament_values_to_variants
for projects and the CLI.
find_variant_column now holds that rule and map_variant_columns applies
it to a variant list, so a change to how missing variants are filled
reaches all three. Each caller keeps its own copy step. There is no
behaviour change: G-code is identical before and after. The one
relaxation is that get_config_index_base no longer reads past a short
id list when its two lists differ in length, which its assert already
rules out.
* Rename variant column helpers to variant index
---------
Co-authored-by: SoftFever <softfeverever@gmail.com>
* fix: include bundled UV binary for arm64
* fix: update unit test CI
* Install unit-test numpy only with the bundled uv
* Simplify the unit-test script's uv lookup
---------
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
The ARM64 jobs pinned CMake 3.31 because CMake 4 dropped pre-3.5 policy
compatibility and its ARMASM support broke Boost.Context. Both are
handled now. deps/CMakeLists.txt sets CMAKE_POLICY_VERSION_MINIMUM on
CMake 4, and Boost.Context uses the winfib implementation on ARM64, so
nothing assembles with armasm.
CMake 3.31 also predates VS 2026. Its InstallRequiredSystemLibraries
treats the v145 toolset as v143, searches only the VS 2017-2022 install
directories and finds no runtime, so the ARM64 installer ships without
msvcp140.dll and vcruntime140.dll. CMake 4.2 and newer find the VC145
redistributable.
get-cmake also installs Ninja, so the ARM64 jobs now use its latest
release instead of the one already on the runner, as x64 does.
The install now fails when InstallRequiredSystemLibraries returns no
msvcp140.dll or vcruntime140.dll, after a configure warning naming the
CMake and MSVC versions. A CMake that predates the Visual Studio in use,
on a developer machine or after the next runner image update, then
stops the installer build instead of shipping one that cannot start.
The build_win.bat prerequisite installer drops its matching 3.31.8 pin.
This is the bulk of the profile changes for U1 that led to the cooling
catiant connection.
The ugly end gcode is used by the printer's UI to complain if normal
nozzle is installed but the file was sliced for highflow.
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
* Make Painted Multi-Material Slicing Deterministic
Painted (multi-material) models sliced to slightly different G-code on
every run: ±1 µm wall coordinates and reordered islands. Hashing each stage
of the segmentation across runs showed the projected painted lines and the
per-layer Voronoi segmentation were stable; the raw top/bottom projections
from slice_mesh_slabs() were not. Three causes, all thread-order dependent:
- slice_slabs_make_lines() appends each slab's intersection lines from a
parallel facet loop and never restored a canonical order, so the loop
start vertices and polygon order from make_slab_loops() depended on
scheduling. Sort every slab's lines with the same key slice_make_lines()
already uses.
- segmentation_top_and_bottom_layers() wrote a layer's shell projections
into neighbouring layers' vectors from the parallel loop, relying on a
parity double-buffer that assumes TBB ranges are exactly one group wide
and aligned, which blocked_range does not guarantee; two threads could
append to the same vector. Each source layer now records its projections
in its own slot and they are gathered per target layer in source order.
- The painted-line sort in post_process_painted_lines() was not a total
order: projections of one span from facets of different colours tied on
every key and the first one won the span. Colour and end points now break
the tie.
Three multi-threaded runs of each painted fixture now give one G-code;
unpainted output is unchanged.
* Test That Slab Slicing Does Not Depend on the Thread Schedule
Projects a dense, tilted sphere with slice_mesh_slabs() on one thread and
then three times multi-threaded, and requires the polygons to match exactly,
vertex order included. Fails without the canonical line sort, passes with it.
A tilted layer runs from the belt to the top of the part, so a wall loop
that starts above the belt still passes along it. Tagging only the path's
first point left such loops out of the band entirely; the band is now
evaluated at each segment, with the tag capped where the fan stops
depending on it.
The clearance test needs the relative-E reset in its layer change G-code to
get past validate()'s other checks, and now asserts the height message. The
fan band test counts cycles rather than commands: the band is decided per
path start, so a cube cycles the fan far less often than a benchy.
* Add Missing Includes Across src/slic3r/GUI
Every GUI source and header now directly includes the headers declaring what it uses, rather than relying on the precompiled header or transitive includes. Generated with clang-tidy misc-include-cleaner, plus one hand edit making CalibrationPanel.hpp self-contained.
* Drop the OS-Specific Includes Added Outside Their Platform Guards
GLib, GTK, D-Bus and POSIX headers are only used inside platform #if blocks, which already include them. Added unconditionally at the top of the file they broke the Windows build.
* Add the clang-tidy Configuration That Generated These Includes
Only misc-include-cleaner's missing-include check, with the headers it must never suggest: per-platform, internal and OS-specific ones that would break other platforms or are not meant to be included directly.
* Match Windows Paths in the clang-tidy Ignore List
Header paths use backslashes on Windows, so every / in a pattern is now [/\\]. The Windows SDK headers are ignored alongside the other OS-specific ones, and the list is one pattern per line. Suggested by @raistlin7447 from a Windows clang-cl run.
Covers the case from Hanif Koh's review of #14394 (belt raft layers below
the object with no lower bound), which the negative-Z bottom layer fix in
layer_initialize() addresses.
With a first layer of about 0.28 mm or more at 45 degrees (or a shallower
belt) the brim band is wider than one bead and its lines go on the nominal
lattice. A lattice line could land where the belt is almost at the band's
print_z; its flow was clamped to half a layer while the nozzle sat nearly on
the belt. Such a line now moves uphill to the 0.75 fraction the single-line
case uses, and a line that lands on the previous one is skipped.
Ported from the Unlayered fork (patch 0007 of its belt port series, found
there by fuzzing first layer heights). The fork's companion fix, restricting
the brim filament to those the writer was handed (0008), is not needed here:
ToolOrdering registers the brim filament on every band's layer, so the writer
always has it. A test pins that with every object a flush target.
EXCLUDE_OBJECT_DEFINE keeps plate coordinates on a belt printer: the frame
after the slicing rotation is undone and before the G-code axis remap and
machine-frame shear, which is where the object stands on the belt.
Raised in Hanif Koh's review of #14394.
The purge tower is a model object the GUI creates and sizes, and libslic3r
only purges into one that exists. A multi-filament belt project sliced from
the CLI without it changed filament with nowhere to purge, silently.
Raised in Hanif Koh's review of #14394.
The cooling buffer's band pass rebuilt positions from the layer's G-code
and tested them against the first-layer plane. The G-code is in machine
coordinates and the plane is in slicing coordinates, so on the shipped
profiles the nearest move was over 100 mm from a 0.2 mm band and the pass
never changed the fan. GCode::_extrude() already knows each path's height
above the belt, so it now tags the band changes and the buffer applies and
strips the tags.
The pass also took the S of every M106 as the part fan, whatever its P
index, and stored that 0..255 value where a percentage was expected (an
auxiliary fan line came back as M106 S651); it now uses FanMover's parser,
which ignores other fans, and converts to percent. It no longer overwrites
the layer's intended speed, only the fan's actual state.
Raised in Hanif Koh's review of #14394.
belt_brim_instances_compatible() runs while the slicing parameters can be
stale, like the rest of the brim predicates, which read the print config.
Raised in Hanif Koh's review of #14394.
Apron bands looked up their filament and nozzle config slot with a running
counter, while object layers use Layer::id(), so band N read the map of
object layer N. They precede layer 0 and now use its assignment.
Raised in Hanif Koh's review of #14394.
The mesh transform is a rotation and an axis permutation, so its
determinant is always 1; rebuilding the forward transform on every
extrusion to divide the flow by it changed nothing.
Raised in Hanif Koh's review of #14394.
enable_prime_tower stays on for any multi-filament project, but a belt
printer never prints the classic tower and the belt purge prism is an
ordinary object that never takes a brim, so every brim on a multi-filament
belt print was refused for nothing.
Raised in Hanif Koh's review of #14394.
validate() skipped the build-volume height check whenever the machine-frame
transform was active, which is every shipped belt profile, so a 400 mm
object passed on a 300 mm printable_height. The transform only changes how
the height is written to G-code; the clearance check from f682ab5cd3
applies regardless.
Raised in Hanif Koh's review of #14394.
Drops the [BELT-DEBUG], [BELTRACE], [BELT-CALIB] and [BELT-PREVIEW] log
lines, the SLIC3R_BELT_DIAGNOSTIC_LOG blocks, and the counters and
temporaries that existed only to feed them. Six of the purge tower lines
logged at warning level, which is Orca's default, on every plan. Raised in
Hanif Koh's review of #14394.
preslice_remap_*, preslice_remap_global and gcode_remap_* describe the
printer's kinematics and are set once by its profile. A wrong value sends
the gantry outside the machine (a user preset with the pre-slice remap in
place of the G-code remap emitted gantry moves to Y=646 mm), so they are no
longer offered in Expert mode.
On a belt printer a lift is a move along the belt axis (0.4 mm / sin 45 =
0.57 mm of belt travel out and back on every hop), not a lift away from the
part. The three belt printer bases now ship z_hop 0, the IR3 V2 leaf no
longer restates 0.4, and the BabyBelt Pro and IR3 V2 filaments stop
overriding the printer with filament_z_hop 0.4. The option stays editable.
PartPlate's containment tests treat the plate as open along Y on a belt printer with
belt_printer_infinite_y, so a long part is no longer flagged outside the plate in Prepare
while the slicer and the G-code checks accept it. The check reads the printer preset
through the app object, which does not exist headlessly, so it is guarded on the plater.
The prism's generator already sets no_brim; PrintObject::has_belt_brim() now also ignores
any brim setting on the prism (belt_purge_tower_object), so a brim on the parts beside it
never blocks purging.
ToolOrdering::has_wipe_tower() reads the first layer's flag. On a belt the first layer may
be a brim apron band, which carries neither object nor support and never gets the flag, so
with a brim the purge plan returned early and nothing was purged. Scan the layers for a
change.
With top_shell_layers = 0 the `top` vector is never filled and erasing its begin() was
undefined (found by fuzzing on a painted object dropped below the plate).
The per-filament island tour was cached by island centroids only. A later layer with the
same centroids but fewer islands (thin walls, negative volumes) reused the stale visit
list, whose catch-all index pointed past the layer's islands, and extrude_perimeters read
freed memory (three fuzz crashes, planar and belt). The per-instance island layout is part
of the cache key and the use site never indexes past the islands.
Found by fuzzing the headless slicer:
- A BBS-style 3MF without Metadata/project_settings.config segfaulted the CLI silently on
the missing printable_height option.
- A project saved with fewer filaments (or filament groups) than --load-filaments overran
the filament variant tables (segfault in the variant match) and then hit an uncaught
ConfigurationError from set_with_restore_2 (std::terminate). The tables are regenerated
for the filaments the project did not know about, the destination vectors grown first
(only from a non-empty source), the match bounded, and a failure becomes a CLI config
error.
A 3MF vertex with a nan/inf coordinate was accepted by both parsers and crashed qhull in
ModelVolume's convex hull while the file was still loading. Both vertex handlers refuse it,
and volume generation checks again whichever parser produced the geometry. The main
parser's _stop_object_xml_parser keeps a message a handler already set.
filament_minimal_purge_on_wipe_tower became variant-keyed upstream, which
invalidated the comment claiming none of full_config's keys were.
WipeTower2::extract_wipe_volumes indexes it by raw filament slot. Not reachable
with shipped presets; noted rather than worked around.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
ComboBox::SetLabel is overridden and writes the text control when the text
control is shown, which it is on an editable combo. Clearing the label after a
pick therefore erased the name. Removed; the label is never written on this
control.
Suggestions are also rebuilt when the list opens. Built in add_row they filtered
against only the rows that already existed, so a row was offered names the rows
below it had taken.
Drops the 1px inset on the Primary label, which matched a frame the name field
no longer has.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
138 commits. The count is large because texture displacement merged with its
whole branch history behind it, going back to July, alongside config and preset
file locking across instances, a foundation for configurable printer agent
connections, and a day of smaller fixes and CI work.
Two conflicts, both the same shape: each side had appended to a sorted list and
git could not choose an order. libslic3r's CMakeLists gained InstanceLock
alongside our IMEXHelpers and IMEXZones, and the preset bundle loading test
gained an include for ParallelResolve alongside ours for IMEXHelpers. Both sides
kept, alphabetical. No logic conflicted.
Verified: 740 targets build clean under -Werror, and the Release suite passes
1740 of 1740, up from 1665 -- the 75 new cases arrived with the merge and all
pass. That mattered more than usual here, since preset loading and config
locking are both areas the IMEX preset code touches.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The name field becomes an editable ComboBox, the same pattern the sidebar uses
for parameters like sparse infill anchor length: predefined entries in a
drop-down, with the text still typeable. The mode table is authored for whatever
hardware the user has, so a closed list would be wrong -- nothing in the slicer
reads a mode's name except as the key a plate stores -- but the conventional
names are worth offering rather than leaving everyone to retype them.
What is offered follows the tool grid rather than a fixed list. Four carriages
get iq-copy and iq-mirror; multicolor needs a Span partner beside the primary
and a second gantry to copy the pair onto, so mc-copy and mc-mirror appear only
on a grid that can hold one. imex_resolve_routing() already refuses a multicolor
mode with no Span on the primary's gantry, and suggesting a name it would then
reject is worse than not suggesting it. Names a row already uses are dropped, so
the list only ever offers what is still free.
Two things about ComboBox matter when it is editable, which nothing else in the
tree does -- the other 78 call sites all pass wxCB_READONLY:
GetValue() returns the drop-down selection whenever there is one, so a name
typed after picking a suggestion would be silently discarded. Every read goes
through GetTextCtrl() instead. Field.cpp reconciles the same way for its own
open enums.
The constructor hands its value to TextInput as the LABEL -- the small
right-aligned slot a unit like "mm" occupies -- because a read-only combo hides
the text control and shows the label in its place. Left there, the name rendered
as a greyed echo beside the hint while the field itself sat empty. The combo is
built empty and the value written to the text control, and the selection handler
clears the label again afterwards, since SetSelection() writes there too.
The mode column widens to 176 to leave room for the drop-down arrow, with the
header spacer deriving from the same constant.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Felix14-v2's review of OrcaSlicer#16019 found the tab behaving as a world of its own.
- Icons: the design_* glyphs were drawn in a fixed light grey, made for the dark ribbon, and the
toolbar re-tinted some of them by rebuilding the bitmap from a wxImage, which drops the HiDPI
scale factor Orca sets on Windows: at 150 % the icons came out half again too large for buttons
that were sized in raw pixels, overlapping and clipped. The glyphs now use Orca's sidebar icon
grey (#949494), which the icon cache maps per theme, nothing is re-tinted, toolbar glyphs drawn
for Prepare's light toolbar use their "_dark" twin, and every size is in DIP.
- Theme: the chrome colours were read once, at construction, and nothing in the tab answered a
theme switch, so switching left light surfaces and unreadable text in a dark tab and the other
way round. The colours are now {light, dark} token pairs; MainFrame::on_sys_color_changed
reaches DesignPanel::on_sys_color_changed, which moves every token colour onto the other
theme's, runs the app's dark pass and re-rasterises the icons. Card borders are StateColors,
resolved at paint time.
- Scale: MainFrame::on_dpi_changed reaches DesignPanel::msw_rescale, which re-rasterises every
icon (buttons, flyout rows, card headers, the tree's image list, now sized from its bitmaps)
and re-measures the Orca widgets.
- Mouse: the canvas no longer forces middle-drag to orbit and right-drag to pan; it reads the
drag actions in Preferences > Control like Prepare. Left-drag is shared with picking, so the
whole-body rubber band takes Shift+left-drag while left-drag is given to the camera.
- FPS counter: ImGui's display size is shared and only refreshed when a canvas sees its own
size change; the Design canvas now re-announces its size when the tab is entered (and the
editor canvas when it is left), as Plater does between Prepare and Preview.
- Viewport text: the status line and the tool readout were top-level popups over GL. A popup
does not follow its frame, so the empty-canvas hint floated over other applications, and the
readout was never taken down with the tab. Both are drawn by the canvas in the tool's ImGui
pass now, with the theme's overlay style.
- Dialogs: messages use MessageDialog/RichMessageDialog; Add/Edit Variable is one Orca dialog
with Name and Expression fields instead of two native text prompts; the Text dialog uses
TextInput, ComboBox and CheckBox (its height is a TextInput: SpinInput is integer-only), and
enumerates the installed fonts once per session. The ribbon's Confirm/Cancel, the reference
pick buttons and the expression buttons are Orca Buttons; the variable actions are icon
buttons like the other cards'.
- Undo: the tab's own Undo/Redo buttons are gone. The top bar's Undo/Redo drive the Design
history while the tab is shown, greyed to what an undo would do, as Ctrl+Z and Edit already
did.
- The first build of the tab logs how long each phase took ("Design tab build: ..."): it is
under a second here but was reported at about fifteen on Windows.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
* Lock Config and Preset Files Across Instances and Write Them Atomically
Every running instance shares one OrcaSlicer.conf and one user preset
tree, and nothing kept their writers apart. Two instances saving at the
same moment, or the cloud preset sync thread writing while the GUI thread
saved, could interleave, and a reader in another instance could open a
preset JSON or .info file between truncate and close and get a partial
file, dropping that preset for the session with a parse error.
Add InstanceLock, a scoped guard that serialises the threads of one
process through a recursive mutex and other processes through an advisory
OS file lock: flock on POSIX, held on the guard's own descriptor so no
other close in the process can drop it, and LockFileEx on Windows. The
outermost guard opens the lock file and closes it on release, so nothing
stays open between saves and a data dir can be removed once nothing is
saving into it; the file itself is kept, since deleting it would let a
third instance lock a fresh file while the second still holds the old
one. It is best effort: when the lock file cannot be opened or locked, or
another instance still holds it after a second, the guard logs once and
lets the write proceed, then leaves the file alone for ten seconds, so a
hung instance never blocks every other one and a holder stuck in a
debugger does not cost a stall per save. The guard sits at the leaf
readers and writers: set_sync_info_and_save() calls save_info() under the
preset collection mutex, so a batch lock around save_user_presets() would
invert the order against the sync thread. The user preset scan reads its
files on worker threads without the guard, since the mutex would
serialise them, and takes it per file in the serial commit step, so a
save never waits for the whole scan. Each read keeps the bytes of the
preset and its .info as they were before parsing; commit compares them
with the disk under the guard and reads a file that changed again, so it
never deletes or writes back over another instance's newer save, nor
installs a .json and .info from two different saves; a preset another
instance removed in the meantime is not installed. Without the guard, in
a cool-down, the scan still loads the presets but leaves their files
alone: an unreadable file stays for the next scan, and a derived
compatible printer is not written back. Read-only scans, which is what
the CLI does, take no lock and create no lock file.
AppConfig holds OrcaSlicer.conf.lock in load() and save(); load is
included because the Windows path restores from the .bak copy. Every
user preset writer and reader holds user.lock: Preset::save(), which
writes no .info when the preset itself could not be written, since an
.info without its preset reads as a cloud deletion request, save_info(),
reload() and remove_files(), each preset the scan commits, the
bundle metadata reads and write, the .info removal after a cloud-confirmed
delete, the orphaned-.info scan on the sync thread, the bundle folder
removal on unsubscribe and the physical printer writers and delete
paths. A bundle import extracts under cache/ into a folder per process
and per import, where no scan reads.
Preset JSON, .info, bundle metadata, physical printer and config files,
and the caches and state files that already used a temporary by hand,
now go through write_file_atomically(), which writes <file>.<pid>.<n>.tmp
beside the target and renames it over, so a reader that never waits sees
a complete old or new file. A symlink is followed; a target that is not
a regular file is written in place; and when no temporary can be created
beside an existing target, or the rename itself is refused, by a Windows
reader holding the file open or a mount that cannot replace in one step,
the helper writes in place as before, since losing the save is worse
than a torn read. On POSIX the rename replaces the
target atomically where the old code removed it first and left a window
with no file at all; only a mount that refuses a one-step replace gets
the old remove-then-rename. A crash between temporary and rename leaves
the temporary behind, which no scan reads. Preset::save() returns
whether it wrote the preset, so the scan counts a compatible printer it
could not write back as an error. A failed config write keeps
the config dirty, and the idle handler waits ten seconds before retrying
while an explicit save always tries.
* Run the Cross-Process Lock Test on Every Platform
The test that checks the guard yields to a lock held elsewhere forked a
child to hold it, so it was left out on Windows. The OS lock belongs to
the handle on Windows and to the open file description elsewhere, so a
second handle in the same process is refused like another instance
would be. The test now holds the lock that way and runs everywhere.
extrude_infill() and extrude_support() reversed the layer's extrusion
entities in place while chaining them, so each export started from the
previous one's reversed toolpaths, and each copy of an object from the
copy before it. The export-time region lists now hold const pointers,
and chaining reverses a clone instead.
The display-name tests call set_user_session(), which persists the session.
The agent they built was in keychain mode, so every run saved a fake
OrcaSlicer/Auth session into the system keychain of whoever ran the tests,
replacing their real Orca Cloud login on any desktop with a working keychain.
Move them next to the other agent tests and build the agent the same way:
encrypted-file mode with a throwaway config directory.
An ironing line spacing of 0 reached the fillers from a 3MF, the CLI or
the per-filament override, which has no GUI guard. Concentric ironing
then never finished slicing, because a zero inset never shrinks the
region, and rectilinear ironing was silently dropped. Tiny positive
values produced an unprintable number of lines.
Top surface and support ironing now clamp the spacing to the 0.05 mm
floor the process GUI guard already enforces, so these configurations
iron at that spacing. Spacings at or above the floor, including every
shipped profile, are unchanged. The concentric filler also returns early
on a non-positive step so no other caller can hang it, and the filament
settings page now resets a too-small override the same way the process
page does.
GCodeProcessorResult declares a copy assignment, which suppresses the implicit
move assignment, so `*result = std::move(processor.extract_result())` binds to
the copy and duplicates the whole moves array: 96 bytes per move, measured at
147.7 MiB in one allocation for a 597k facet model at 0.08 mm, with the source
staying alive until the export returns.
Add the move assignment. It assigns exactly the same 39 members in the same
order as the copy, including the ones the copy deliberately leaves alone, so a
target that carries its own filament maps and nozzle type keeps them. A move
constructor is intentionally not added: the type holds a std::mutex, so it is
neither copy nor move constructible today and a partial one would leave ten
members uninitialised.
reset() now releases the two print sized vectors rather than clearing them. Its
callers are the paths that discard a result, so the memory went back only at the
next slice.
ViewerImpl::reset() does the same for the two vectors sized by the print, which
is what keeps a discarded preview resident, and counts m_vertices_colors in the
reported CPU memory, where it was missing.
simplify_path, simplify_multi_path and simplify_loop each copied the whole
PrintConfig, 10520 bytes with around 200 heap owning members, once per extrusion
entity, from inside the parallel_for over all layers. LayerRegion.cpp already
binds a reference a few hundred lines above. Three more sites in Layer.cpp do the
same thing.
The tree support changes in the same area: draw_circles() dropped layers from an
owning vector without deleting them, so every re-slice leaked the layers above
the topmost overhang. std::stable_partition rather than std::remove_if, because
remove_if leaves moved-from duplicates of kept pointers in the tail and deleting
those would be a double free. The collision and avoidance accessors returned
ExPolygons by value out of a cache that hands them out by reference, several
times per node per layer. The preview cache is released when the support step
finishes instead of at the start of the next slice; nothing reads it in between.
set_shared_object() now releases the layers the object still owns before it
starts aliasing another object's. Both clear functions are no-ops once the alias
is set, so this cannot free the sharee's layers.
PrintObjectSlice.cpp: eps was a function local static inside a parallel_for, so
slice_closing_radius was latched once per process. This one changes output for
projects that override it per object, or on a re-slice after editing it. The old
behaviour was undefined: whichever object won the race set it for the session.
Stores a painted cube, points its component back at the object that holds it and
expects the load to fail. Without the bound the test does not finish: the work
list grows until the process is killed.
# Description
With Xcode 27, building deps on macOS 26 fails at the Python install
step with a segfault, and a libpython built with Xcode 27 crashes on
macOS 12–26 the first time anything calls `os.pipe()`, which every
plugin `subprocess` call does. The macOS 27 SDK declares `pipe2()` and
`dup3()` as macOS 27-only, and CPython 3.12 calls them without a runtime
check once configure finds them, so on older systems they resolve to
NULL. This keeps CPython on the `pipe()`/`dup2()` fallbacks it already
uses with older SDKs; upstream fixed it in 3.13+
([python/cpython#153711](https://github.com/python/cpython/issues/153711))
but not in 3.12.
No change for builds with Xcode 26 or older, or on Linux and Windows.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
Rebuilt deps with Xcode 27 on macOS 26.6: the Python install step now
completes, the installed libpython no longer imports `pipe2`/`dup3`, and
CPython's `test_os`, `test_subprocess` and `test_posix` pass, apart from
one test that needs `_testcapi`, which `--disable-test-modules` leaves
out. Running CPython's configure against the macOS 26.5 and 27.0 SDKs
gives a byte-identical `pyconfig.h` for 26.5 with and without this
change, and for 27.0 with it. The x86_64 cross-build path was configured
on arm64 to confirm both of its configure runs pick up the change.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
_generate_current_object_list expands component references through a work list
with no bound. An object whose component points back at itself, or a pair that
point at each other, makes the list grow until the process runs out of memory:
a few hundred bytes of XML take the slicer past 20 GB of resident size.
Bound the expansion by the number of objects in the file. A reference chain
longer than that has to revisit an object, so this rejects every cycle and no
acyclic file, however deeply nested. A second bound on the number of expanded
components stops an acyclic graph that fans out exponentially.
The plugin test fixtures start the interpreter before the test points
data_dir at its temporary directory, so PythonInterpreter creates
{data_dir}/python/packages and {data_dir}/log with an empty data_dir: a
python/ and log/ folder in whatever directory the tests run from. When that
is the test binary's folder, the next run's embedded-interpreter tests took
the stray python/ as their home and failed to start Python.
Give each fixture that initializes the plugin manager its own temporary
data directory, set up before initialize(), and only use the python/ folder
next to the test binary as the interpreter's home when it holds a standard
library.
With stealth mode off the home page asks for the login status every 2 s,
and while nobody is logged in that ends in clear_user_secret(), which
opened the system keychain and deleted the OrcaSlicer/Auth entry on the
UI thread every tick. A working keychain cost a D-Bus round trip per tick;
a keychain that never answers blocked the UI for 25 s per tick. It also
deleted a login another running instance had just saved, and ignored
use_encrypted_token_file, so opting out of the keychain did not help.
Remember whether this process read a secret from the store or wrote one,
and only then touch the store when a logged-out poll asks for a logout.
A logged-out instance never touches the keychain, and in encrypted-file
mode the poll no longer opens the keychain at all. A secret this process
cannot read, such as a token file encrypted for another OS user sharing
the data directory, is left alone by the poll. An explicit logout still
wipes both backends, so a token stranded by switching the token storage
option cannot sign the account back in later.
# Description
Include variant-aware validation, temperature and pressure controls,
named nozzle selection, and filament sidebar refresh.
## Support nozzle-specific filament cooling and tuning
Filament presets can require different cooling and tuning for Standard
and High Flow nozzle variants. This change makes part and auxiliary
cooling, pressure advance, temperature limits, and multitool ramming
settings follow the selected variant, with matching UI controls and
profile validation.
The Snapmaker U1 profiles in #15755 / Snorca illustrate why this
matters:
| Profile | Standard cooling | High Flow cooling |
|---|---|---|
| Snapmaker ABS | Part fan: 15–15% | Part fan: 10–60% |
| Snapmaker PETG HF, 0.4 mm | Part fan: 20–40% | Part fan: 30–60% |
| Snapmaker PETG-CF | Part fan: 0–20%; auxiliary: 0% | Part fan: 5–40%;
auxiliary: 20% |
| Snapmaker PLA Matte | Auxiliary: 80% | Auxiliary: 100% |
These differences need to survive variant selection, editing, and
slicing. Shared single values continue to apply across variants, and
short filament arrays fall back to element zero, matching the loader.
### Additional changes
- Preserve named nozzle selections when refreshing the diameter
selector.
- Enable or disable the pressure-advance input for the selected variant.
- Initialize profile-validation slicing with the printer’s declared
nozzle volume type.
- Refresh filament controls after preset loading.
Profile changes remain separate in `u1-hf`.
### Validation
- All 13 focused variant-tool tests passed.
- Broader checks encountered an existing Windows path-separator
assertion failure, reproduced with upstream code.
- C++ build and GUI validation have not been run.
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
## Summary
Adds a new paint-style **Texture Displacement** gizmo that stamps
grayscale
height-map textures onto a model's surface and turns them into real
relief -
engraved or embossed detail - either as a live preview or baked into
actual mesh
geometry.
You paint where a texture applies, stack up to **8 blended texture
layers**
(image-editor semantics: Add / Subtract / Multiply / Divide), choose how
each is
projected onto the surface (Triplanar / Cylindrical / Spherical / LSCM
unwrap /
From-view), and - for the LSCM projection - lay the charts out by hand
in a new
dockable 2D **UV Editor** pane. Coarse models can be **Subdivided** or
**Remeshed** first so there are enough vertices to carry fine detail,
and the
result is committed with **Bake**, restricted to the painted area only.
The tool only ever affects the **painted** region; everything left
unpainted
keeps its original surface, and bake blends the relief seamlessly into
it with no
remeshing or hole-filling at the seam.
---
## User-facing features
- **Paint the affected area** with Brush (circle/sphere), single-Face,
or
Connected-area flood fill; or **Select whole model** - reusing the
existing
`TriangleSelector` / `FacetsAnnotation` painting machinery, one full
mask per
layer slot.
- **Up to 8 texture layers**, each with its own paint mask, texture, and
parameters, combined in slot order like image layers.
- **Per-layer controls:** Depth, Tile size, Rotation, Midlevel
(bidirectional
emboss/engrave), Smoothing, Edge smoothing, Invert, Blend mode, Tile
(Repeat / Mirrored-repeat / decal), and Projection.
- **Projection methods:** Triplanar (blended, seam-free across sharp
edges),
Cylindrical, Spherical, **Unwrap (LSCM)** - a real CGAL conformal
parameterization - and **From view** (slide-projector decals).
- **UV Editor pane** for LSCM layers: move / rotate / scale / snap / cut
/ join
islands, average texel density, Checker and Distortion overlays, manual
mark-
seam workflow, live model update while dragging.
- **View modes:** Normal (true displaced geometry = what Bake produces),
Fast
(GPU bump-shaded approximation of the active layer), Checker, Distortion
heatmap, and an independent Wireframe toggle.
- **Mesh prep:** Subdivide (1–5* uniform 1->4 split) with cyan-wireframe
preview,
and CGAL isotropic **Remesh** to a target edge length.
- **Texture library:** 10 shipped seamless 512×512 grayscale height
maps, plus
import of any PNG/JPG/BMP (converted to an 8-bit grayscale map and
copied into
the user data dir so app updates can't clobber it).
- **Bake** runs in the background (off the UI thread); the preview is
free to
explore and only Bake changes the real mesh.
---
## How it works (implementation)
- **Bake is accumulate-then-displace and topology-preserving.** The
output mesh
has exactly the input's vertices and triangles in the same order; only
displaced vertex positions differ. Each layer is evaluated against the
**base
mesh** and folded per-vertex into a shared accumulator via its blend
mode, then
every touched vertex moves once along its precomputed undisplaced
normal. This
replaced an earlier sequential re-mesh-per-layer design that was the
root cause
of the "second layer never applies" bug and made blend modes impossible.
- **Boundary vertices are pinned.** Any vertex shared with an unpainted
triangle
is never displaced, which is what keeps bakes seamless with zero
remeshing.
- **LSCM unwrap** uses a new `MeshBoolean::cgal::parameterize_lscm()`
built on
CGAL's already-vendored `Surface_mesh_parameterization` package - **no
new
external dependency**.
- **Fast GPU preview** perturbs the shading normal from the height
gradient
(analytic mm-per-mm slope for triplanar; Mikkelsen's
screen-space-derivative
method for the conformal LSCM path), so apparent depth matches the bake.
- **Background jobs:** preview compute and bake both run off the UI
thread on the
shared job worker (queued, not `replace_job`, so a preview never cancels
an
in-flight bake); a generation counter discards stale results.
- **UV Editor** shares the app's single real `wxGLContext` and reuses
the
registered `flat`/`flat_texture` shaders; island drags update one affine
matrix
per island rather than re-uploading geometry.
---
## Backward compatibility & constraints
- **Feature is fully gated behind the new gizmo** - it adds no new
default
behavior and does not touch existing slicing, profiles, or defaults.
Models
that never open the tool are unaffected.
- **Cross-platform** - pure `libslic3r` / `libslic3r_gui` /
`libslic3r_cgal`
code; no new dependency and no `deps/` rebuild. (Built and tested on
Windows;
no platform-specific APIs introduced.)
- **`.3mf` compatibility:** **baked** relief round-trips fine, since it
becomes
ordinary mesh geometry via the existing serialization path. **Unbaked**
paint
masks and layer definitions are **not yet serialized** - see
Limitations. No
existing project data is affected.
---
## Testing
Unit tests in `tests/libslic3r/test_texture_displacement.cpp` - **run
and
passing** (7 cases, 116 assertions). Coverage:
- `decode_height_texture` round-trip
- Empty-layer no-op
- Full-cube uniform displacement
- Boundary-vertex pinning on a hand-built fan mesh
- Regression: a **second layer over the same area actually contributes**
(the
bug the bake rewrite fixed)
- Table-driven check of all four blend modes
- The lowest painted layer ignoring its blend mode
`BUILD_TESTS` is `OFF` in the checked-in cache; enable to run:
```bash
cmake -S . -B build -DBUILD_TESTS=ON
cmake --build build --config Release --target libslic3r_tests -- -m
./build/tests/libslic3r/Release/libslic3r_tests.exe "[TextureDisplacement]" --order rand
```
---
Building deps with Xcode 27 on macOS 26 failed at the Python install step,
and a libpython built with Xcode 27 segfaulted on macOS 12-26 whenever a
plugin started a subprocess.
Community work from Rob Niccum, adding 0.04mm Ultra Fine and 0.08mm Extra Fine
for the 0.4 nozzle, a toolchange prime on the three IQEX machines, and
recompressed cover art and build plate model that are byte-identical in content.
Bundle version bumped to 02.04.00.13 here rather than in his branch, so the two
of us were not editing the same line while his review was open. Without the bump
the updater refuses the bundle outright, and nothing in CI catches that.
Verified rather than assumed, since his own validator predates the parallel
printing code and the fine tiers changed shape during review: the full profile
check passes all five stages tree-wide, and both new tiers slice clean on all
four 0.4 machines, 8 of 8. His last revision dropped a bottom_shell_thickness
override that was making those tiers thinner against our common process rather
than thicker against his; the emitted G-code now carries the inherited 1.0,
which is what that fix was for.
Co-Authored-By: Rob Niccum <klober81@users.noreply.github.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Eleven commits, including a typed-config G-code export speedup, a printer agent
refactor that generalizes the infrastructure beyond Bambu, gyroid optimization,
and three CLI crash fixes.
One conflict, in PrintConfig.cpp. Upstream introduced a new_def macro and began
migrating the placeholder table onto it, adding curr_bed_type that way in the
same block where this branch had added imex_mode, imex_mode_index and
imex_mode_gcode in the older def = this->add(...) form. Both sides are kept and
ours are converted to the macro, which expands to the same three statements and
wraps label and tooltip in L() exactly as before, so nothing changes about what
is registered or what is translatable.
Note for anyone building this branch: the agent refactor adds a dependency,
LibDataChannel, so the deps tree needs dep_DataChannel built before the app will
configure. A distribution package of the same name will be found first if one is
installed, and the resulting error names a missing RelWithDebInfo location
rather than the wrong package, so point LibDataChannel_DIR at the dependency
prefix if that happens.
Verified: 789 targets build clean under -Werror, and the Release suite passes
1665 of 1665, up from 1630 before the merge -- the 35 new cases arrived with it
and all pass.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Addresses the interface notes on the IDEX/IQEX modes editor.
Add Mode moves from below the rows to the top of the panel, beside a "?" button
that now carries the overview text as its tooltip. At the bottom the button
shifted down the page every time a mode was added, so where it sat depended on
how many modes already existed. It is an Orca Button in the Confirm style, width
matched to the mode column it creates a row in, and the panel opens on the
legend rather than on a paragraph.
Remove moves out of the right-hand column, where it sat one icon away from
Edit -- a destructive control beside the one pressed most -- to under the name
field it deletes, and its icon becomes a boxed minus rather than an X, which
read as "close". Reset joins Edit in the right-hand column, which is now
top-aligned so the icons hold position regardless of row height. Tool tiles are
square at 24px, and the header spacer tracks that width so the column titles
stay over their columns when the grid changes shape.
The two text fields were landing on GTK's near-black default border, invisible
against the panel: measured 45,45,49 against a 43,43,43 background, where the
settings fields above use 74,74,81. wxTextCtrl cannot color its own border, so
each sits in a one pixel frame taking the color TextInput derives for the
theme, and carries wxBORDER_NONE so Windows and macOS do not draw a native edge
inside it. The G-code boxes also take the monospace face EditGCodeDialog uses.
Bed zone fills drop to roughly half opacity in the Standard theme. They cover
whole quadrants for a whole session, so at swatch saturation they dominate the
scene. The collision strip is dimmed less, since it marks where a head hits
something. The deuteranopia, tritanopia and high contrast themes keep their
alphas: those are chosen for discriminability, which is the opposite trade.
Also fixes the icon size never applying. All four ScalableButton call sites
passed eight arguments, so the size bound to use_default_disabled_bitmap and
bmp_px_cnt kept its default of 16. Both are passed now.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Fan speeds, the recommended nozzle temperature range, minimal purge and
multi-tool ramming flow now take one value per extruder variant. Each
single value is repeated for every variant the filament lists, so every
preset loads exactly what it did before.
set_filament_pressure_advance's declaration said -1 omits the tool qualifier,
without qualification. That holds on Klipper, Marlin and BBL but not on
RepRapFirmware, where no-tool keeps the historical `D0`: a bare M572 applies to
whichever tool is selected and errors when none is, so omitting the qualifier
would make pressure advance depend on tool-selection state for every RRF user,
none of whom are using IMEX. The behavior is deliberate and unchanged; only the
declaration overstated it. GCodeWriter::set_pressure_advance and the index-space
notes in IMEXHelpers.hpp already described it correctly.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
In a parallel mode no tool changes occur, so two loops address each active
carriage explicitly: pressure advance before the print, and the second-layer
drop off the initial-layer temperature. Neither had coverage, and both fail
silently -- a carriage missing from one emits nothing at all, so it holds the
initial-layer temperature for the whole job, or runs on whatever pressure
advance the firmware was last given. The new case pins that the carriages
addressed are exactly the ones the mode declares active, each with the values of
the slot physical_extruder_map routes its head to.
The test needs filament_self_index, set here on imex_7x4_printer() so the whole
file has it. Production authors that key 1..n; its all-1s default collapses
every per-filament vector to filament 1's value through get_config_index_base(),
which leaves a per-slot assertion comparing a value against itself.
Also bounds the second-layer loop on filament_diameter alone. The bound belongs
in slot space, and filament_diameter is the one per-filament vector never
expanded per variant; the previous min() against nozzle_temperature mixed the
two index spaces without changing the result. The pressure advance loop already
bounds this way, so the two now read alike.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The lane map fetched over Moonraker was applied to whichever printer preset
happened to be edited when the response arrived, with no check that it was the
machine polled. Not a race, as the whole chain runs on the GUI thread with the UI
blocked, but a steady-state mismatch: two IMEX printers of different models with
the same logical extruder count both pass every existing guard, and the map is
written to the wrong preset and dirties it with no user action.
The pairing is now evaluated when the callback runs, against the selected device,
on the predicate update_sync_status() uses. Capturing an identity at connect time
would instead ask whether the edited preset had changed since then, and would
reject the user who selects a machine and only then switches to its matching
preset.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A plate can carry its own parallel mode, or leave it on Primary and inherit the
process preset's. set_imex_mode() erases the key on Primary precisely so the
preset's value survives the config merge, and the slicer honors it: Print::validate()
and the G-code path both read the merged object config, so such a plate slices, and
emits, in the preset's mode.
Three places that describe that plate did not resolve it the same way. They read
the plate's own value and stopped, so on a plate left at Primary:
the multi-material conflict badge stayed dark on a plate validate() will refuse,
which is the one invariant the comment above it claims to keep;
the bed-temperature and filament-type warnings never ran, so the job went out in
copy or mirror with a mismatched bed and no notice;
the plate tooltip reported Primary for a plate about to print in another mode.
The geometric badge beside the first of those already resolved correctly, because
it goes through the zone layout, so one badge fired while its neighbour stayed
dark on the same plate.
The fallback had been written out by hand four times. Three are now collapsed onto
one accessor, get_effective_imex_mode(); the fourth is in libslic3r, which cannot
call a GUI method and resolves the two modes itself from arguments. The remaining
callers of the raw get_imex_mode() want the raw value and keep it: the accessor
itself, the zone layout call that passes both modes separately, the mode menu and
the left-click cycle, which act on what the plate stores, and the reset that looks
for plates whose own mode was removed.
The tooltip is the one place the two readings meet. It headlines the stored mode,
because it sits on the button whose menu and click act on that value, and names
the inherited mode after it when they differ -- so one control no longer says
three different things while still telling the user what will actually print.
Reachability, honestly: nothing in the UI writes the process preset's copy today,
so this needs a hand-edited preset, a vendor process profile or a project that
carries one. It is latent rather than live -- and it stops being latent the moment
a process-level mode selector exists. Note the key lives in the process preset, so
one value there would govern every Primary plate in every project using it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The default 0.4 mm z-hop is a 0.57 mm move along the belt axis and its
return tripped the back-step check. Shipped belt profiles print without a
z-hop, so the test does too.
check_multi_extruder_gcode_valid() compares each object's max Z with
printable_height. On a belt printer machine Z is belt travel (a 3DBenchy
on the BabyBelt Pro runs from Z=197 to Z=309 on a 69 mm printable_height),
so every belt export set the over-height error bit and the CLI refused the
plate with -102 "G-code in unprintable area". The preview already skips its
ToolHeightOutside warning for the same reason; the export check now does
too. The XY printable-area check is unchanged.
The header tags lost their belt_ prefix in the Part 3.2 rename (20
characters now), but the parser still skipped 25, so every axis read as
pos_x. Found in Hanif Koh's review of #14394.
A scarf joint begins one layer height below the current layer and ramps
up along the wall. On a tilted belt that start is a step backwards along
the belt axis, into the previous layer's wall at the seam: 0.283 mm per
0.2 mm layer at 45 degrees. With an aligned seam the nozzle rams the same
spot on every layer. A BabyBelt Pro benchy with seam_slope_type=external
showed 601 such back-steps from layer 107 on, and in the field the belt
"jumped backwards" and the head knocked the part loose.
Belt printers now skip the scarf in GCode::extrude_loop, and the process
tab greys the scarf controls out for them, as it already does for arc
fitting. The regression test slices a cube on a belt with the scarf
enabled and checks the belt axis never steps back by a layer pitch.
Brings in upstream/belt-printer (the Sept 14 main merge) plus Hanif Koh's
21 review-fix commits from PR #15685, on top of the MachineKinematics
refactor and the purge-prism / tree-support / first-layer-speed fixes.
Conflict resolution:
- BeltGCodeWriter is gone (kinematics refactor), so Hanif's plate-offset
fix for it is ported into GCodeWriter: the first-layer-plane checks in
travel_to_xy / travel_to_xyz / _travel_to_z now evaluate the plate-local
point, and BeltGCode::init_belt_writer hands the stored plate origin to
the writer it installs.
- init_belt_writer(Print&) takes Hanif's signature; the BBL flag is set on
the surviving writer by GCode::_do_export.
- The shared emit_belt_brim_bands() loop keeps the BeltFloorObjectGuard the
local branch added, so apron bands classify first-layer height against
their own object.
- eager_lift keeps effective_type: it now carries set_force_normal_lift().
- GCodeWriter's initializer list follows Hanif's member order with
m_kinematics in its declared position.
- TreeSupport::detect_overhangs uses Hanif's clamped build_plate_tilt_slope()
for the non-belt path and the belt shear for the belt path.
Printers without a named nozzle variant no longer switch profiles when their
current diameter is picked in the sidebar. The filament tab reads the selected
variant through one helper.
Fan speeds, multi-tool ramming, the tower interface and flush temperature
fallbacks and the custom G-code placeholders now use the extruder variant a
filament prints with on each layer, instead of reading by filament id.
main moved the CAD docs into docs/HLSD/design-tab.md and the offer
generator and atlas into scripts/CAD/ (#15803).
- The four docs this branch had edited are deleted as on main. What the
branch changes about the design goes into the HLSD doc: the right-click
is judged by drift over the whole press, with no time budget; a Text
feature stores its outlines as well as its string, font and height; a
new section on how the Design canvas renders bodies (studio lighting
through the shared phong shader, B-rep edge ribbons).
- gen_offer_table.py keeps both main's atlas validation and this
branch's L() markers on user-facing strings.
- The Text verb's hint is now changed in scripts/CAD/tool_atlas.json as
well, so `gen_offer_table.py --check` passes.
- DesignPanel.cpp keeps the DesignTextDialog include and main's new
path in the offer comment.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
* Add developer flag for printer agents
* Parse user print info on the UI thread to prevent heap corruption (#119)
get_user_print_info()'s HTTP fetch can run on a worker thread (e.g. BindJob),
but parse_user_print_info() mutates userMachineList (insert/erase/delete
MachineObject). on_machine_alive (SSDP) mutates the same maps on the UI thread
without locking, so parsing off-thread races the map and frees MachineObjects
out from under it -> heap corruption.
Keep all device-list mutation on the UI thread: parse inline when already on
the main thread, otherwise marshal via CallAfter so it stays serialized with
on_machine_alive.
* Prevent loss of user access code on LAN reselect
Keep user access code intact to maintain access rights even if
device slot is unpopulated, ensuring continuous connection
and status message reception.
* Harden send flow and separate upload failure recovery (#111)
* fix(send): harden FT send path + IP pre-flight UX
* Remove early returns
* Working Moonraker and Qidi printer agent transport (#104)
Folds the Qidi AMS box-mapping print
overrides (apply_box_mapping +
start_* wrappers) that the transport
fix builds on.
* Add support for runtime error status in plugins
Distinguish a loaded plugin whose
capability errored (RuntimeError,
warn-styled, stays checked) from a
load-time Error. Status now derives
via resolve_plugin_status(); enum
ordinal keeps dialog sort priority.
Unloading clears stale errors.
* Resolve duplicate agent ID conflicts
Reject a printer-agent capability
whose agent ID is already owned by
another capability or built-in:
flag the plugin error, disable the
capability, and warn the user
instead of silently ignoring it.
* fix: checkbox should depend on plugin is_loaded status
* Replace fake-enum printer agent dropdown (#121)
A dedicated PrinterAgentChoice field
reads rows straight from the live
agent registry and stores the agent
id string, replacing the fake-coEnum
index mapping. The field moves to
TabPrinter and registers with the
searcher so UnsavedChanges renders
it; the PhysicalPrinterDialog copy
and its update hook are removed
(#125). switch_printer_agent now
resolves ids via
resolve_printer_agent_id.
* Reset device selection on agent swap or unload (#124)
set_live_printer_agent centralizes
the swap: deselect the machine,
clear stale sidebar state and the
previous agent's Other Devices, then
install the new agent (or null when
its provider vanished). Plugin
load/unload callbacks refresh the
dropdown and re-run agent selection.
load_last_machine no longer falls
back to the first available machine.
* Gate agent mode behind use_printer_agents toggle
Replace per-printer auto-activation
(is_current_printer_agent_plugin)
with a global experimental AppConfig
toggle, default off: legacy
print-host behavior is unchanged
until the user opts in. The toggle
drives device-tab routing, print
button defaults, connect-button
visibility and sidebar layout, and
dedups machine-select dialog opens.
* Track BBLPrinterAgentPlugin.py
* Add printer-agent and plugin status tests
Ports the agent lifecycle, duplicate
agent-id, built-in-id clash and
status-resolution tests. The loader
runs on a detached worker thread, so
the lifecycle tests live in their own
executable. Tests install the
production unload-side registry
wiring themselves (no GUI in the
test binary) and register agents
manually so concurrent loads stay
deterministic.
* fix: pin HTTP to prevent connection refusal
Set `use_ssl` to false to ensure Moonraker
connectivity, as the service uses HTTP rather
than HTTPS, preventing connection issues. Initialize
device info early for reliable name resolution.
* Bring Moonraker device panel to feature parity
The monitor panel showed wrong or missing
data for Moonraker printers, and its
controls did nothing.
Push payload now carries layer number and
total layers. Remaining time replaces the
wrong total_duration - print_duration
formula. The chamber light toggle maps to
Klipper SET_PIN / SET_LED, and pause,
resume and stop post to
/printer/print/{action}. Task thumbnails
resolve via /server/files/thumbnails onto
a new MachineObject thumbnail url.
Filament sync switches to pull mode so the
agent is queried on demand.
Not compiled or run.
* Stop blocking print on unreported nozzle data
* fix: make Klipper macro lamp control reliable
* Surface Moonraker webcams and gate unrunnable controls
* Keep Bambu AMS dialect out of the agent waist
M620 is Bambu firmware dialect, not a
neutral command. Composing it in
MachineObject let non-Bambu agents
(Moonraker/Klipper) forward it and
report success on firmware that
cannot run it.
Agents now own the dialect: the
default refusal on IPrinterAgent
returns not-supported so the UI
can say so; BBLPrinterAgent keeps
the byte-identical composition.
* Fix multi-color filament logic
Reuse color decoding across functions to improve
code readability and maintain consistency in
multi-color filament handling.
* Move Moonraker commands off the UI thread
Pause/resume/stop, g-code sends, temps, and
light ran synchronous HTTP on the UI thread,
freezing the app up to 10s per click on slow
or unreachable printers.
Run them on a single agent-owned FIFO worker
so g-code ordering is preserved, while command
translation stays synchronous so unsupported-
command dialogs still work.
Add a pending-disabled state to the pause,
resume, and abort buttons for Moonraker-family
printers: the icon only flips once the
WebSocket reports the real state, which also
rules out double-click races.
* Show Snapmaker U1 camera in Device tab
The U1 exposes no /server/webcams/list entry;
its camera only captures after an explicit
camera.start_monitor RPC, which the Moonraker
websocket executes unauthenticated but only
answers over MQTT - so the call is fire and
forget.
Start the camera when the camera view is
shown and renew every 300 s: the printer
retires the capture task at ~362 s and
stop_monitor is accepted but ineffective,
so teardown is simply to stop renewing.
Frames land in monitor.jpg as still JPEGs
(~2 fps at interval 0), so the webview loads
a local HTML wrapper that repolls with a
cache buster.
* fix: start stream when camera URL changes
* fix: stop Qidi slot parse throwing on null
* Keep printer-agent progress in sync
Keep the shared task progress aligned with agent
reports that lack Bambu cloud task identity.
Release the lazily allocated task during reset to
avoid leaks when machine objects reconnect.
* docs: document the printer-agent subsystem
* fix: merge access codes into one
* Reconcile implementation split with PR tip
* feat: abstract remaining gcode commands in devicemanager
* refactor: abstract bambu specific protocol to printer agent
* refactor: push bbl workflows to bbl printer agent
* remove unused
* fix: default impl
* fix callback error
* fix: remove redundant cache
* specify api for getting file transfer url
* revert file transfer abstraction
* fix: ams filament mapping workflow
* feat: update qidi to use subscription based filament sync mode
* fix: resolve stubgen byte header conflict
* fix: ams sync info and periodic ams sync via subscription workflow
* fix: skip filament sync dialog if filamentSyncMode is none
* feat: parse nozzle information for qidi and moonraker printer agents
* fix: extend access code requirements t 0, 8 or more characters.
* remove irrelevant docs
* fix: remove heavy includes from IPrinterAgent
* fix: defer filesystem and camera abstractions
* fix: remote do_fetch_filament_info from tests
* cleanup moonraker and snapmaker printer agents
* fix: access codes regression
* fix: tests
* fix: printer agent switching on preset change
* fix: remove unused variable
* fix: snapmaker U1 SelectMachineDialog blocking print
* fix: merge artifact
* fix: clear up some unrelated changes
* feat: connect to cloud printer and monitor
* feat: connect to cloud printer and monitor
* feat: generic camera stream support for http snapshot and rtsp
* fix: build & access code UI
* feat: generic camera stream support for http snapshot and rtsp
* fix: build & access code UI
* feat: connect to cloud printer and monitor
* feat: connect to cloud printer and monitor
* fix: build errors
* feat: camera via webrtc
* fix: build
* fix: cmake
* feat: remove frame assembler and change config to set protocol
* fix: orcaprinteragent refactor
* fix: LAN paths and camera stream
* feat: use ffmpeg to render http camera stream
* fix: make model_id/dev_type optional instead of blocking
* fix: connect via ip dialog
* feat: LAN impl for Orca Printer Agent
* fix: model_id resolution method for non bambu printers
* fix: ffmpeg http camera stream jittering due to incomplete frames
* fix: revert sdcard check
* feat: check printer storage status before sending
* fix: moonraker printer agent hang on printer power cut
* fix: shim layer for any compatibiliity changes
* fix: cloud printers were using the wrong MQTT endpoint
* feat: cloud download via HTTP
* temp: doc for intended change
* fix: warnings
* fix: warnings
* fix: camera auto-play on startup
* fix: split infra from impl
* fix: uninitialized ams state blocking print
* Fixes nullptr deref
* Log first before std::move
* fix: printer agent virutal optional functions
* fix: parameterize orcaslicer_copy_test_dlls() for printer_agent_plugin_tests
* Revert "fix: parameterize orcaslicer_copy_test_dlls() for printer_agent_plugin_tests"
This reverts commit 2f566e3779.
* Guard libdatachannel. Remove unused code
* fix: unit tests & unused variables
* fix(ci): deps build order for datachannel
* Resolve printer agent first before getting cloud printer agent
* fix(ci): set depends openssl
* fix: re-include apply header guarded by ifdef __APPLE__
* fix(ci): add libdatachannel to flatpak manifest
* fix: add internal_developer_mode chekc back to MediaPlayCtrl::load()
* fix: invoke js clearInterval on WebMediaController::stop
* fix: change rtc log level
* fix: inject provider, agent id and generation to get_user_print_info to ensure correct metadata
* fix: revert moonraker specific behavior
* fix: remove stale comment
* fix: use ORCA_CLOUD_PROVIDER instead of hardcoded string
* fix: remove hardcoded ICE servers
* feat: enable https camera stream mode
* fix: move non-mandatory printer agent function stubs to IPrinterAgent
* fix: dedupe compatible printer type check
* fix: stop the correct media controller
* fix: scope get_my_machine_list to printers listed under the current printer agent
* fix: move printer agent plugin tests into test_plugin_lifecycle.cpp
* fix: always build bundled DataChannel dep
* fix: disable unused DataChannel media support
* revert: filament sync work
* fix: wrap command_* with small wrapper
* fix: regression bug, connecting to bambu needs bblp username
* fix: default impl for vendor agnostic gcode commansd
* refactor: media controller playback routing and ownership
* fix: bump libdatachannel ver & update flatpak to use tar instead
* fix: stop flatpak DataChannel build from re-cloning over the sandboxed network
* fix: update windows ffmpeg prebuild
* fix: update printer agent plugin API
* fix: shift camera signaling channel to network agent
* fix: follow external-packages for flatpak libdatachannel deps & add flatpak path to use source_dir
* feat: add printer-agent.md doc to HLSD
* fix: guard DeviceManager command dispatch when no printer agent is bound
* fix: port BBL implementations from #15711
* refactor: connect_printer api and dialog
* fix(tests): make omitted printer agent operations answer like a missing agent
* fix: preserve printer agent defaults in PrinterAgentPluginCapabilityTrampoline
* test: cover printer agent default command dispatch
* fix: validate windows FFmpeg avformat library
* feat: extend optional printer model warnings to calibration & ams workflows
* fix: handle malformed printer progress values safely
* fix: clear webview document on stop
* chore: reduce diagnostic logging level to trace
* refactor: centralize printer compatibility checks
* tests: add device manager integration coverage
* tests: cover WebMediaController lifecycle with wxWebView stub
* fix: make integration tests headless
* refactor: collapse command_ams_refresh_rfid and command_ams_refresh_rfid2
* refactor: make printer connection SSL agent-specific
* fix: persist input printer host and port
* fix: use correct device id for Moonraker connections
* fix: make moonraker gcode commands asynchronous
* fix: preserve moonraker device names
* fix: add include for non BBL_RELEASE_TO_PUBLIC path in BBLPrinterAgent
---------
Co-authored-by: Andrew <159703254+andrewsoonqn@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
Co-authored-by: Lam Wei Lun <weilun.lam@gmail.com>
When a 3mf is tagged as written by OrcaSlicer or BambuStudio, the CLI
treats it as a project and reads the printer settings stored in the
file. #14580 guarded four of those reads, but printable_height still
called opt_float() without a check, so a 3mf whose
project_settings.config is empty or missing crashed with a null
dereference. The handy models OrcaBadge.3mf and OrcaSliced.3mf are both
like this.
Skip the read when the option is absent, like the reads around it. The
bed-size logic further down already treats an old printable height of 0
as unknown and uses the current printer's.
OrcaSlicer#16019's Windows unit-test jobs failed on one test, "Hole
standards: ... 82° countersink": CTest passes the test name to Catch on
the command line, the ° arrives in the ANSI code page, the filter
matches no test and the run counts as failed ("No test cases matched",
"No tests ran"). The name now says "82 degree"; it was the only test
name in the suite outside ASCII.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
Expand shared filament and process values to their declared variant widths while preserving existing tuning. Bump the Snapmaker bundle version to 02.04.00.22.
Validation: Snapmaker profile check passed with zero errors and warnings; verified all expanded entries repeat their original values.
Co-authored-by: Codex <codex@openai.com>
Use the shared unsigned variant index for adaptive pressure advance and volumetric speed controls, resolving missing identifiers and the ambiguous string accessor.
Co-authored-by: Codex <codex@openai.com>
Refinement now tapers off away from the paint, and the unpainted surface is kept out of
simplification, so what one bake leaves unpainted is still the model's own triangles when a later
bake paints there. A second bake used to refine the slivers and long triangles the first one left
around its stroke, and came out many times denser, with walls off the texture's lines.
Once the budget is met the progress fraction stops moving, and a cancel was only checked when it
moved, so Cancel did nothing until the flat-face merging finished. It is now also checked every 16k
steps.
Meeting the triangle budget by merging flat faces alone no longer raises the warning, and the
warning now quotes the budget instead of the triangle count left after the flat faces were merged.
A vertex pinned on flat ground could collect a fan of thousands of slivers, and validating each
collapse next to it walked the whole fan, turning a second of simplification into minutes. Collapses
that would leave more than 64 faces around a vertex are now skipped.
The Windows builds of OrcaSlicer#16019 failed in DesignSketchTool.cpp:
windows.h defines near (and far) as empty macros, so
`Vec2d near = m_op_anchor;` reads as `Vec2d = m_op_anchor;` under
clang-cl. The variable is now called closest.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
* Fill Settings Missing From a CLI Project From Its System Presets
A project saved before a printer or process option existed has no value
for it. The GUI takes such keys from the project's system preset; the
CLI left them at the option default, so e.g. extruder_clearance_dist_to_rod
sliced as 40 instead of the P1S's 33.
The CLI now resolves the project's system printer and process presets by
name and copies the keys the project lacks, skipping preset bookkeeping,
print-host keys, the extruder variant layout and keys the legacy handler
drops. PresetBundle::resolve_system_preset finds the vendor through its
manifest or preset cache, so it also works in release builds, which ship
vendors as caches only.
* Load a CLI Project's Printer and Process Settings as the GUI Does
The CLI filled only the keys a project lacked from its system preset.
The GUI builds a project preset differently: the project's values go
over the default preset, without the print-host keys, and every key
the project does not list in different_settings_to_system is refreshed
to its base system preset's current value. After a profile update the
two sliced the same project differently.
That step now lives in Preset::load_external_config, which takes plain
configs and gets the base preset from a callback, so the collection
lookup stays in PresetCollection. PresetBundle::project_different_keys
builds the kept-key set from a project's escaped entry, adding the
preset bookkeeping keys, and is used by both the GUI and the CLI.
PresetCollection::load_external_preset calls the shared step with no
change in behaviour.
The CLI now builds the project's printer and process configs with the
same step, passing the system preset from resolve_system_preset. That
drops the hand-kept skip list for print-host and variant-layout keys
and the legacy-key check: the shared step already excludes print-host
keys and maps per-variant values onto the base preset's variant layout.
The --uptodate path and a printer or process given on the command line
are left as they were.
Because the GUI's kept-key set always holds the bookkeeping keys, the
refresh runs for every project that names a system preset, so the CLI
now loads that preset's vendor on every such run.
Reported: the Revolve axis could only be the sketch plane's X or Y axis
through its origin, so a half-profile drawn beside a centerline, the
usual way, could not be revolved about that centerline.
- CadFeature::revolve_axis_entity names a Line of the profile sketch to
revolve about (a construction centerline, or an edge of the profile
itself); -1 keeps revolve_axis. Appended at the end of the framed
recipe, so existing projects load and rebuild unchanged. Revolve and
Surface Revolve resolve their axis in one place (revolve_axis_of); an
index that no longer names a line fails with a reason.
- SketchEngine::make_revolve takes the world axis. A profile with points
on both sides of it is refused with "the profile crosses the revolve
axis"; MakeRevol failed there with no reason.
- The Axis list of both cards reads Plane X, Plane Y, then every line of
the sketch, named as the constraint list names them (Centerline E4,
Line E3). A fresh revolve preselects the sketch's centerline when it
has exactly one. The gizmo turns about the chosen axis and draws it
dashed; construction lines no longer pull its centre.
Tests: a rectangle beside a construction centerline revolves into the
tube of the expected volume along that line; about its own edge, into a
cylinder; an axis through it is refused with the reason; a stale axis
index fails with a reason; the axis survives save and load. The
truncated-recipe test accounts for the new tail field.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
Reported: once extruded, a part is hard to read; the lighting says
little about its shape. Both lights of the object shaders sit near the
camera, so the sides of a part come out in almost the same tone, and
nothing marks where one face ends and the next begins.
- The phong shader gains a studio lighting model, chosen by a new
lighting_model uniform: a sky/ground hemisphere in world space (up
faces cool and bright, down faces warm and dark), a key light from the
upper left and a weak fill from the right, a plastic-like highlight,
and a darker base with a faint sheen toward the silhouette so curved
faces read as round. GLCanvas3D::set_studio_lighting() makes a canvas
draw its objects with it whatever the realistic-view preferences; only
the Design canvas turns it on. Every canvas sets the uniform on each
use, 0 for the slicer's, so they render as before.
- Every B-rep edge of a body is drawn as a thin dark line over it, depth
tested and pulled a few pixels toward the eye so it wins against the
faces meeting at it and hides behind the faces in front. The seam of a
closed surface and degenerate edges are left out
(GeometryEngine::display_edges). Edges are sampled once per shape and
kept across recomputes that leave a body unchanged; bodies faded by
body focus get fainter edges, and a dress-up previewing its result
alone hides them with the bodies.
Tests: display_edges gives a box its 12 edges at their lengths, a
cylinder its two round rims without the seam, a cone its base rim
without the seam or the apex.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
The context menu still said the text outline is added to the sketch as
editable lines, which is what it did before Text became its own
feature. Only SVG goes into the open sketch now; the comments on that
path say so.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
Reported on the rig: Text showed no preview where it would go, its
dialog was pinned over the middle of the window (GNOME attaches a modal
dialog to its parent and it cannot be moved), and once confirmed the
text could not be edited and did not appear in the feature tree. Inside
an open sketch it became loose lines of that sketch.
- Text is always a feature, "Text N" in the tree. A new text goes on the
plane of the open sketch (committed first when it holds anything,
closed when it is empty), else centred on the picked face, else on the
reference plane.
- The dialog is modeless and opens at the top right of the window. The
feature is created at the first character and redrawn on every change,
so the text appears in the view where it will be as it is typed. Enter
inserts it, then the usual move/scale gizmo and Confirm; Esc, Cancel or
closing the dialog takes it out again (undo to the checkpoint taken
when it appeared).
- CadFeature keeps text_string, text_font (the WxFontUtils descriptor)
and text_height, appended at the end of the framed recipe. Editing a
Text feature reopens the dialog with them and redraws the outline in
place, keeping its placement. The outlines are still saved, so the
project opens the same on a machine without that font.
- The MCP control refuses writes while the Text dialog is open.
Tests: the text parameters survive a save and load along with the
outline; the truncated-recipe test accounts for the new tail fields.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
Offset worked on the single entity under the pointer. On an outline made
of many short entities (a text glyph, an imported shape) that is a
segment a few tenths of a millimetre long, and the starting distance was
a tenth of that: the ghost was too small to see, and a typed distance
moved one invisible segment. It read as "Offset does nothing, no
preview".
- The pick takes the chain the entity belongs to (connected_loop, same
construction state), highlights all of it, and offsets it as one
outline through offset_entities, which already joins and trims chains
at their seams.
- The starting distance is a twentieth of the outline's size, so the
preview is visible at once.
- The arrow reads its side off the ghost, since the engine may walk the
picked entity backwards in the chain; it keeps pointing at the offset
copy, flipped for a negative distance.
- A single entity still gets its Parallel/Concentric constraint. A chain's
offset is placed as geometry, since its joined entities no longer map
one to one onto the originals.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
A sketch drawn on the rig extruded to walls with no caps. Every joint of
its loop met, so the loop analysis called it closed and MakeFace
accepted it, but the loop crossed itself: an arc left the top line's end
heading back over it and crossed it again 2.5 mm on. A second arc left a
0.28 mm line tangent to it but the other way, a cusp. The prism of that
face is an invalid solid, and it was shipped as a body.
- SketchEngine::wires_to_face checks the face it builds and, when OCCT
calls it invalid, fails with "the profile crosses or folds back on
itself, so it does not bound one region". The extrude reports that
instead of producing the broken body.
- sketch_loop_defect() judges a closed loop of lines and arcs exactly:
any contact between two of its entities away from the joints they
share, or a joint where the curve turns straight back (a cusp; OCCT
still builds that one, but it is never what was meant). It returns the
point.
- The sketch uses it on every region: the loop is tinted red, the point
gets a marker, and the status line says what the red means the first
time one appears. The MCP loop report lists the defects and no longer
calls such a profile buildable.
Tests: the rig's profile, with each defect and with both, from a
recording of the real entities. The analysis names the cusp at its joint
and the crossing on the top line, in either traversal order, and
passes ordinary tangent and collinear joints. The extrude refuses every
crossing variant with the reason, and the same arcs swept the other way
round extrude to a valid solid.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
A solid edge could only be picked one at a time, and Fillet/Chamfer took
either that one edge or a whole face group. Rounding three chosen edges
meant three features, whose edge ids each resolve against a body the
previous one had already changed.
- Shift+click (or Ctrl+click) on an edge of the body already picked adds
it to the selection, or removes it; the same modifiers that extend a
sketch selection. The whole set is highlighted. A plain click replaces
it, as before.
- Fillet/Chamfer dresses every picked edge in ONE feature at one size,
all ids resolved against the same body. The card says "3 edges", the
status line and the offer header name the count.
- CadFeature gains dressup_edges, appended at the end of the framed
recipe, so existing projects load and rebuild unchanged. dressup_edge
keeps the first edge, so an older build opening a newer project still
dresses that edge instead of falling back to the face group.
- The MCP fillet/chamfer verbs take `edge` as one id or an array.
Tests: a fillet on the four picked top edges equals the Top face group
exactly; the list survives save/load; two opposite chamfers remove
exactly twice one; a missing id fails with a reason. The truncated-
recipe test accounts for the new tail field.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
Drawing a line or a rounded rectangle crashed the Linux AppImage: the
constraint list labels were formatted from narrow literals holding "—",
"·" and "°". wx converts a bare char* with the current locale, and the
AppImage's AppRun sets LC_ALL=C, so the conversion failed, the format
string came out NULL and wxString::Format dereferenced it
(wxFormatConverterBase<wchar_t>::Convert, from constraint_label via
rebuild_constraint_list). A build started under a UTF-8 locale never
showed it.
These literals, and the "…" of the interference report, now go through
wxString::FromUTF8, as the panel's other non-ASCII literals already do.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
- Text opens a dialog instead of a bare text entry: any installed font
(bold, italic), the height in mm, and a live outline of exactly what
will be inserted with its size. Enter inserts, Esc cancels; the last
font and height are remembered. text_to_regions gains an overload
taking a loaded font.
- The offer menu opens with a greyed title naming what the rows act on
("Flat face 4 of Body 2", "Sketch line", "Nothing selected").
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
- Offer table: user-facing strings carry the L() marker so xgettext
extracts them; DesignOffer.hpp, DesignSketchTool.cpp and
SketchInlineEditor.cpp are listed in localization/i18n/list.txt.
- Offer: model-mode Constrain sits in the same row as the sketch one;
Interference is wired; Rib shows its R key; a verb that accepts the
selection but is blocked by the document stays greyed with its reason
instead of vanishing from the submenu; one refusal wording per verb.
- Extrude infers Join when the profile touches a solid (new
CadDocument::body_touching_sketch) and on face push/pull; New body in
free space. Revolve/Sweep/Loft/Boolean use the same result words.
- Interference and volume/area reports go to the status line in mm3/mm2
instead of modal dialogs; Delete Body no longer asks (it is undoable).
- New feature names match the card header ("Extrude 3"), translated;
"Coordinate system", "Angle (°)", center/color spelling, translated
face and length readouts, slot hints say width.
- CAD gizmos in Prepare are selectable only with the CAD feature on; the
sketch auto-close setting is stored per design.
- Docs: confirm/cancel rules, enabling the feature and MCP in
design_tab.md; drift-only right-click in interaction-model.md; the
portability note rewritten to describe the integration as it is.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
Keyboard and mouse
- Esc drops what is pending (picks, a dimension's first point, an edit-op or transform) and
never applies it; Enter applies a ready edit-op/transform, ends a polyline/spline chain,
ends an armed tool, and confirms a feature card exactly when its ✓ is enabled.
- Right-click only abandons the gesture in progress; with nothing pending it opens the offer
in every tool (Trim, edit-ops, transforms, Dimension, Constrain, TransformArt, move gizmo).
Clicking empty space no longer commits. The offer needs no timing, only a still press.
- Delete removes only an explicit selection. Undo/redo inside a sketch go through the same
route as the buttons (whole shapes, with redo); Edit > Undo follows the shown tab.
- The canvas no longer handles Delete/Esc/Ctrl+Z itself (Backspace in a value field deleted
the geometry it measured); F is in the panel's key map.
- Value fields: a refused value keeps the field open with the reason; click outside and Tab
commit; an untouched field commits the exact value; any decimal separator is accepted;
lengths are always mm; validation is the same for every editor.
- Snapping: the marker shows only where the click will actually snap; pick tolerances are
one set of pixel budgets (Constrain picks within reach; no mm floor on labels).
Messages and consistency
- set_status(kind, text) gives every status line its own colour and glyph; kernel errors are
translated into sentences and formatted, not concatenated; sketch refusals go to the status
line instead of the per-frame HUD that erased them.
- Hints describe the gestures that now work; Dimension shows its second step; Constrain uses
the sketch palette (red means conflict only); the straight slot's value is its width.
- Hole/Thread/Project keep the user's pick or refuse up front; circular pattern opens with its
own preview; thread fields use the nominal diameter and the ISO internal depth.
Integration
- MCP loads the project's recipe before touching the document, refuses to mutate it while the
tab is busy, never runs a request that already timed out, only replaces a socket at its
path, caps line length and removes the socket at exit; not started in the G-code viewer.
- Hiding a feature keeps later body references on their bodies; a design keeps its modeling
origin across printer changes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
Sketch layer
- Trim/extend of an arc (and a circle opened into an arc) rewrites p0/p1 from the new angles;
the wire builder, the solver and snapping read them.
- Tangency binds the arc end that touches the line (or the other arc) instead of always the
start, so a fillet tangent to both legs solves; circle-circle / circle-arc tangency uses the
centre distance instead of CURVE_CURVE_TANGENT, which aborts on a circle.
- Point-on-line distances and circle-line tangency keep the side the geometry is on; a point
on an arc's rim uses PT_ON_CIRCLE.
- The partitioned solve keeps constraints onto the origin/axes and counts free entities' DOF.
- Zero-radius circles get no solver primitive and build no wire; constraints the solver cannot
apply are reported in SketchSolveResult::skipped.
- EllipseArc: mirror no longer yields the complement; after a solve its angles and ends are
re-derived; on the XZ plane it is no longer built mirrored.
- Offset: a circle follows the "+d = left of travel" rule (it shrinks, like a CCW arc chain);
chains are joined at the weld tolerance. Bridge end pole fixed (G1, no cusp). Negative-scale
transforms keep arcs and ellipses on their ends. Inference tolerances aligned with the weld.
Model layer
- Body references follow the body across delete / reorder / hide (resolved by the feature
that made it); datum-plane ordinals are re-pointed; an index past the end is an error, not
"the last body". New set_feature_enabled(). A move that puts a consumer above its input is
refused.
- Threads made from now on read thread_radius as the nominal major radius (internal: bore to
minor, groove to major; external: groove cut into the rod); older recipes build as before.
Bad thread parameters say why. Circular patterns span their angle end to end (new ones);
add_pattern pivots on the modeling origin.
- clear() drops variables; expression fields the GUI offers are bindable (thread_diameter,
helix_*, thicken_thickness, ...); deg()/rad() in expressions; the recipe saves the modeling
origin and body colours; names/colours follow bodies by identity.
- Boolean and dress-up failures throw instead of returning the input; one produces_body();
hole standards corrected (82° inch countersinks, UNC names, #10-24); legacy profile solve
validates indices and writes back only on success; v4 recipes read with a frozen field list.
Tests cover each of the above.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QK4VgguuCAk2hZLWgcjJb9
Twelve defects found by the 2D design pass.
- ONE erase path repairs the dimensions' cached constraint indices. Removing a
constraint renumbered m_constraints and left every later DimAnnot.con pointing
one slot short, and set_dimension_value checked only the range, not the slot's
identity — so editing a dimension's value could overwrite an unrelated
constraint. All five mutators (badge delete, fillet/chamfer, Move/Rotate/Scale,
the two drop paths) now route through erase_constraints().
- apply_dimension validates BEFORE moving or recording. A value outside a case's
threshold moved nothing and then recorded the constraint anyway, handing the
solver a number it could never satisfy; the socket guarded against this, the
tool did not.
- A Distance and its zero case are one dimension slot: typing 0 and then 5 used
to leave a Coincident AND a Distance on the same operands.
- A dimension whose constraint the solver rejected is named: the label renders in
the refusal colour and the commit says the sketch is over-constrained, instead
of showing a number the geometry does not have.
- Six silent refusals now speak: fillet/chamfer on a non-corner and on an
overrunning radius, offset on an ellipse or spline (the kernel's own reason),
a rejected array binding ladder, a dimension pick on an unsupported entity.
- set_tool commits a ready transform instead of dropping it, the rule the ready
edit-op already followed.
- Constraint releases are reported, and roles_of is one function again (the two
copies had already diverged on EllipseArc).
- Labels no longer collide: every label is its own centred ImGui window at an
anchor whose offsets are multiples of the text height, so on a feature smaller
than one text height a line's Length and Angle labels landed on the same spot.
draw_text, the one function all of them pass through, now pushes a colliding
label clear of the ones already drawn this frame.
Verified: build and LTO link on behemoth (exit 0, new binary); on the rig,
badge-delete took constraints 3->2 with dof 9->10 and the geometry untouched;
the draw-then-edit chain committed a typed 70 to exactly 70.0 mm; and a 5.4 mm
selected line renders "5.4 mm" and "21.8°" as separate readable labels.
NOT yet exercised: the stale-index corruption itself (needs a middle delete with
a labelled dimension after it), the poison-value guard through the field, the
transform commit, and the red refusal label.
The design pass over the rest of the 2D tool suite: Select/Measure,
Dimensioning, Constraints, Modification (Trim/Extend/Split/Offset/Mirror),
Dress-up (Fillet/Chamfer) and Transform, each with its FSM, its event routing,
its C++20 blueprint and the defects it exposed.
Thirty-two findings, every one verified against the tree with file:line, plus
nineteen open scope calls and a closing note that orders the six cross-cutting
changes by what unblocks what. Batches 1-8 were delivered in-session; only
their cross-cutting results (six archetypes, two defect classes) survive here.
Merged by /bot merge on behalf of @peachismomo (id 52488812).
Grants: resources/profiles/OrcaFilamentLibrary/filament/Elegoo, resources/profiles/OrcaFilamentLibrary.json, resources/profiles/Elegoo, resources/profiles/Elegoo.json
Head: ff6a2f2056
Since the CLI can open an OpenGL context (#15745) it renders plate
thumbnails on export, and the viewport restore at the end of
render_thumbnail_internal (#15674) then reads the plater through the wx
application. The CLI has neither, so every --export-3mf on a machine
with a display died with a segmentation fault after the first
thumbnail. Skip the restore when there is no application or no plater;
the GUI path is unchanged.
Include variant-aware validation, temperature and pressure controls, named nozzle selection, and filament sidebar refresh.
Co-authored-by: Codex <codex@openai.com>
Upstream moved pressure advance onto the extruder variant: enable_pressure_advance,
pressure_advance and the four adaptive keys joined filament_options_with_variant,
the repeated inline blocks in set_extruder() became a helper, and the lookups
moved from the filament id to get_filament_config_index().
All three conflicts were the same collision, because this branch had modified two
of those same inline blocks to pass a tool qualifier so each carriage is addressed
explicitly in parallel modes. Taking either side whole would have lost something:
upstream's drops the qualifier and leaves parallel carriages with no pressure
advance, ours drops the per-variant indexing and reads the wrong column on a
multi-variant printer. The helper now takes an optional tool, defaulting to -1,
which omits the qualifier. imex_pem_tool_for() already returns -1 off IMEX and in
primary mode, so non-IMEX output is unchanged, and the three call sites that never
passed a tool keep upstream's behavior exactly.
The third conflict was two test cases appended at the same place. Both are kept.
Separately, one defect that merged cleanly and so was not flagged: the loop that
emits pressure advance for secondary carriages at the start of a print still
bounded and indexed those vectors with a raw filament id. They are variant
expanded now, so their length is columns rather than filament slots -- the value
read was the wrong column, and the bound no longer sat in slot space, letting an
out-of-slot filament through. It now bounds on filament_diameter and translates
with get_filament_config_index(), which is what the sibling second-layer
temperature loop already does.
Verified: both changed translation units compile clean under -Werror. The merge
was resolved independently twice and the two resolutions agree on every line of
code. Not yet run: the Release test suite and a parallel-mode slice sweep.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Every extruder on every model carried the same yellow. The key exists so the
plater, the preview and the filament mapping can tell toolheads apart, which
matters most on the machine this bundle exists to demonstrate: four independent
heads that were indistinguishable at a glance.
Yellow, blue, plum and orange, assigned in tool order, so the two-tool models
take the first two. Red and green are avoided as a pair because they are the
hardest to separate for the commonest color vision deficiency.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
- IQEX (0.4/0.6/0.8): retract_restart_extra_toolchange 1 -> 0 on all four
tools. With the silicone ooze blockers the parked nozzle stays primed,
so an extra 1 mm after each toolchange can over-extrude at the restart.
- Add 0.04mm Ultra Fine and 0.08mm Extra Fine @Xplorer 0.4 on the existing
fdm_process_xplorer_common ladder for all four 0.4 printers. Same speeds,
accelerations and line widths as the 0.12mm Fine tier; only the layer
height, shell layers, bottom_shell_thickness 0.6 and (0.04) top solid
infill flow differ. Support stays off, as on the other tiers.
- Xplorer_buildplate_model.stl: decimated from 82,508 to 10,000 triangles
(4.1 MB -> 500 KB). Same bounding box and origin, so bed_model is unchanged.
- Covers: one distinct 240x240 image per model instead of the same image
copied four times (39 KB each -> about 11 KB each).
load_obj emits the second triangle of a quad from corners 0, 2 and 3,
but read its texture coordinates from corners 0, 1 and 2, so half of
every textured quad sampled the wrong part of the texture. The corner
indices are now passed down to where the coordinates are read.
A mesh with inward-facing triangles is flipped after loading, which
swaps corners 1 and 2 of every face. The texture coordinates were left
as they were. They are now swapped along with the corners.
* Escape Project Metadata in the Project Page and Restrict Accessory Opening
The Project page rendered the model and profile name, author, description
and accessory file names from the 3MF as live HTML. Names, authors and file
names are now set as text, and the file list is built from DOM nodes with
bound click handlers instead of concatenated markup. Descriptions can
legitimately carry rich-text HTML, so they are rebuilt from an inert
DOMParser document, keeping only plain formatting tags, http(s) links and
http(s) images, with every other attribute dropped.
Opening an accessory from the page now only launches regular files that
lie inside the project's extracted auxiliary directory. The containment
check is a new libslic3r helper, is_absolute_path_within_root, built on
is_path_within_root so symlinks leading out of the root are rejected too.
* Tighten Project Page Description Rendering and Keep More Formatting
Link and image URLs in descriptions must now start with an http or https
scheme as written and parse as such with the URL parser. Preview images are
built as DOM nodes like the file list, and accessory names show their full
text as a tooltip.
Descriptions keep more plain formatting: del, ins, figure, figcaption, dl,
dt, dd, caption, q, abbr, kbd and wbr, plus alt, title, width and height on
images, colspan and rowspan on table cells and start on ordered lists.
Numeric attributes must be plain integers. Embedded YouTube players become
a link to the video.
* Confirm Before Opening Program Attachments and Load Only HTTPS Images
Opening a project attachment whose type runs as a program or script
(executables, installers, shortcuts, shell and PowerShell scripts, macOS
command files and apps, Linux desktop entries) now asks for confirmation
first. The check lives in libslic3r as is_executable_file_name and ignores
the trailing dots and spaces Windows strips from file names.
Images in project descriptions are kept only when they load over https,
so opening the Project tab no longer issues plain-http requests.
* Open Project Attachments Through One Guarded Helper
The Edit Project Info view launched attachments directly, without the
checks the project page has. Both now call
desktop_open_project_attachment, which checks that the file is inside
the auxiliary directory, asks for confirmation where needed and then
opens it.
The auxiliary root was built through encode_path, which returns code
page bytes on Windows, while boost::filesystem reads a narrow string as
UTF-8. With a non-ASCII temporary directory the root never matched and
no attachment opened. It is now built from the UTF-8 path directly.
The list of program extensions could not be kept complete and let
unknown types open without a prompt. It is replaced by
is_safe_to_open_file_name, a list of plain document, image, model and
video types that open directly. Everything else asks first.
* Stop Malformed Network Responses from Crashing the App
Duet, MKS and UltiMaker parsed print host replies with boost read_json
inside the HTTP completion callback with no try, so an HTML or truncated
reply threw out of the Physical Printer Test button and terminated the app,
or killed the upload queue thread. The five identical copies of the parser
(ESP3D's and Flashforge's were unused) are replaced by one shared
PrintHost::get_err_code_from_body that reports a non-JSON reply as an error.
The upload queue now catches a failing job per job, so one bad upload no
longer leaves later jobs queued forever.
Flashforge read material station slots with nlohmann value(), which throws
on off-type fields or non-object entries. The parsing moves into
Flashforge::parse_material_slots, which reads fields leniently with the
existing try_parse_json_int and skips bad entries.
UserManager::parse_json parsed the payload before its try block; the parse
now happens inside it.
* Keep UploadFinished Paired with UploadStarted When an Upload Throws
The exception from a throwing upload was caught around perform_job, so
the UploadFinished lifecycle event was skipped and plugins saw an
upload start that never finished.
The catch now sits around the upload call. The error is reported
through the job's error callback and UploadFinished is fired with an
error code, as for any other failed upload. The worker keeps running
for the next job. The started, upload and finished sequence moved to
PrintHostJobQueue::upload_job so it can be tested without the dialog.
* Confine Updater Archive Extraction to the Target Directory
The preset updater extracted downloaded archives by appending each entry
name to the cache directory, and the network plugin installer did the same
for the plugin folder, without checking that the result stays inside it.
Move the updater's extraction into libslic3r as extract_archive_confined,
which validates every entry with is_path_within_root before writing
anything and fails the whole archive if one entry resolves outside the
target. The plugin installer now rejects such an entry the same way. Well
formed archives extract exactly as before.
* Harden Archive Extraction Against Symlinks
The plugin installer now creates a symlink entry only when its target is
relative and, joined to the link's own directory, passes
is_path_within_root, via the new is_symlink_target_within_root helper.
Before writing any entry it checks the destination with symlink_status, so
an existing symlink, dangling or not, is replaced rather than followed, and
it creates parent directories inside the existing error handling.
extract_archive_confined replaces a symlink at a destination file the same
way.
is_path_within_root now ignores a trailing separator on the root, which
previously made every path fail the check.
* Validate Plugin Symlink Targets Before Replacing Existing Files
A symlink entry's target is now read and checked before anything already
at its destination is removed or renamed aside, so an archive rejected
for its link target leaves the installed plugin files in place.
* Reject Paths with an Embedded NUL When Confining Extraction
is_path_within_root compared each component with "..", so a name such
as "..\0" passed the check. The filesystem calls stop at the NUL and
act on a shorter path than the one that was checked: a symlink target
read from a plugin archive as raw bytes was created as "..", pointing
out of the plugin directory.
A path containing a NUL is now rejected before anything touches the
filesystem, which covers every caller, including entry names taken from
the Unicode Path extra field.
* Tune pressure advance separately for each extruder variant
Pressure advance, adaptive pressure advance and its model can now take a
different value for each extruder variant of a filament, such as Standard and
High Flow nozzles, like the other per-variant filament settings. Projects
saved with one value per filament apply it to every variant of that filament,
and the addnorth BBL filaments in the Orca Filament Library are updated to the
per-variant layout.
* Move Nozzle type to each extruder's settings page
Editing other settings on an Extruder page of a single-extruder
multi-material printer no longer triggers the nozzle diameter prompt.
* Fix command-line slicing when a filament leaves out a per-variant setting
* Percent-Encode Local File URLs for Embedded Web Pages
The Home tab, setup wizard, Project tab and other embedded pages were
loaded from file:// URLs built by pasting the resources path into a
string. A '#', '%' or '?' in the install path was then read as a URL
fragment, escape or query, so the pages failed to load, for example a
portable install under D:\#OneDrive showed a directory listing instead
of the setup wizard.
Add file_url_from_path(), built on wxFileSystem::FileNameToURL, and use
it wherever a local page or image URL is built from a path. Queries such
as ?lang= are appended after the path is encoded. The wizard's printer
cover images are passed to the page as file URLs too.
* Encode the Login Error Page URL and Cover More Windows Path Forms
The login dialog's error page was still loaded from a raw resources path;
it now uses file_url_from_path like the other local pages.
The Windows file URL tests now also cover a resources path joined with a
forward-slash relative path, as the callers build them, and a UNC path.
* Build the Flush Dialog Page URLs with the Shared Helper
WipingDialog and NozzleListTable still called
wxFileSystem::FileNameToURL directly. They now go through
file_url_from_path like every other local page, so the URLs are built
in one place.
Adds a test for a resources directory with a '#' in its name, which the
plugin page check did not recognise before.
Support variant-specific cooling, pressure advance, purge, ramming and temperature ranges. Keep shared U1 values as singletons and verify element-zero fallback for missing variant entries. Preserve preset migration aliases and validate high-flow selections using their declared nozzle volumes.
Co-authored-by: Codex <codex@openai.com>
libslic3r_gui calls webkit_* directly, and OrcaSlicer.cpp and libspnav
call Xlib, but both libraries were only linked when FLATPAK was set.
The default build links because the bundled static wxWidgets lists
them in wx-config. A shared wxWidgets does not, so any build against
one, like the Flatpak build or a distro package, fails with undefined
webkit_* and X* symbols.
Pressure advance, adaptive pressure advance and its model can now take a
different value for each extruder variant of a filament, such as Standard and
High Flow nozzles, like the other per-variant filament settings. Projects
saved with one value per filament apply it to every variant of that filament,
and the addnorth BBL filaments in the Orca Filament Library are updated to the
per-variant layout.
Adds the four Xplorer configurations, their twelve machines and fifteen
processes, so the parallel printing modes can be exercised on a real printer
rather than only on hand-built configurations.
Brings in nine commits, including stricter slice validation of custom G-code and
filename formats across system profiles, OBJ and DRC import hardening, and the
Ender-3 V3 SE extrusion mode fix. None of them touch the files this branch
changes.
The Xplorer ships in four configurations that differ in how many toolheads they
carry and how those heads are arranged: Single, IDEX with two heads on one
gantry, Dual Gantry with one head on each of two, and IQEX with two on each of
two. Each is offered at 0.4, 0.6 and 0.8 mm, giving four machine models, twelve
machines over a shared base, and a five-tier process ladder per nozzle.
The dual-gantry machines park the second gantry's tools over the plate, so the
area both gantries can reach is 57.5 mm shallower than the plate itself. Those
two declare 400 x 342.5 where Single and IDEX declare 400 x 400. That depth is
what the parallel print modes divide into equal zones, so it has to be the
reachable area rather than the physical one.
The bed textures are drawn to match. Texture coordinates are normalised per
axis, so a canvas whose aspect differs from printable_area is stretched and
anything drawn outside the plate is pulled onto it; each viewBox equals its own
model's area. They mark the real reach limits and nothing else, because the
print zones are computed and drawn per plate at run time and a static copy of
them only disagrees with the live one.
Motion limits are the firmware's: 5000 acceleration, 300 mm/s, 100 on Z, 120 on
the extruder, and a square corner velocity of 5 rather than a jerk, which Klipper
does not have. Every per-extruder and per-variant value is written at full width,
because padding a short array repeats its first value rather than its last pair
and would otherwise hand every extruder past the first a normal-mode figure in
its silent slot.
The processes are based on a profile tuned on the hardware. Line widths are
expressed as percentages of the nozzle so one statement serves all three sizes,
and the ladder varies only what belongs to layer height. Extruder variants are
declared rather than left to a default that would have capped volumetric flow at
a figure describing a plain V6.
Verified by slicing all fifteen tiers on all four models: every one completes,
emitted accelerations and widths match what the profiles resolve to, no
coordinate leaves the printable area, and the two pre-existing Troodon models in
the same bundle are unaffected. The bed graphics, the print zones and the
volumetric cap resolve through paths the command line does not exercise and want
confirming in the application. Values tuned against hardware the author does not
have - the 0.8 ladder in particular - are a starting point rather than a result.
Co-Authored-By: Dan_3dp <corexy.diy@gmail.com>
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
They stand on their own: nothing in the texture displacement feature
calls them, and nothing in them knows about textures. Reviewing a
slicer-wide parallelization next to a new gizmo helped neither.
They now live in perf/slicing-optimizations, based on current main.
* Validate OBJ Texture-Coordinate Indices
load_obj read the texture coordinates of a face without checking the
vt index, so a face referencing a vt past the end of the list read out
of bounds and crashed, and a face vertex with no vt read index -1.
Out-of-range or missing indices now fall back to a zero UV. The face
keeps its entry in the per-face UV list, so the following faces stay
aligned, and the geometry loads as before.
Negative (relative) vt indices were also rebased by dividing the float
count by 3, but each vt stores two floats.
* Reject DRC Meshes Without Positions or with Invalid Face Indices
load_drc dereferenced the POSITION attribute without checking that the
mesh has one, and trusted the decoded face indices, which the Draco
decoder does not check against the point count. Both now fail the load
cleanly. A failed vertex conversion is treated the same way.
The libslic3r tests link Draco so they can encode the malformed meshes
in-test.
* Keep OBJ Texture Coordinates That Carry a W Component
The vt parser stopped reading the optional third component when texture
coordinates were cut down to u and v, but the check that nothing is left
on the line stayed. A legal "vt u v w" line was therefore rejected and
silently dropped, shifting every later texture index. The w component is
parsed again and discarded.
The texture coordinate stride is now a named constant, OBJ_TEXCOORD_LENGTH,
used by the parser and the importer, so the relative-index rebase cannot
drift from the storage layout again.
# Description
The Ender-3 V3 SE machine start G-code performs its purge using
absolute-style extrusion positions:
```gcode
G1 ... E15
G1 ... E30
```
However, the machine start G-code does not explicitly initialize the
positioning or extrusion mode before these commands.
OrcaSlicer emits the printer's custom `machine_start_gcode` before its
own generated `G90` / `M82` or `M83` preamble. This means the purge can
inherit the extrusion mode left active by the printer.
For example, if `M83` relative extrusion is still active, such as after
a cancelled print where normal end G-code was not executed:
- `E15` extrudes 15 mm
- `E30` extrudes another 30 mm
Instead of the intended 15 mm followed by another 15 mm.
This change explicitly adds:
```gcode
G90 ;Absolute positioning
M82 ;Absolute extrusion mode
```
before the purge sequence so startup behaviour is deterministic and does
not depend on inherited printer state.
The change is applied consistently to all Ender-3 V3 SE nozzle variants:
- 0.2 mm
- 0.4 mm
- 0.6 mm
- 0.8 mm
No print speeds, temperatures, retraction values, machine limits, or
other profile settings are changed.
# Screenshots/Recordings/Graphs
Not applicable. This is a machine start G-code profile fix.
## Tests
- Confirmed all four Ender-3 V3 SE profiles use the same `E15` / `E30`
purge sequence.
- Confirmed the current profiles do not explicitly issue `G90`, `M82`,
or `M83` before that purge.
- Confirmed OrcaSlicer emits `machine_start_gcode` before its generated
positioning/extrusion-mode preamble.
- Verified the modified JSON for all four machine profiles parses
successfully.
- Verified the added commands make the purge explicitly use absolute XYZ
and absolute extrusion state.
# Description
Follow-up to #15950, where a `filename_format` using
`initial_no_support_extruder` shipped broken because the profile
validator's slice sweep never expands `filename_format`. The sweep now
expands every custom G-code and `filename_format` text shipped in any
system profile. Each printer's slice also fires the pause, template
custom G-code and clumping-detection hooks and names the output file,
and the first printer shipping a `printing_by_object_gcode` also slices
by object. Beyond each printer's default process and filament, every
compatible system process and filament carrying a template text no
earlier slice has expanded is sliced once, which takes the sweep from
1,110 to 1,248 slices (about 49 s locally, up from 43 s). While the
sweep runs, the placeholder parser also resolves variable names inside
`{if}` branches a slice does not take, so one expansion checks every
branch.
The stricter sweep found two profile bugs, fixed here: the Anycubic
Kobra X filament change G-code carried an unreachable block reading
variables only Anycubic's own slicer defines, and the Wanhao France D12
template custom G-code had an unterminated `{if}`, so adding a template
custom G-code on those printers failed the slice. It also fixes a parser
bug where a declaration such as `{local a = layer_height + 1}` failed to
parse inside a branch that is not taken.
No change to slicing output for templates that already worked: the
untaken-branch check is enabled only by the validator's slice mode, and
the parser fix only lets previously rejected templates parse.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
New placeholder-parser cases cover the parse fix and the untaken-branch
check: off by default it changes nothing; on, it rejects undefined names
in branches not taken, accepts names declared there, and leaves boolean
expressions (compatibility conditions) alone. The full sweep passes on
all system profiles, and planting an undefined variable in an untaken
branch of a printer's start G-code, of a non-default filament's start
G-code or of a non-default process's `filename_format`, or reverting
#15950, each fails it and names the preset. `scripts/check_profile.sh`,
`libslic3r_tests` and `fff_print_tests` pass.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Multi-material previews now open in Filament view, where toolpaths are coloured
by the filament printing them, and a marker carrying a fixed palette colour can
sit on toolpaths of that same colour. The plan already resolves which physical
head each carriage is, so in that view every carriage - the primary marker
included - takes its head's filament colour, from the accessor the ghosts
restamp themselves with, and marker, ghost and toolpath agree on what is loaded
where. The primary marker is the one every preview draws, so it goes back to its
own colour when no parallel mode is active.
Every other view keeps the Okabe-Ito palette, which is what identifies the
carriages when colour means something else, and so does a head whose filament
does not resolve. Each head is resolved once per frame and shared by its marker
and its toolhead box, the way the ghosts already hoist the map they all read.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
X-Ray takes both shaded passes over and returns, and it draws only the volume
collection, which the ghosts are not part of. They disappeared while their
picking pass, hover tooltip and filament picker all kept answering, so a plate
in a parallel mode offered an invisible click target.
They go through the X-Ray shader rather than their own, so a ghost reads as one
more see-through body, which is what the mode is for. The pass supplies z_range
and clipping_plane itself, since the volume collection sets them for the whole
pass it runs and the vertex shader discards everything outside z_range. It also
draws two-sided, so a hollow ghost shows its far wall like a real body does, and
hands the shader an opaque colour: X-Ray derives coverage from the view angle
and multiplies the colour's alpha into it, so a ghost carrying its own
translucency as well would composite far fainter than the body it mirrors.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The two conflicts are both places where upstream landed on top of IDEX/IQEX
code. In GCode.cpp the relocated file header block meets the IMEX placeholder
block, and the placeholders are set first: file_start_gcode is processed through
the placeholder parser now, which throws on a name it does not know, so a script
naming {imex_mode} would abort the export if the header ran first. In
test_multifilament.cpp both sides appended a case at the end of the file.
The 3MF importers read XML entries into a single expat buffer whose size is
an int, while the archive extraction used the entry's 64-bit declared size.
The two could disagree for entries declaring more than INT_MAX bytes.
Reject such entries before allocating, and use one size for the buffer, the
extraction and the parse. This applies to the BBS importer, the PrusaSlicer
importer and the PrusaSlicer fingerprint probe. The load now fails with an
error instead.
* Reject 3MF Plate IDs Below 1 Instead of Indexing Before the Plate List
The plate importer copied each plater_id from model_settings.config into the
1-based plate list after checking only the upper bound, so plater_id="0"
wrote to plate_data_list[-1] and crashed on load. Both copy sites now reject
ids below 1 with the same "invalid plate index" error already used for ids
past the end.
* Drop Malformed 3MF Paint Data Instead of Reading Past the Bitstream
Painted facets are decoded from a bitstream a nibble at a time with no bound
check, so a truncated or corrupt paint string in a 3MF (for example split
codes with no children behind them) read past the end and crashed on load and
slice. A one- or two-side split naming side 3 also indexed past the triangle's
vertices.
Every nibble read now goes through a bounds-checked reader. Loading validates
each triangle's tree and drops a malformed one with a warning, so the stored
data, used extruder states and later decoding all agree. deserialize() also
unwinds and clears any triangle whose tree is incomplete or malformed, and
has_facets() stops at a truncated triangle. Valid streams decode unchanged.
* Sanitize Server-Supplied Download File Names
The URL downloader used the file name from the Content-Disposition header
as given, without the cleaning and unused-name search applied to the
URL-derived name.
Reduce the header name to a sanitized base name with the new
sanitize_file_basename helper, which splits on both path separators and
rejects names made only of dots and spaces. Run the result through the
same unused-name search as the URL-derived name, now shared in
find_unused_filename, and fall back to the URL-derived name when nothing
usable remains.
* Sanitize Download Names Before Choosing an Unused One
The unused-name search probed the name as given and sanitized the
result afterwards, so a name whose special characters are replaced
could be mapped onto a file that already exists.
Move the search into libslic3r as find_unused_filename, sanitize first
and probe the name that is actually written. The download marker path
is shared through download_marker_path. Restore the last tried name in
the error reported when no free name is found, and cover the search
with unit tests.
* Keep Downloads on an Unused Name Until They Complete
When the server supplied the name, the download marker stayed under the
URL-derived name, so the adopted name was not reserved against other
downloads. The final rename also replaced any file that took the name
while the download ran.
Move the marker to the adopted name before any data is written, and
check the name again right before the final rename, picking the next
free name if it is taken by then.
* Sanitize the File Name of Model Import Links
The model import took the file name from the link as given and only
avoided an existing file with a substring match on the folder listing.
Reduce the name to a sanitized base name, falling back to untitled.3mf,
choose the name with the shared unused-name search, and check it again
before the final rename.
* Handle Filesystem Errors When Finishing a Model Import Download
Choosing the final name and moving the downloaded project into place
could throw from inside the download callback. Any such error now removes
the temporary file and reports the existing import failure message.
* Non-crossing infill optimization
* test triangles
* test grid
* cleaning
* Align and clip rectilinear infill paths
Generate infill coverage in the pattern's local frame, rotate triangular patterns by layer, and clip centerlines to the surface vicinity. Start closed outlines outside the surface so clipping splits them cleanly.
* Update test_fill.cpp
* Update multiline-infill.md
---------
Co-authored-by: Ian Bassi <ian.bassi@outlook.com>
# Description
Extruder variants (Standard, High Flow, extra high flow) now work on any
printer. Any vendor profile can declare them, and a multi-variant
filament picks up the right variant on every printer. In the sidebar,
users can switch the printer variant and set the nozzle volume type of
each extruder on multi-extruder printers. This also fixes a later
filament printing at the first filament's temperature on a P1S or X1C
with a High Flow nozzle. The profile checks now reject variant arrays of
the wrong size and outdated variant strings. Every shipped vendor
profile passes them, and the orca-profiles skill documents the rules.
Slicing output changes only where the wrong variant was used before.
# Screenshots/Recordings/Graphs
<img width="393" height="218" alt="Screenshot 2026-09-28 at 11 28 07 PM"
src="https://github.com/user-attachments/assets/8bee4b0e-c38c-4039-8c11-096ace6fbad0"
/>
<img width="448" height="267" alt="Screenshot 2026-09-28 at 11 28 37 PM"
src="https://github.com/user-attachments/assets/e767cd97-a5d0-4728-b010-c8ea2bc94ca7"
/>
<img width="746" height="603" alt="Screenshot 2026-09-28 at 11 28 54 PM"
src="https://github.com/user-attachments/assets/23c8af3b-c679-46e5-9fd0-2e43df0449b3"
/>
https://github.com/user-attachments/assets/10d75d72-86a3-42e8-8a95-b1627bd58e91
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The block ran only when flush_length_4 is -1392 and read ace_t_box_vector / ace_t_slot_vector, which only Anycubic's own slicer defines. It emitted comments only, so the printed G-code is unchanged.
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
This PR re-orders the OFL OTA auto-publish workflow by using successful
cron runs as checkpoints, marking and waiting for every dispatched
profile publisher to finish, and clearing the pending queue once
centrally only after all publishers succeed.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Merged by /bot merge on behalf of @peachismomo (id 52488812).
Grants: resources/profiles/OrcaFilamentLibrary/filament/Elegoo, resources/profiles/OrcaFilamentLibrary.json, resources/profiles/Elegoo, resources/profiles/Elegoo.json
Head: d4a18c633e
* Add missing TopTools includes to GeometryEngine.cpp
* Add missing TDF_LabelSequence include in STEP.cpp
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
This PR fixes some problems with the current profiles for Blocks
printers
- **Fix**: start gcode for the Blocks 0.6mm RF50 printer where the start
gcode command had no newline separation on two commands so they were
concatenated and would make printing fail.
- **Added**: multi material plates selection for the RF50 printers,
right now only textured PEI and smooth high temp plates are available,
but in the future more to come.
- **Naming consistancy** for Blocks Pro S100 processes and filaments,
target labels used bare `@Blocks` while RD50 and RF50 used
`@Blocks_RD50` and `@Blocks_RF50`. Renamed Pro S100 to
`@Blocks_Pro_S100` to match the structure.
- **Filament restructure**: replaced Blocks tuned-generic filaments
(e.g. `Generic PLA @Blocks`) with real Blocks-branded filaments. One for
each printer model. Every material keeps the same coverage we already
had.
- **Fix**: `sparse_infill_pattern` was misspelled `sparse_infill_patter`
on 13 `@Blocks_RF50` process presets. Harmless due to inheritances but
fixed non the less.
- **Fix**: `Blocks RD50 V2 0.4 nozzle.json` was missing
`printer_variant`.
- **Normalize** all Blocks profiles to standard formatting (tab
indentation, key order) and bumped `Blocks.json` version.
- **RF50 retraction length** increased on all RF50 machines.
## Tests
- [x] `orca_profile_tool.py check --vendor Blocks` - clean
- [x] `orca_profile_tool.py check` - clean
- [x] `normalize` / `generate-id` / `update-index --dry-run` — fully
settled, no pending changes
- [x] `check_profile.sh --vendor Blocks` → `profile_tool` check — passes
- [ ] `validate_system` / `validate_slice` /
`validate_filament_subtypes` / `validate_custom` —
could not run locally (dev host's glibc 2.36 is older than the nightly
validator binary's
required 2.38); needs CI or a compatible host
- [x] In-app confirmation that the Blocks bundle loads
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The segments come in print order, bottom layer first, which seen from above is back to
front: every hidden fragment is shaded before the one that covers it, and on an integrated
GPU that overdraw is most of the frame. Drawing the instances last to first whenever the
camera looks down lets the depth test reject the hidden fragments instead. Side views and
views from below keep the print order, and the shadow-caster pass is unchanged.
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
Avoid calling Print::output_filename() when a slice is canceled or
fails, since unresolved filename placeholders can throw before G-code
export completes.
Instead, we should use the model name in ctx.name and the stable model
ID in ctx.id, consistently across slice and G-code export events.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Fix#15885
Synchronize the plater filament controls after preset loading, even when the internal filament list already matches the nozzle count.
Co-authored-by: Codex <codex@openai.com>
Fix size_t/%d mismatch in MsgDialog button keys
`m_buttons.size()` is a `size_t`, which does not match the `%d` conversion in
printf-style variadics. Build the key with `std::to_string` instead; exact
for any size_t, no behavior change for realistic counts.
Co-authored-by: yw4z <ywsyildiz@gmail.com>
Guard the smooth-normals preference read when CLI thumbnail generation initializes geometry without a wx application. Use the existing flat-normal path in that case and preserve GUI preferences.
Use DPI-aware 20 DIP vertical scrolling for general-purpose GUI
scroll areas.
Keep list-based views aligned to their item height so one wheel increment follows the visible row rhythm.
This makes Preferences and other dialogs scroll consistently across
Windows DPI settings.
get_dpi_for_window's pre-8.1 fallback and get_font_list_by_enumeration
both called GetDC without a matching ReleaseDC, leaking a GDI handle
each call. get_dpi_for_window runs on every mouse-move over the 3D
viewport, so the leak exhausts the per-process GDI handle limit and
hangs the app within minutes on Windows 7/8.
Co-authored-by: Fernando Marino <f.marino@rheagroup.com>
Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
The OFL OTA workflow sent an ISO-8601 timestamp to the pending-clear
endpoint, which only accepts Unix epoch seconds. This caused the
scheduled workflow to fail with HTTP 400.
Use `data -u +%s` so that request matches the endpoint contract.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
If a vendor runs `/bot merge` while the cronjob is running, it might be
dropped because the table might be cleared before `post_merge_profiles`
completes. Instead we can add a timestamp so that we don't accidentally
drop any PR merges that occur while the OFL cronjob is running.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The bake is covered in the libslic3r suite, but nothing carried its
result through the slicer - where a mesh it cannot use shows up as a
crash or as a print with no toolpaths, which is what was reported on
macOS and Linux.
Paints a cube, bakes it as the gizmo commits it (mesh replaced, hull
recomputed, paint cleared), slices it and checks the G-code has
toolpaths and layers. Once under its budget and once capped, since a
capped bake also goes through the simplification and the repair.
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
The previous implementation checks against the last successful
`post_merge_profiles` which can trigger when a normal OTA update for
non-OFL profiles are made. This causes the check to fail when OFL
changes are made before other regular profile changes are made in
`resources/profiles/<vendor>`
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Merged by /bot merge on behalf of @peachismomo (id 52488812).
Grants: resources/profiles/OrcaFilamentLibrary/filament/Elegoo, resources/profiles/OrcaFilamentLibrary.json, resources/profiles/Elegoo, resources/profiles/Elegoo.json
Head: b73e9df4f4
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
- `OrcaFilamentLibrary` (OFL) has no `FOLDER_MERGERS` delegation and
changes far more often than other vendors, so it can't go through the
existing merge-triggered, human-published OTA flow.
`post_merge_profiles.yml` gains an explicit `vendor` + `auto_publish`
`workflow_dispatch` path that bypasses the `FOLDER_MERGERS` check and
calls the OTA auto-publish API directly, making the update live with no
human step.
- `ofl-ota-cronjob.yml` drives that path daily across `main` and every
`release/vX.Y.Z` branch, checkpointed against
`post_merge_profiles.yml`'s own per-branch run history so a failed
publish is retried rather than silently dropped.
- Fixes a pre-existing bug in `post_merge_profiles.yml`: a single vendor
with no `FOLDER_MERGERS` grant in a push used to zero out the *entire*
vendor batch, silently dropping other, properly-authorized vendors'
publishes too. It now drops only the ungranted vendor, with a warning.
## Notes
Once v2.5.0 stable is released, we need to remove branch checking
against main in both these workflows.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
# Description
This brings back the Bambu Lab profile sync with BambuStudio, reverted
from main, along with its bed-model offset fix. The synced printers and
filaments share their start and end G-code and dual-nozzle settings
through template files that Orca did not read, so those printers had no
machine G-code. Orca now loads these templates the same way BambuStudio
does, so every synced preset comes in complete. `profile_include_dump`,
a new developer tool, compares the result with BambuStudio's.
Other vendors' profiles are unaffected. Bambu Lab profiles move to
version 02.08.00.10, so installs that took the earlier sync update too.
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
Unit tests cover how template settings combine with inherited and preset
settings, loaded from JSON and from the preset cache. The profile
validator passes on every shipped vendor, and every Bambu Lab preset
gets the same template values in Orca as in BambuStudio.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
A scarf joint split the loop at exactly the scarf length, so the remainder
of the segment the split landed in became the first flat segment, often a
fraction of a millimetre. The seam insertion also leaves segments of a few
micrometres at both ends of the loop, which a scarf extrudes through where a
plain loop would start or stop. With junction deviation the planner treats
such short moves as tight corners, limited by the Z axis acceleration at the
end of the ramp, and slows to a third of the wall speed or nearly halts at
the seam.
Extend the ramp to the next vertex when the remainder would be shorter than
half a scarf step, capped at a millimetre, beyond which planners treat a move
as ordinary; the scarf only grows, never shrinks. Drop the vertex next to the
seam point at either end when that segment is shorter than an eighth of a
common line width, so the loop still starts and ends at the seam; a trimmed
path loses its arc fitting and prints as line segments. Clamp the scarf
length to the trimmed loop so a scarf covering a whole loop still ends at
full flow. The descending pass reuses the same path, so both wedges stay
consistent, and a flat part that would collapse to a single point is dropped.
The global arrange branch, taken by --arrange with all plates selected, only
reserved the prime tower when filament ids had been given on the command
line for STL input. A project carries its filament use per plate and its own
tower positions, but that set was empty for it, so the tower was never an
obstacle: the arranged pile was centred over it and the slice then failed on
a G-code path conflict.
When no filament ids were given, count the filaments each plate uses and
reserve a tower on every plate that needs one, keeping the project's tower
position instead of resetting it to the default. Only a tower the slicer will
print is reserved: the prime tower must be enabled, and a by-object print
gets none unless a smooth timelapse needs it, as the per-plate arrange
decides. Overflow beds are sized for the busiest plate. The STL route is
unchanged.
G-code generation collects errors raised per object, such as an empty
first layer, into one SlicingErrors exception whose own message is just
"Errors". The CLI's generic handler printed that word and recorded the
generic slicing error text, so a headless caller had nothing to act on.
Let Print render the per-object messages with each object's name, and have
the CLI catch SlicingErrors ahead of the generic handler, print that text
and record it as the result's error string. The exit code is unchanged. A
unit test lifts a cube off the bed and checks the message names the object.
Since #15811 the settings page is built when its tab is shown, so on a
Home start that opens a project the Quality page is built on screen. Its
flow-compensation-model field is a multiline text view that GTK maps as
it is created and that is hidden, because compensation is off, before
GTK first allocates it. The loading dialog's wxWindowDisabler then
desensitizes the frame, and GTK crashes in
gtk_text_layout_cursors_changed() on that text view.
On GTK the page view is now hidden while a page is built, so a control
that starts hidden is never realized and GTK realizes it when it is
shown. The deferred build is unchanged.
For clang-cl, CMake sets the depfile flags to the gcc-style -MD, -MT and -MF,
passed through as -clang: arguments. ccache does not parse those, so a cache
hit writes only the object. Ninja has no depfile to read, so it records zero
header dependencies for that object and does not rebuild it after a header
edit. The stale object is still linked into the library and the DLL. sccache
0.15.0 reproduces the depfile on a hit and is unaffected.
Use /showIncludes instead. CMake already does that for MSVC, and ccache
reproduces it on a hit. A make-rules override sets the flags, because CMake
includes the override after Platform/Windows-Clang.cmake. It is forwarded to
each dependency because every one configures as its own CMake project, and it
is only set when the file exists, because scripts/flatpak/make_deps_tar.sh
packs deps/ alone.
In a build with a warm cache, 678 of 790 objects had no recorded headers.
Object code does not change. One GUI translation unit compiled both ways is
byte-identical apart from the COFF timestamp.
Extend the existing U1 filename format to the base process preset and let Benchy presets inherit it. Bump the Snapmaker catalog version.
Co-authored-by: Codex <codex@openai.com>
* preset updater: refresh installed vendors from resources regardless of enable_ota
Since c4fea8ad24 the resources check in check_installed_vendor_profiles()
sat behind enabled_config_update, which now follows enable_ota, a hidden
flag that defaults to off. The hotfix a93c6ea67b then made that gate skip
installed vendors entirely, so a new build's newer vendor profiles were
never installed over an existing vendor unless OTA had been turned on.
Before the gating the update URL always had a default, so the comparison
effectively always ran.
The resources shipped with a build are not an over-the-air update. Judge
installed vendors against them, and drop the ones no longer enabled,
regardless of the flag; enable_ota keeps gating the online sync.
* preset updater: stop reinstalling the filament library on every launch
check_installed_vendor_profiles() put OrcaFilamentLibrary on the install
list unconditionally, so every launch recopied the whole vendor from
resources and, in a build that ships the profile JSONs rather than a
preset cache, then re-parsed and re-cached it: about 0.3 s of a dev
build's startup, and a 3 MB copy in a release build, for a vendor that
had not changed.
The library was special-cased because it is never in the enabled-vendor
list. Treat it like the default bundle instead: always wanted, and
reinstalled only when the resources carry a newer version.
libpng reports a corrupt image by longjmp()ing to the jump buffer the
decoder sets. Both decoders set theirs in a frame holding a PNGDescr and
a vector, and with exceptions enabled MSVC unwinds the stack as part of
longjmp - so returning from that frame afterwards crashed. A truncated
texture segfaulted the Windows test run and was fine everywhere else.
The calls that can fail now sit in two helpers that own nothing but
pointers, and the C++ objects stay in the callers' frames.
texture_displacement_patch_frame picks the world axis least aligned with
the patch's average normal. With no patch there is no normal either, and
the axis came out of the +Z fallback for it - so it was +X, while its
test asked for the upright frame the fallback is meant to give. An empty
patch now answers +Z outright.
The Windows build stopped on test_kdtree.cpp: it calls std::iota without
including <numeric>, which libstdc++ happens to pull in anyway. Added
there, and the same for <limits> and <algorithm>/<cmath> where the
recent changes rely on them being included by something else.
bridge_over_infill's helper for splitting polygons by proximity named
its locals near and far. The Windows headers define both as macros that
expand to nothing, so "Polygons near;" declared nothing and the uses of
it did not compile. Renamed; no behaviour change.
* fix: resolve filament vendor/type across split base presets
WebGuide's filament list dropped presets whose vendor and type came from different ancestors, and looped forever on inherits cycles. Walk the full inherits chain with a path-scoped guard, fill only missing values, and skip presets that still lack vendor or type. Add tests for split-base resolution and validate the real profile tree.
* Merge branch 'main' into feat/filament_dialog_fix
* Fixes unit test
* Merge branch 'main' into feat/filament_dialog_fix
* Merge branch 'main' into feat/filament_dialog_fix
* revert tests/fff_print/test_gcodewriter.cpp changes
* Add Sovol Zero nozzle profiles and filament updates
Preserve inherited tuning with Zero-only overrides so shared
Sovol defaults and other printers remain unchanged. Retain migration
aliases while using canonical names in compatibility references.
Co-authored-by: Codex <noreply@openai.com>
* Fix Sovol Zero bed model reference and ignored settings
Point the Zero model at the shipped bed mesh. Remove unsupported filament
and process keys that the loader silently discards, preserving effective
tuning and the existing preset names and IDs. Bump the Sovol bundle version.
Co-authored-by: Codex <noreply@openai.com>
* Fix Sovol Zero filament identity and vector formatting
Add hardened-steel nozzle variants with disjoint filament compatibility, retain generic product IDs and preserve renamed preset aliases. Convert 43 scalar vector fields to arrays without changing tuning values and bump the Sovol bundle version.
Co-authored-by: Codex <codex@openai.com>
* Keep only shipped names in the Sovol Zero profile renames
The Zero machines, processes, machine model and SPEEDBENCHY filament carried renamed_from entries for names that only ever existed inside this branch (Sovol ZERO ..., " - Brass", " - Hardened Steel"). Drop those, and keep only the redirects that matter to existing users: the historical Sovol Zero material names and the three deleted hardened-steel presets. The Sovol Zero machine model also sheds the setting_id, instantiation and from keys the model loader ignores.
---------
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: Codex <codex@openai.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
Unwrapped, it ran past the edge of the panel and pulled the layout with
it. Counts under a million now read in thousands, so a 119 k budget is
no longer shown as 0.1 M.
They described the implementation, and a few no longer matched the code.
Each now says what the control does and in what units; five controls that
had no tooltip have one. The catalogs need regenerating.
Nothing removed the even mesh over a texture's flat parts unless the
budget was lowered until simplification had to run. Collapses stay under
a thirty-second of the resolution: on the test map 387584 -> 369876
triangles, with the print unchanged (0.02% of the extrusion).
An earlier attempt was reverted for a 21% heavier result. That was the
colour despeckle switching itself off, not the collapsing: it keyed off
the face parent map, which any collapse empties.
Relief preserved from an earlier bake is counted on top of the budget
instead of eating into it: baking a second area went from ~214 k
triangles to the ~1000 k it was given.
When the resolution needs more triangles than the budget allows, it is
said before the bake under Resolution and again afterwards.
Polymaker, Overture and eSUN presets move from brand subfolders to
BBL/filament/, matching BambuStudio's layout; their names and ids are
unchanged. Default filament lists keep pointing at presets that exist,
0.2 mm nozzles keep a default PLA, and user presets based on removed or
renamed presets still load. Generic SBS is no longer in the Bambu bundle;
new selections on those printers fall back to OrcaFilamentLibrary's
Generic SBS @System. The profile tool now recognises BambuStudio's new
filament id/name maps and support_recommended_params.json as data files.
The Polymaker, BETA, COEX, Overture, addnorth, Numakers, FusRock and
AliZ presets that BambuStudio does not ship now live under
OrcaFilamentLibrary/filament/<Brand>/BBL. The BBL bundle now holds
only the filament presets BambuStudio ships, plus a few Bambu and
generic ones.
New "<product> @BBL base" presets carry the values these presets used
to get from BBL's own bases, so their settings on Bambu Lab printers
are unchanged.
The nightly found its build with a filtered run listing (branch=main,
status=success) and trusted the first result. GitHub serves filtered
listings from a run search index that has intermittently returned
weeks-old results, so some nights tested a build from weeks earlier and
reported its differences as regressions. The same filter also matched
fork PR builds whose branch is named main.
The build is now picked from the unfiltered listing, which stays
current, and filtered here: a successful build_all run of this
repository on the requested branch. Fork PR builds are excluded by
repository. A feature branch is normally built only for its PR, so this
repository's own PR builds stay eligible, but a PR build compiles the PR
merged into its base rather than the head commit the later jobs check
out, so a push or dispatch build of the branch is preferred when the same
page of the listing has one. A scheduled run fails instead of testing a
build more than 48 hours old, and every run names the build it tested
in the job summary.
Manual runs scan further back, so a branch that last built weeks ago
can still be tested, and a new build_run_id input pins one build_all
run, read directly rather than through a search.
# Description
This PR introduces lifecycle events to the plugin API.
For all plugin capabilities, you can define a `on_lifecycle_event`
function in the plugin that takes in a event enum and a small payload
for some generic information on the lifecycle event.
The idea is to keep the payload generic and small, and if you want to
get more information, you should invoke other more targeted APIs to get
more information.
For example, lets say you are keeping track of the the `ObjectAdded*`
event hook for model transformation, addition or deletion. The payload
would tell you the name of the model, and you should use a targeted API
such as `orca.host.plater().model()` to get more information on the
model. This is the overall design principle of the API.
Currently the lifecycle events are the following:
```cpp
enum class LifecycleEvent {
// Project (3mf)
NewProject,
ProjectOpened,
ProjectBeforeSave,
ProjectAfterSave,
ProjectClosed,
ProjectDirtyChanged,
// Slicing pipeline
SliceStarted,
SliceGeometryFinished,
GCodeExportStarted,
GCodeExportFinished,
SlicingJobComplete,
// Plate/model editing
ObjectAdded,
ObjectDeleted,
ObjectTransformed,
ObjectChanged,
ObjectRenamed,
PlateCreated,
PlateDeleted,
PlateSelected,
PlateRenamed,
// Preset
PresetSelected,
PresetSaved,
// Printer/device
PrintStateChanged,
DeviceOnlineChanged,
DeviceDiscovered,
DeviceSelected,
DeviceConnected,
DeviceDisconnected,
UploadStarted,
UploadFinished,
// Print/send jobs
PrintJobStarted,
PrintJobFinished,
SendJobStarted,
SendJobFinished,
};
```
This is an initial draft and lifecycle events can be included later on.
[orca_telegram_notifier_plugin_any.py](https://github.com/user-attachments/files/31220973/orca_telegram_notifier_plugin_any.py)
If you're familiar with telegram bots, after you install the telegram
bot, in the config of this plugin, you can enter the Bot ID and the Chat
ID with said bot.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
For this plugin, I am testing it with a telegram bot that sends me a
message on lifecycle event.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The CLI turned "align to Y axis" on for every i3 printer with no way to opt
out. With rotations forbidden the pre-rotation is the result, so every object
ends up turned 90 degrees from how it was loaded. The GUI defaults the
checkbox the same way for i3 printers, but lets the user untick it.
Add --align-to-y-axis. When it is not given the printer-structure rule still
applies, so existing calls are unchanged; the CLI's own options are filled
with defaults after parsing, so the keys the user typed are remembered to
tell the two apart.
* Answer the Preview's Per-Frame Time Query From a Cached Sum
The G-code preview's cost is linear in the number of toolpath vertices, and on a
tall multi-filament print the wipe tower dominates that count: it emits a roughly
constant 160-180 moves on every layer whatever the object is, measured at 57-61%
of all moves on a three-filament print.
Four places scanned or allocated across the whole vertex array. None of them
needed to.
get_estimated_time_at re-accumulated the estimated time from vertex 0 on every
call, and its caller is the tool marker tooltip, which ImGui re-renders every
frame while the properties panel is unfolded. It now starts from a running sum
kept at each layer's first vertex, built at load in vertex order, and adds only
that layer's vertices: the same additions in the same order, so the float result
is unchanged, at a cost of one float per layer and time mode rather than per
vertex. At the 351k vertices of a 636-layer test print the call scanned the whole
print (238us); it now scans one layer.
update_view_full_range walked from vertex 0 to find where the layer range starts,
on every slider tick. It now starts at the first vertex of that layer. The index
is derived from the vertices rather than from Layers::Item::range, because
Layers::update folds a vertex whose layer_id arrives out of order into whichever
bucket is open, which makes that range the wrong answer in general; the index
costs four bytes per layer, not per vertex.
update_colors_texture allocated one float per vertex of the whole print on every
slider tick. It now reuses a buffer.
render_legend fetched the layer Zs and the per-layer times from inside loops over
the custom G-code items, and built whole vectors only to test them for emptiness.
The times are hoisted, the Zs are built lazily so a print with no colour change
does not pay for them at all, and the emptiness tests use the existing counters.
No rendering behaviour changes.
* Draw the Preview's Toolpath Segments From an Index Buffer
The preview's frame cost is dominated by one call: a single instanced draw of
every visible toolpath segment. On a tall multi-filament print the wipe tower
supplies most of those segments, which is why the preview of a large tower is
slow and why shrinking the layer range speeds it up again.
That draw is not fill bound. Shrinking the model to about a fortieth of its
screen area moved the frame from 419 ms to 401 ms, so the cost is per segment,
not per pixel, and it is paid in the vertex shader: five texelFetch calls plus
several cross/normalize per invocation.
Each segment is a box of eight corners, but it was submitted with
glDrawArraysInstanced over a 24 entry array, so every corner was transformed
once per triangle that touches it and the shader ran 24 times per segment. The
same 24 entries are now an element buffer over the eight distinct corners, which
lets the post-transform cache reuse them and drops the shader to 8 runs per
segment. The triangles, their winding and the vertex_id each corner receives are
unchanged.
Measured over 100 frames on the 636-layer, 351k-vertex three-filament fixture,
the segment draw goes from 381 ms to 322 ms per frame. That is a software
rasterizer, where triangle setup dominates and understates the win; the drop in
shader invocations is the transferable part.
Verified by loading the same project in this build and in a build of the parent
commit and comparing the canvas across three states - the default view, a
rotated camera, and a reduced layer range: pixel identical in all three. The
rotated case matters because the shader picks its corner offsets from the camera
direction. The only pixels that differ anywhere on screen are in the G-code text
panel, which prints a per-process object id that varies between any two runs.
* Add Prusa CORE One MMU3 profiles
Dedicated MMU3 CoreOne profiles (like those used on prusaslicer 3).
Tested with latest coreone and mmu3 firmware, and works as well as prusaslicer.
Correct model default materials to reference compatible MMU3 filaments.
Co-authored-by: Codex <noreply@openai.com>
* Consolidate the CORE One MMU3 generic filaments
The bundle shipped two families covering the same four MMU3 variants for
each generic material: a standalone `Generic X @MMU3` and a
`Prusa Generic X @CORE One MMU3`. The `Prusa Generic` spelling was renamed
away from the rest of the tree, so re-name the CORE One-tuned family to
`Generic X @Prusa CORE One MMU3` (they inherit the CORE One tune and now the
shared generic product id) and drop the standalone files, which only carried
raw material-base values. Repoint the model default_materials and re-register
the index.
* Fix the CORE One MMU3 0.4 default process
default_print_profile named `0.20mm Speed @COREONE0.4 + MMU3`, but the
preset is `0.20mm SPEED @COREONE0.4 + MMU3`. Preset lookup is case-sensitive,
so the intended default never resolved and compatibility selection silently
picked another tier.
* Normalise the CORE One MMU3 process names
Match the bundle's all-caps quality ladder: `Fast Detail` -> `FAST DETAIL`,
`Speed` -> `SPEED`, `Structural` -> `STRUCTURAL`, `Balanced` -> `BALANCED`.
Filenames now equal their preset name, as every pre-existing Prusa process
file does, and the index is rebuilt for the renamed entries.
---------
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
Collecting every candidate within the radius into a vector cost more
than the search itself. Same order, so the seams are unchanged;
align_seam_points ~19.6 s at 0.1 mm / 2000k, was ~21.
detect_surfaces_type, process_external_surfaces and the vertical shells
each waited on their own heaviest layer in turn. The LOTR map plate
slices in ~10.5 min at 0.1 mm / 2000k, was ~11.5; ~87 s at normal
settings, was ~97.
#15668 saves the compiler cache on cancelled and failed builds and then
drops the older entries for the leg on the ref. actions/cache/save only
warns when its tar fails, so a cancelled build whose ccache directory
was still being written saved nothing, the drop ran anyway and deleted
the leg's last good entry. The next run on main restored nothing and
compiled cold, and so did every PR that restored in the gap. Run
35405244634 (Flatpak x86_64, 2026-09-18) did this to
ccache-Flatpak-x86_64-35397824860-1; between 13 and 18 September 9 of
87 cancelled main build jobs did the same.
Look the new entry up before deleting anything, and keep the older ones
when it is not there.
MultiPoint had no rvalue constructor, so the derived move constructors
bound to the const reference and copied; append reserved exactly, so
collecting pieces one by one was quadratic. Colour segmentation ~3 s at
0.1 mm / 2000k, was ~40, and ordinary prints gain too.
Only the slices within the deepest shell offset decide the result, so
the work is done per tile of the face. Top and bottom segmentation
~130 s at 0.1 mm / 2000k, was ~180.
ClipperLib slows down with the number of edges on a scan line, and a
layer cut through a fine relief has tens of thousands of pieces.
detect_surfaces_type ~50 s at 0.1 mm / 2000k, was ~145.
Adds the LulzBot Mini 1 (single extruder, 0.5 mm nozzle, 2.85 mm filament)
to the existing Lulzbot vendor, which previously shipped only TAZ models.
Values are taken from LulzBot's own current slicer configuration
(github.com/lulzbot3d/CuraLE) rather than estimated:
- geometry and custom g-code from resources/definitions/single_mini_mini_1.def.json
and resources/gcodes/mini_1/{mini_1_start,mini_1_end}.gcode
- filament temperatures and cooling from resources/materials/*.xml.fdm_material,
overridden by resources/quality/single_mini/<material>/*.inst.cfg
The start g-code reproduces the Mini's nozzle wipe and four-washer G29 probe,
with the filament-type temperature conditionals used by the sibling TAZ profiles.
One deliberate departure from current CuraLE: the pre-wipe retract is 30 mm,
the value used by Cura LE 4.13.x, rather than the 4 mm current CuraLE uses.
The Mini probes by electrical contact between nozzle and washer, so the nozzle
must stay clean through the wipe and all four touches. At 4 mm the melt zone
stays full and can ooze onto a washer, which caused auto-levelling failures on
hardware; 30 mm empties it. The cost is a ~24 s purge at print start.
The three filament presets follow docs/HLSD/filament_id.md rule 4: a vendor
tuning a generic material inherits Generic X @System, keeps the Generic X base
name and declares no filament_id, so identity stays with OrcaFilamentLibrary.
Their compatible_printers is the Mini alone, disjoint from Generic X @Lulzbot
(TAZ only) as rule 3 requires. They exist because temperature is a filament-scope
setting, so LulzBot's values need printer-scoped presets. All six plate types
carry the same temperature, because the Mini has a single bed and curr_bed_type
can hold a stale value carried over from another printer.
Strictly additive: no existing profile is modified, so TAZ behaviour is
unchanged and no migration is required. Lulzbot.json is bumped to 02.04.00.05.
Verified with scripts/orca_profile_tool.py check and scripts/check_profile.sh
on the full tree (profile tool, system validation, slice, filament subtypes and
custom-preset fixtures all pass), by diffing sliced output against LulzBot's own
Cura LE g-code for the same model and filament (temps, retraction, speeds and
the full wipe/probe sequence match), and by printing a 3DBenchy on a Mini 1 over
OctoPrint.
Claude-Session: https://claude.ai/code/session_012aLyqsXB7FKqwQdE2QF7Et
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
The merge took anything from 3 to 38 minutes at 0.1 mm / 2000k, now
~2.5. Every region is grouped with the islands it overlaps, so the
result is the same.
Preset::get_printer_type matches a preset's printer_model against every vendor's model names and returns the first hit, so two bundles declaring one name make the lookup order-dependent, and the Add Printer list shows the printer twice. The existing name check is per bundle, because base profiles share names across vendors by design; this one covers the global machine_model namespace and runs over the whole tree, like the setting_id and filament_id checks.
Register high-flow filament defaults for every U1 nozzle variant in the
wizard, and select compatible filaments for the standard and mixed-nozzle
printers. Size the shared U1 extruder vectors to its four physical tools.
Replace stale 3x flow claims with the source-specific filament guidance.
Bump the Snapmaker bundle version.
Co-authored-by: Codex <noreply@openai.com>
* Rework Ultimaker profiles
The Ultimaker profiles where not usable by default
* Fixes
* couple small fixes
* Rework Ultimaker profiles
The Ultimaker profiles where not usable by default
* Fixes
* couple small fixes
* Fixes
* fix errors
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
Fixed bug to ensure filename format respects multi-extruder filament selection
Improved filename format to ensure that the selected filament is included in the filename and not the [0] filament
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
This is needed for the AnkerMake M5 and M5c to have correct print time estimates. Fixes a random super-long time on the M5 touchscreen.
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Fix incorrect printable_area for Raise3D Pro3 and Pro3 Plus profiles
All Raise3D Pro3/Pro3 Plus machine profiles (Left, Right, Dual) shared
the same 340x300 printable_area regardless of single vs dual extruder
use. This overstated single-extruder X travel by 40mm and failed to
shrink the Dual profile to the real nozzle-overlap zone.
Corrected to Raise3D's published build volume specs:
- Single extruder (Left/Right): 300 x 300 mm
- Dual extruder: 255 x 300 mm
printable_height (300 for Pro3, 605 for Pro3 Plus) and origin (0,0)
are unchanged; both were already correct.
Source: https://www.raise3d.com/pro3-series/
Co-authored-by: Jon Ashton <ashtonj@zentechman.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
* Increase max volumetric speed for CR-ABS filament
* Update filament settings for CR-PETG profile
* Increase max volumetric speed from 16 to 18
* Increase max volumetric speed from 16 to 18
* Increase max volumetric speed for CR-PLA Matte
* Adjust filament temperature and speed settings
* Update filament settings for ENDER FAST PLA profile
* Modify pressure advance and filament load settings
* Adjust filament cost, speeds, and pressure advance
Updated filament settings for Hyper PETG @K2 Plus.
* Update filament cost and pressure advance values
* Increase max volumetric speed from 10 to 14
* Enable pressure advance in filament profile
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
A colour texture baked at 0.1 mm / 2000k made the top layers thousands
of islands and slicing never finished. Colour segmentation runs per
island, the merge subtracts piece by piece, the support check tests only
nearby islands, and the travel ordering finds crossings through a grid.
* profile: reorganize the ZR Ultra family and fix the Ultra S tool count
- Adds a new fdm_ultra_common profile to hold all common ZR Ultra toolchanger attributes.
- The S variants are the base machines plus an enclosure heater and filtration, so each now inherits its matching ZR Ultra profile instead of duplicating the per-nozzle values
- Also fixes Ultra S 0.6 and 0.8 - original were declared a single nozzle_diameter entry, so OrcaSlicer treated four-tool machines as single-extruder.
- ZR Ultra S 0.8's retraction_minimum_travel now matches the base Ultra.
- nozzle_diameter stays declared on each S variant rather than inherited, even
though the value is identical to its parent's. Several profile consumers read
these files without resolving `inherits` -- the config wizard's loader and the
web Profiles page among them -- and 1012 of the 1013 instantiated machine
profiles in the tree declare it, so this is the format's expectation rather
than redundancy. The same rule is enforced for filaments' compatible_printers
by scripts/orca_extra_profile_check.py.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix errors
---------
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
fix: install all Windows runtime DLLs
Keep the runtime DLL list local while it is assembled.
The previous PARENT_SCOPE assignment exported the initial list before the OCCT DLLs were appended. CMake then created a local list containing only the appended OCCT entries, causing the generated Install manifest to omit GMP, MPFR, WebView2, FreeType, and FFmpeg DLLs.
Export the completed list only after all runtime DLLs have been added.
* add eryone config
* add eryone config
* add eryone config
* Add 13 Eryone filament presets for the Thinker X400 0.4 nozzle
Rework the stale author branch onto the current Eryone bundle. Main already ships
the Thinker X400, so the duplicate "Eryone Thinker X400" machine and process family
from the branch is dropped and the new filaments are pinned to the existing
"Thinker X400 0.4 nozzle" variant. Hand-typed setting_id/filament_id values are
replaced with generated ones, the bundle version is bumped, the stray .info sidecars
are removed, and Eryone PETG-CF's filament_settings_id is corrected to match its name.
* fix errors
---------
Co-authored-by: Eryone <technical@eryone.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
The slice check centres its cube on the bed, puts the prime tower beside
it, then pulls the tower alone inside the printable outline. On a bed too
narrow for the estimated footprint that pull drags the tower back over
the cube: Volumic EXO42 IDRE MIRROR MODE (189 mm wide, 87.6 mm estimate)
logged "gcode path conflicts found between WipeTower and cube" in every
run, and three ~105 mm beds were left with 0.15 to 3.3 mm of clearance.
The cube and the tower's footprint are now pulled inside as one rigid
pair, so the clearance between them is fixed by construction. A bed too
small for the pair keeps the old placement, and presets that were never
clamped keep their exact layout.
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
A Python printer agent plugin could take the host down, and one of its
operations could never report its result. This PR fixes both in
`PrinterAgentPluginCapabilityTrampoline.hpp`.
## Changes
### A faulty printer agent no longer throws into the GUI
`IPrinterAgent` reports failure through return values, and none of its
callers catch. A Python `raise`, a missing override or a wrongly typed
return from a printer agent plugin therefore escaped the trampoline as a
C++ exception.
Every trampoline operation now catches, logs `Printer agent plugin
'<key>': <operation> failed: <error>`, and answers with what
`NetworkAgent` returns when no printer agent is set. `BBLPrinterAgent`
returns the same values when the Bambu plug-in is unavailable:
- `-1` for every `int` status code
- `false` for `start_discovery` and `fetch_filament_info`
- `""` for `get_user_selected_machine`
- an empty `AgentInfo` for `get_agent_info` (registration already
rejects an empty agent ID)
- `FilamentSyncMode::none` for `get_filament_sync_mode`
`ORCA_PY_AGENT_OVERRIDE(ret, name, ...)` derives the fallback from the
return type through `printer_agent_failure<ret>()`, so the call sites
carry no fallback values of their own.
An exception is the safety net for plugin bugs, not an error channel. A
plugin reports an expected failure by returning a code, as the Bambu
plug-in does. A raise is logged as a failure and collapses to the
generic `-1`, so the GUI shows the generic message instead of the
specific one (`-18` cancelled, `-4020` FTP upload failed, …).
### `bind_detect` results now reach the host
`detect` is an out-parameter (`detectResult&`). pybind11 casts a
reference argument to an override with a copy, so a plugin that filled
in `detect` wrote to a throwaway object and the host always saw an empty
`detectResult`. It is now passed so that Python edits the caller's
struct. Plugins see the same `DetectResult` argument as before.
## TODO
- Expose the `BAMBU_NETWORK_*` return codes to Python (the
`orca.printer_agent` binding and the generated stub from
`scripts/generate_orca_python_stubs.py`). Plugins can already return
them, but only as hard-coded numbers.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
- New `tests/slic3rutils/test_plugin_printer_agent.cpp`: an agent whose
operations raise, one that omits them, and one that returns the wrong
type all answer like a missing agent, and the interpreter stays usable.
A working agent's answers reach the host unchanged, including
`request_bind_ticket`'s out-param and the fields a plugin writes into
`bind_detect`'s `detect`.
- The `bind_detect` check failed before the fix (`"" == "192.168.0.2"`)
and passes after.
- `slic3rutils` passes under `ctest` (144/144); full Release build clean
on Linux.
- End to end on Linux with a test plugin whose chosen operations raise
(`start_discovery`, `get_filament_sync_mode`, `disconnect_printer`):
selecting the plugin's agent in the printer preset and switching back
logged each raise as a `Printer agent plugin '…': <operation> failed`
line, and the app kept running and closed cleanly (exit 0). Without the
guard, the first raise (`start_discovery`, on selecting the agent) ended
the app with `Uncaught exception` and SIGABRT (exit 134); that run used
a build whose printer-agent files are identical to `main`.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
pybind11 copies a reference argument to an override, so a Python
printer agent that filled in detect wrote to a throwaway object and the
host always saw an empty detectResult. Pass it as a pointer so Python
edits the caller's struct.
Add the one-all-printer-preset-per-product rule to the orca-profiles skill:
color is chosen at runtime, a material family is a new product and a color is
not, and CI does not catch per-color presets so it stays a review call. Note
that @System is the all-printer convention rather than an enforced check.
orca.host.ui.create_dock_panel(html, title, width, height, on_message,
on_close, dock) hosts plugin HTML in a pane of the Plater's dock manager,
next to the sidebar, and returns a UiDockPanel handle
(post/show/hide/close/is_open). The arguments follow create_window(). The
panel uses the window.orca bridge of plugin windows, restores its position
and size from the saved window layout, hides with the Plater off the Prepare
and Preview tabs when floating, and is closed with its plugin; plugin panes
are removed in MainFrame::shutdown().
The web view hosting moves out of PluginPage into a shared WebPanel base:
bootstrap page and swap to the plugin HTML, theme, element-default and
bridge scripts, window.orca message parsing, delivery to the page, and live
re-theming, also re-applied on every load after the swap. Pages tabs and
docked panels both derive from it. Pages tabs now re-theme in place on a
theme change instead of being reloaded, and a window.orca call a host does
not support is logged.
What the hosts share no longer lives in one of them: the bootstrap page, the
base URL and the plugin-window bridge move to Widgets/WebHosting, used by
WebDialog and WebPanel alike. The Plater restores plugin panes with a new
saved-layout parser, GUI/AuiPaneLayout, kept in its own small header so
slic3rutils can test it without pulling in the Plater.
The web hosting classes carry no plugin name, so other hosts can reuse them:
PluginWebDialog becomes WebDialog (its bootstrap page moves to
resources/web/dialog/WebDialog), and destroy_for_plugin(),
load_plugin_content() and plugin_defaults_user_script() become
destroy_silently(), load_page_html() and element_defaults_user_script().
Includes a sample plugin (sandboxes/orca_dock_panel_plugin_any.py) and
binding and layout-helper tests in slic3rutils.
IPrinterAgent callers do not catch, so a Python raise, a missing
override or a wrongly typed return from a printer agent plugin escaped
into the GUI. Each trampoline operation now logs the failure and
answers with NetworkAgent's no-agent value: -1 for status codes, the
empty value otherwise.
* Add PlastAR and Printalot filament vendors
Two Argentine brands from Printalot: PlastAR (budget PLA) and Printalot
(ABS). Each has a tuned @base plus per-color presets covering the
manufacturer's color lineup.
PlastAR PLA: flow 0.98, 1.75 mm, nozzle 220 C (190-230), bed 60 C; 14 colors.
Printalot ABS: flow 0.94, density 1.05, retraction 0.2 mm, max volumetric
18 mm3/s, nozzle 250 C (240-270), bed 100 C; 14 colors.
Per-color presets inherit their @base (no per-color parameter changes).
Passes scripts/orca_extra_profile_check.py.
* Ship PlastAR and Printalot as a single Printalot filament vendor
Place both Argentine lines in one OrcaFilamentLibrary/Printalot bundle and follow the library's brand shape: a non-selectable @base plus one all-printer @System per product, which is where a library preset meant for every printer belongs. The former per-color presets are dropped, leaving PlastAR PLA and Printalot ABS with their PLA and 250 C ABS material values. setting_ids are minted by the tooling; filament_ids resolve through the @base roots.
---------
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
* Add Polymaker PLA Pro filament profile
Polymaker PLA Pro is present in the tree only as printer-scoped variants
under Snapmaker U1 and Anycubic Kobra S1, so it is invisible to every other
printer. This adds it to the Orca Filament Library as OGFPM020 so it is
selectable generally.
Values are taken from the manufacturer's published print settings and
cross-checked against the two existing vendor profiles:
density 1.23 g/cm3 both existing profiles agree
softening 55 C both existing profiles agree
nozzle 220 C (210-230) manufacturer's stated range
max volumetric speed 15 Snapmaker 15, Anycubic 16
Flow ratio is set to 0.96, matching the Snapmaker profile. Worth noting for
review: the two existing profiles disagree here -- Snapmaker 0.96, Anycubic
0.85 -- because flow ratio depends on the extruder as much as the filament.
Any single value in a vendor-neutral preset is a starting point users should
calibrate; 0.96 is closer to the generic PLA baseline than 0.85 is.
setting_id assigned by scripts/assign_vendor_setting_ids.py.
OrcaFilamentLibrary version bumped 02.04.00.03 -> 02.04.00.04.
* fix errors
---------
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
A coEnum config value has two representations: the typed ConfigOptionEnum<T> a
config cloned from the static classes carries, and the ConfigOptionEnumGeneric
that a config assembled from the option definitions creates - which is what a
preset, a project's own settings and the CLI all hold. imex_cfg_enum() accepted
only the first, so every IDEX/IQEX reader took the option default instead: a
printer saved as rear-left came back front-left in the settings, the bed zones
and the carriage markers, while the preset on disk still held rear-left.
Read the generic form too, keyed on the value map it carries, since only T's own
map yields a T. The last reader that matched on the coEnum tag alone and cast
across the two hierarchies now goes through the helper with everything else.
Declaring the three keys the static classes were missing is what lets a change
to imex_tool_layout invalidate the slice it moves, which it never did before.
The two that are only ever drawn stay out of that: a colour scheme and the
advisory margin bands do not reach a slice, so changing one must not discard it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
set_hover_tooltip records one string per frame, so the multi-material warning
replaced the mode tooltip instead of joining it, and hovering the icon on a
conflicted plate no longer said which mode was active or that clicking cycles
it. The two are composed into one string, paragraph separated.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The estimate no longer reports a tower for a single filament whose flush matrix
purges, so the comments that justify calling prime_tower_is_printed() instead of
reading a depth now cite what still holds: it reads neither enable_prime_tower
nor print_sequence. GCodeViewer's comments name render_scene(), the function
that replaced the render() they still pointed at, and the pass contract mentions
the toolhead boxes it draws.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* Add Meltingplot vendor profiles (CHX 350 + MBL series)
Adds the Meltingplot vendor bundle with 5 printer models:
- CHX 350 (880x422x943, RepRapFirmware/Duet)
- MBL 133/136/308/480 (RepRapFirmware/Duet)
Each model ships 0.4/0.6/0.8/1.0/1.2 nozzle variants. Layer height
presets span 40-60% of the nozzle diameter and deliberately avoid
heights that divide the Z screw lead evenly (CHX: 1204 ball screw,
4 mm lead; MBL: TR4x2 trapezoidal screw, 8 mm lead) so periodic
screw error does not repeat as banding.
All printers use 2.85 mm filament, so the bundle carries its own
filament library (PLA, StoneFil, ABS, TitanX, ASA, PETG, PA6 CF HT,
STYX 12, PP 9-2, PET-CF15, Extrudr PLA NX2 Matt, Igus iglidur J260)
with the 2.85 mm diameter pinned once in the vendor-local
fdm_filament_common base.
Machine limits mirror the printers' firmware configuration; machine
limit emission is disabled so the values only feed the time estimator.
* Add per-nozzle volumetric flow limits and nozzle size in filename
Cap filament volumetric speed per nozzle (0.4: 14, 0.6: 20, 0.8: 26,
1.0: 30, 1.2: 40 mm3/s). Filaments whose limit exceeds a nozzle's
capability get a nozzle-scoped child profile that only lowers
filament_max_volumetric_speed; filament_id is inherited so all nozzle
variants remain the same material.
Also include layer height and nozzle diameter in the g-code filename.
* Set CHX 350 machine time cost to 5 EUR/h
* Inherit Meltingplot filaments from OrcaFilamentLibrary generics
Per the profile creation guide, vendor filament profiles now inherit
from Generic * @System instead of vendor-local fdm_filament_* copies.
Each profile carries only Meltingplot-specific deviations plus the
2.85 mm filament diameter; the seven local fdm_filament_* base files
are removed.
Deliberate pins where the library default would change tuned behavior:
- PLA/ABS/TitanX keep filament_max_volumetric_speed 20 (0.6 nozzle cap)
- ABS/PET-CF15 keep close_fan_the_first_x_layers 1
- STYX 12 sets required_nozzle_HRC 0 (unfilled PA12, density 1.02)
* Remove Meltingplot StoneFil filament profile
* Meltingplot ABS: use library fan settings
* Meltingplot TitanX/ABS: use library fan and nozzle temperature settings
* Remove STYX 12, PP 9-2 and Igus iglidur J260 filament profiles
* Meltingplot ASA: use library fan_max_speed
* Derive per-nozzle volumetric speed caps from physical limit formula
Caps now follow V = 34*d * m / S: the 34*d limit line calibrated on
Micro Swiss Volcano plated copper at 2.85 mm, material factor m from
the OrcaFilamentLibrary generic ratios (PLA/ABS/ASA 1.0, PETG 10/12,
PETG-CF 11.5/12, PA-CF 8/12, PLA Matte 11/12), safety factor S = 1.25
covering batch/moisture/wear/diameter tolerances.
Each material base now serves the 1.2 nozzle at its maximum cap;
children @0.4-@1.0 lower the cap per nozzle. Machine variant
default_filament_profile entries point at the matching preset.
* Scale slow_down_layer_time per nozzle from bead cross-section
Minimum layer time follows the derived cooling rule: heat per layer
scales with bead cross-section (layer height x line width), so the
0.8-nozzle-tuned values are scaled by q(d)/q(0.8) using each nozzle's
standard preset geometry. Bases carry the 1.2 value, children @0.4-@1.0
their nozzle-specific value.
* Meltingplot: lift Z above the print before the end macro
Add a G90 / G1 Z{min(max_layer_z + 5, printable_height)} F600 move at the
start of machine_end_gcode, before M98 print_end shuts the printer down and
parks it. Besides clearing the part, the explicit Z coordinate lets the Duet
file-info parser pick up the object height.
* Meltingplot: cap default layer height at 40% of outer wall width
A 45 deg slope shifts each bead outward by exactly one layer height. With
the Slic3r bead model (rectangle (w-h) x h plus half-circles of diameter h)
the flat contact band to the layer below is w - 2h, so the maximum
self-supporting angle is atan((w-h)/h). Support cannot compensate: the
stair tread at 45 deg is only h wide, far below one support line width, so
no support line ever fits underneath a single step.
All ten machine variants defaulted to h/w = 0.455..0.524, i.e. at or past
the h/w = 0.5 point where the flat contact vanishes. The 0.8 nozzle default
(0.44 mm at w = 0.84) reached atan(0.4/0.44) = 42.3 deg and failed on 45 deg
slopes even with support.
Default presets now satisfy h <= 0.40 * outer_wall_line_width, which leaves
one third of the bead's flat underside in contact at 45 deg and puts the
geometric limit at 56.3 deg. To keep the defaults on sensible ladder rungs,
outer wall width is widened where it was the binding constraint:
chx_small 110 -> 115%, chx_large 105 -> 110%, mbl_04 105 -> 115%.
Ladders are unchanged; the thicker rungs stay available for parts without
steep overhangs.
* Meltingplot PLA: raise slow_down_layer_time by calibration factor 1.5
Print feedback on CHX 350 with a 0.8 nozzle at 0.44 mm layer height showed
curling at the retraction points with the derived value of 25 s. There is no
cooling headroom left to trade: fdm_filament_pla pins fan_min_speed =
fan_max_speed = 100, and additional_cooling_fan_speed is inert because no
Meltingplot machine enables auxiliary_fan. Minimum layer time is the only
remaining lever; 35 s prints cleanly at that geometry.
The cross-section rule only fixes the ratios between nozzle sizes, so this
is a calibration of the PLA anchor itself: both PLA ladders are scaled by
1.5 (Extrudr 9/20/38/56/83 s, Meltingplot PLA 8/17/30/45/66 s for
0.4/0.6/0.8/1.0/1.2). The Extrudr base fan_cooling_layer_time goes 60 -> 100
so the 1.2-nozzle value stays below it.
Other materials are left unchanged - only the PLA anchor has a real
counter-example so far.
* Meltingplot: drop layer heights that cannot print 45 degree walls
A preset whose layer height exceeds ~45% of the outer wall line width has
less than ~6 degrees of reserve on a 45 degree wall (max self-supporting
angle atan((w-h)/h)), and support cannot compensate because the 45 degree
stair tread is only h wide - far below one support line. Those presets are
not usable in practice, so shipping them only invites failed prints.
Removes the 22 rungs above h/w = 0.45 and prunes them from the vendor
index. What remains per nozzle (default first):
CHX 350 0.4: 0.18 | 0.6: 0.24 0.28 | 0.8: 0.32 0.36
1.0: 0.44 0.48 | 1.2: 0.48 0.56
MBL 0.4: 0.18 | 0.6: 0.24 0.28 | 0.8: 0.36 0.34
1.0: 0.44 0.48 | 1.2: 0.48 0.56
The 0.4 nozzle keeps a single rung: 0.20 mm would qualify (h/w = 0.435) but
divides both screw leads evenly (4/0.20, 8/0.20) and stays excluded by the
Z-artifact rule, and 0.22 is already at 0.478.
Also corrects the MBL 0.8 default to 0.36 mm, which at w = 0.92 sits at
h/w = 0.391 and therefore still satisfies the stricter 40% default rule.
* Meltingplot: require 5 degrees of reserve on 45 degree walls
Replaces the two separate thresholds (h <= 0.40 * outer wall width for
defaults, 0.45 for ladder rungs) with a single requirement stated on the
design constraint itself: every shipped preset must reach
a_max = atan((w - h) / h) >= 50 deg
i.e. 5 degrees of reserve on a 45 degree wall, equivalent to
h <= 0.456 * outer_wall_line_width. The default is simply the thickest rung
that passes.
The 11 degrees the 40% rule produced were more margin than the failure mode
warrants, and cost throughput for no return. For calibration: across 940
machine defaults from 60 vendor bundles in this repo, the median reserve is
+1.8 deg, 71% ship less than 5 deg, and 8% cannot geometrically produce a
45 deg wall at all.
Adds the rungs the relaxed threshold makes available - 0.30 (0.6 nozzle,
both series) and 0.38 (0.8 nozzle, both series) - and restores the 0.30 mm
MBL preset removed in 5b4862b3 with its original setting_id. 0.20 mm on the
0.4 nozzle and 0.50 mm on the 1.0 nozzle would also qualify but stay
excluded because they divide both screw leads evenly.
Defaults are now (0.4/0.6/0.8/1.0/1.2) 0.18/0.30/0.38/0.48/0.56 for both
series, at 50.2 to 57.3 deg. The 1.2 nozzle deliberately keeps 0.56 rather
than the 0.60 the rule would allow: solidification time scales with h^2 and
bead weight with h*w, so the largest nozzle is the one that should not sit
on the minimum.
* Meltingplot: disable precise wall by default
Precise wall was inheriting OrcaSlicer's global default of true. Disable it
bundle-wide in the vendor process root so it applies to CHX 350 and all MBL
models. Note it was already ignored on the CHX small-nozzle branch, which
uses the inner-outer-inner wall sequence.
* Meltingplot CHX: raise max extruding acceleration to 6000
The 4000 limit silently clamped the profiles' 6000 values for default and
inner wall acceleration, since GCodeWriter caps print acceleration against
machine_max_acceleration_extruding regardless of emit_machine_limits_to_gcode.
Raising it to 6000 matches the X/Y limits so the configured values are emitted
as-is. The firmware's own M201 still governs what is actually driven.
* Meltingplot PA6 CF HT: raise part cooling floor to 40%
Sharp 90 degree corners were washing out layer by layer while straight walls
and overhangs printed cleanly. Overhangs are fine because overhang_fan_speed
forces 100% regardless of layer time; everything else fell back to
fan_min_speed, since at the layer times seen in practice neither branch of the
layer-time fan interpolation applies.
Raise fan_min_speed to 40% and set reduce_fan_stop_start_freq so that floor is
actually applied instead of dropping to zero. Also lower the auxiliary fan to
40% and raise the overhang cooling threshold to 95%, matching the values
validated on the print.
* Meltingplot: re-mint filament ids with orca_id_tool
Upstream now requires every filament_id to be a minted "OF" id derived
from the (filament_vendor, filament_type, name) triple and recorded in
scripts/filament_id_snapshot.json. The hand-made ids this bundle shipped
with (MPFPLA0, EXTNX2M, ...) were rejected by orca_extra_profile_check.
Regenerate them with "orca_id_tool.py --generate --vendor Meltingplot"
and record the result with --update-snapshot. No setting_id changed, and
the ids were never released, so no existing project file refers to them.
* Meltingplot: normalize bundle for updated profile checks
Drop obsolete keys (silent_mode, adaptive_layer_height, tree_support_with_infill) and rebuild the vendor index in canonical order, as orca_profile_tool.py normalize / update-index write them.
* Meltingplot: list per-nozzle filaments in default_materials
The unsuffixed filament presets only cover the 1.2 mm variants, so the 0.4-1.0 mm printer presets had no compatible entry in their model's default_materials. The validator's check_printer_default_materials rejects that, and load_installed_filaments would auto-install no filament for those variants on first run. Name the @0.4/0.6/0.8/1.0 nozzle presets alongside the base ones for PLA, PETG, ABS and ASA.
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
* Add Flashforge Adventurer A5 printer profiles
Adds system profiles for the Flashforge Adventurer A5 to the existing
Flashforge vendor: the machine model with 0.25/0.4/0.6/0.8 mm nozzle variants,
14 process presets, and 43 filament presets, plus the printer cover image.
All presets inherit from existing Flashforge bases already in-tree
(Adventurer 5M Pro nozzle profiles, AD5M Pro process/filament presets,
Flashforge generics); no new base assets are introduced (the A5 reuses the
existing 5M-series bed/hotend models and buildplate texture). Every inherits
reference and sub_path was validated to resolve.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Assign setting_id to Adventurer A5 presets (fix profile check)
The new A5 filament and process presets were missing the deterministic
setting_id required by OrcaSlicer's profile validator. Generated via
scripts/assign_vendor_setting_ids.py (61 instantiated presets). Verified with
scripts/orca_extra_profile_check.py: 0 errors.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* Fix Flashforge Adventurer A5 profiles so the bundle loads and validates
The new A5 presets were not in canonical form, several filaments resolved
no or the wrong filament_id, and the model and machines pointed at
non-existent or AD5M Pro defaults. Run the standard profile tooling
(normalize, update-index, generate-id), give the branded filaments the ids
of their existing Flashforge products, and repoint the defaults at
A5-compatible filaments and processes. Bump the Flashforge bundle version.
---------
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Co-authored-by: yw4z <ywsyildiz@gmail.com>
Co-authored-by: SoftFever <103989404+SoftFever@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
The estimate now returns no tower for a lone filament whatever the flush matrix
says, so the discrepancy this case pinned between it and normalize_fdm_2 is gone
and the sibling case covers what remains.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The merge of upstream main (dc021c6ef6) resolved the GCodeViewer.cpp conflict
from #15674 by keeping both sides. That left a stray `}` in
SequentialView::render_overlay, so the file no longer compiles. It also kept
the IMEX carriage update and the toolhead-box GL draw in the ImGui overlay
pass, which #15674 no longer uses for 3D geometry.
The carriage update and the toolhead boxes now run in render_scene next to the
primary marker, and render_marker draws the secondary markers. render_overlay
goes back to what upstream has, without the duplicated marker-position block
and the unused bottom_margin the merge left behind.
The same upstream change renamed PartPlate::show_tooltip to set_hover_tooltip;
the two IMEX call sites follow it.
A layer split into ~1000 colour fragments (a colour texture baked over a
large top face) made several per-fragment loops redo whole-layer ClipperLib
work, so slicing took ~33 min; it now takes ~3 min with the same output.
- make_fills: clip the layer's no-overlap area to each expolygon's box
before intersecting
- discover_vertical_shells: small-piece filter compares only against the
nearby part of the layer
- bridge_over_infill: whole-layer union/diff/intersections restricted to the
candidate's neighbourhood; fill boundary expanded once per spacing; anchor
tree built only from lines crossing the scan range; bbox pre-check in the
collision test; limiting outline taken directly instead of through
expand(..., 0.3 * flow.spacing()), which offsets by 0.135 scaled units
(flow.spacing() is in mm) and only cost a whole-layer pass per candidate
- Bake job: include GUI.hpp for show_error()
- Texture panel: select/erase whole model next to the paint tools, brush
size slider back on the same row
- GMP: name the parameters in the configure probe from
0001-GMP_GCC15.patch. Unnamed parameters are a hard error on GCC 10
and older, so configure failed there.
- TBB: turn off the automatic hwloc search so a system libhwloc-dev is
not picked up.
- scripts/linux.d/debian: run apt update before checking which
webkit2gtk package exists, and prefer webkit2gtk-4.1 since the build
requires it. Add automake, which MPFR's autoreconf needs.
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
When a PR is merged using the bot with `/bot merge`, it doesn't invoke
`post_merge_profiles.yml` because post_merge_profiles.yml only starts on
a push or manual dispatch.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Plugin dialog content is injected with SetPage, so on the WebKit backends a
reload (context menu, keyboard shortcut or location.reload()) re-fetches the
SetPage base URL instead of the injected document, and the plugin UI is gone
for good: load_plugin_content() returned early once m_content_loaded was set.
Re-inject the plugin HTML when a main-frame load after the initial swap is
neither that swap nor a page the plugin linked to. m_own_page_load marks the
load our own SetPage caused, and the URL test recognises the reload: the
injected document and the directory a reload re-fetches both report the base
URL, so a load of any other URL is left alone. The test ignores a fragment the
page navigated to, and undoes the escaping the web view applies to what the
resources path holds.
A load reaching the base URL is not enough on its own, because WebKitGTK reports
a navigation that never committed against the document that stayed and then
finishes that document again: a link to a missing file therefore arrives as a
load of the base URL and reads exactly like a reload. So the re-injection also
requires a navigation to the base URL to have committed, which a reload always
does and a failure never does.
A bootstrap page that cannot be loaded is still not recovered from: WebKitGTK
substitutes a stock error page for it, and that load supersedes the swap
whichever way the swap is ordered around it. The file ships, so this is a
broken-install path; nothing here makes it worse than it already was.
No separate MSW path is needed: wxWebViewEdge ignores the SetPage base URL, so
its documents report about:blank and the test never matches there, and WebView2
reloads NavigateToString content from its own history entry anyway.
- Bake: each layer is sampled only on its own painted area; analytic
projections used to stack every layer over every painted region, so the
top layer's texture showed on all of them (colour sampler too)
- Auto resolution follows the texture's texel size and sharpness again
- Unwrap: charts cut by each face's own normal (a cube gives 6 islands, not
12 triangles); non-disk charts (tubes, closed shells) are split until they
flatten; connected nets test real triangle overlap, grow from the largest
chart and are packed side by side
- UV edits are stored per unwrapped copy, so dragging a seam vertex no
longer moves its copies in neighbouring islands
- UV pane: tool strip with unwrap settings moved in from the panel, sharp
HiDPI icons, clearer island/edge/selection drawing with hover, texture
picker from the thumbnail, texture no longer lost on reopen (GL state
from the 3D view, background upload retries)
- Panel: whole-model select/erase as icons in the tools row; inactive
layers' paint shown muted; colour textures shown in colour in the picker
- Built-in displacement texture library
- Tests for unwrap segmentation, connected nets, UV edits and per-layer
sampling
* Add Ultimaker S5 Profile - WIP
* Add Ultimaker physical printer base
* Ultimaker API WIP for base update.
* Fix undefined reference error.
* Add UMS5 profile, add placeholders for testAuth stuff.
* Fix non-const func definitions, make func names align with style guide.
* Localization stuff? IDK if this does anything or is required.
* Add cover image
* Auth code cleanup, add UI button for auth cred generation, implement various auth checks and tests.
* Fix auth stuff
* Clean up code
* Get upload code sort-of working, fix typo
* Update printer settings and start/end gcodes
* Add makeGriffinCompatible preprocessor script to prevent machine crash, update machine profile.
* Fix buildplate size, fix time missing bug, add S5 buildplate model
* Correct capitalization to UltiMaker
* Fix display bug, fix capitalization bug
* Implement credential generation button and logic
* Fix generate auth creds button, add todos, fix capitalization.
* Actually fix generate auth credentials.
* Fix generate auth creds message.
* Update UM S5 machine limits
* Revert accidental commit.
* Fix postprocessor, clean up code.
* Update presets for multi-extruder printing.
* Add Single Extruder and Fast profiles.
* Code cleanup
* Register and validate the UltiMaker S5 presets
* Show the Generate API Key button only for UltiMaker print hosts
The PR added it to the Physical Printer dialog for every host type, where pressing it runs an ordinary connection test and reports "API Key created". Gate it on the selected host type instead.
* Drop unused lambda captures in the UltiMaker host code
clang promotes -Wunused-lambda-capture to an error under the project's -Werror, so the file failed to build on macOS; the five callbacks do not touch this.
* Fix the Windows build of the UltiMaker print host
---------
Co-authored-by: SoftFever <softfeverever@gmail.com>
The validator, fixture archives and unpacked fixture trees no longer live under <repo>/.test/check_profiles, so every worktree on a machine shares one copy instead of re-downloading. macOS, Linux and Windows each use their own user cache dir, named apart from OrcaSlicer's per-user dirs. The now-unused /.test/ ignore rule is dropped.
# Description
A plate using one filament no longer reserves or draws a prime tower in
the plater just because multiple filament slots are configured. The
shared footprint estimate gated "no tower" on the purge volume being
zero, but on non-Bambu printers with the shipped defaults that volume
comes from the SEMM flush matrix, which reads every configured slot and
stays nonzero even when only one filament is used — so the preview
showed a tower the print could never contain. The estimate now decides
from the filament count alone, with wrapping detection and smooth
timelapse staying the only reasons a single-filament plate keeps a
tower, matching `Print::has_wipe_tower()`. The gate also moved above the
flush-volume scan so the lone-filament path no longer pays for it.
No change to slicing output — toolpath generation already omitted the
tower for a single used filament; this removes the phantom preview,
placement reservation, and validation footprint.
# Screenshots/Recordings/Graphs
Before the fix:
https://github.com/user-attachments/assets/c5b0ea61-127d-408e-bafe-f1f76c58b996
After the fix:
https://github.com/user-attachments/assets/ac6632b3-5d44-4013-aca3-73f6cfd1139c
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
Added a shipped-defaults single-filament case to
`tests/libslic3r/test_wipe_tower_estimate.cpp`: it estimates a 5.04
mm-deep tower before the fix and zero depth after. `[WipeTowerEstimate]`
and the fff_print `[WipeTower]` suites pass.
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The no-tower case was gated on there being no purge volume, but the SEMM flush matrix reads every configured slot and is nonzero even when only one filament is used, so the plater preview drew a tower the print would never contain.
* Fix Publish 3MF dialog styling
Use Orca's shared checkbox widget in the Publish 3MF dialog. Keep checkbox labels clickable, preserve toggle event propagation and disabled-row state, and use Windows-only double buffering on the tab page hosts to reduce flicker during page switches.
Also align the dialog's guide-link color with existing Orca dialogs.
* Fix Publish 3MF dialog styling
Use Orca's shared checkbox widget in the Publish 3MF dialog. Keep checkbox labels clickable, preserve toggle event propagation and disabled-row state, and use Windows-only double buffering on the tab page hosts to reduce flicker during page switches.
Also align the dialog's guide-link color with existing Orca dialogs.
* match font size and left margins
---------
Co-authored-by: yw4z <ywsyildiz@gmail.com>
* init
* drop usage of StaticGroup for ExtruderGroup
* completely remove StaticGroup from project
* fix alignment of "Not installed" text
* fix crash on linux while clicking edit button
* Fix background color on macOS
---------
Co-authored-by: Noisyfox <timemanager.rick@gmail.com>
Merged by /bot merge on behalf of @inslogic3d (id 321604763).
Grants: resources/profiles/OrcaFilamentLibrary/filament/INSLOGIC, resources/profiles/OrcaFilamentLibrary.json
Head: 22a51b32f0
libslic3r_version.h defines ORCA_CHECK_GCODE_PLACEHOLDERS from the CMake
option, 0 by default, and GCode.hpp then forces it to 1 whenever NDEBUG
is undefined. Clang reports the second definition under
-Wmacro-redefined, which is on by default, so since -Werror (#15660)
every clang Debug build fails at the libslic3r files that include
GCode.hpp. Release and RelWithDebInfo define NDEBUG and never compile
the override, which is why no CI configuration sees it. The override
dates from #3861 and has warned in every Debug build since.
Undefine the macro before overriding it. Debug builds get the same
value the redefinition already produced, so nothing else changes.
Follow-up to #15749, refs #15748.
fix(build): drop two unused lambda captures that break non-Release builds
Both handlers capture this and never use it. They sit inside
#if !BBL_RELEASE_TO_PUBLIC, which CMakeLists defines as $<CONFIG:Release>, so
the code compiles in Debug and RelWithDebInfo but not in Release. Clang warns
on an unused capture under -Wall, and since every warning became an error the
two captures fail any clang build that is not Release - an Xcode scheme on its
default Debug configuration, for instance. The CI matrix builds Release, where
the block does not exist, and GCC does not implement the warning at all, so
nothing in CI can see it.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
On windows, after installing and loading a plugin, if you try to
unsubscribe from a loaded plugin, and on refresh, it will be an orphaned
plugin.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
On unsubscribing from a loaded cloud plugin
<img width="1418" height="862" alt="image"
src="https://github.com/user-attachments/assets/9ba8353c-1645-4b21-86ff-2ed235f4d259"
/>
On plugin refresh
<img width="1418" height="862" alt="image"
src="https://github.com/user-attachments/assets/fd58c144-7970-4448-ba3b-e72e52980467"
/>
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
Colour height maps now displace the same way up as grayscale ones. A
truncated or corrupt PNG in the texture folder or a project file now
loads as an empty texture instead of aborting the app, and a 16-bit
colour PNG is converted on load rather than displacing as noise.
Emit U1 nozzle flow metadata through profile end G-code
Append filament_volume_type footer comments through the standard and High Flow U1 machine profiles. Generate one comma-separated value per logical filament using existing placeholder expressions, preserving the existing end commands. Mixed-diameter standard profiles inherit the standard footer.
Bump the Snapmaker bundle to 02.04.00.17. No C++ or filament-schema changes. Validation intentionally not run at user request.
Co-authored-by: Codex <codex@openai.com>
Select purifier mode in U1 start G-code using the existing min_vitrification_temperature placeholder. Cover standard, high-flow and mixed-nozzle profiles without new slicer settings or C++ changes. Document the thresholds; retain the branch-wide Snapmaker version bump.
Co-authored-by: Codex <noreply@openai.com>
Add separate HF presets for 0.2, 0.4, 0.6 and 0.8 mm nozzles, preserving
standard printer settings and machine G-code. Provide 141 HF filament
presets with volumetric ceilings 3x their non-HF baseline. Generic presets
inherit active Orca material definitions; remove unregistered Snorca-only
Generic profiles and their obsolete PLA Wood dependent.
Add 36 diameter-specific HF processes, all using Snorca's 0.4 mm HF speeds:
600 mm/s inner walls, 500 mm/s outer walls, and 600 mm/s solid/sparse infill.
Preserve other process settings and keep standard-nozzle speeds unchanged.
No extra-high-flow tier or C++ changes are included.
Native profile validation passed. HF material slicing checks passed, and
all 36 HF processes sliced with verified speed settings and tool changes.
Document calibration sources, flow ceilings and remaining physical tuning.
Generate canonical preset IDs and record the eight new Snapmaker generic
claims in the filament ID snapshot. Regenerate the vendor index.
Co-authored-by: Codex <noreply@openai.com>
# Description
Adds the CI half of the profile OTA pipeline: a push-triggered workflow
that
rebuilds a vendor's binary preset cache (`<vendor>.opc`) whenever its
profile
changes on `main` / `release/*`, and publishes it as a versioned release
asset
for OrcaCloud's OTA Manager to pick up.
### `.github/workflows/post_merge_profiles.yml` (new)
Push-triggered counterpart to `check_profiles.yml` (which only gates
PRs):
- Diffs the push to find which vendors under `resources/profiles/**`
changed.
- Reads the Orca version from `version.inc` and each vendor's 4-part
`version`
from `resources/profiles/<vendor>.json` (fails the run if it isn't
`A.B.C.D`).
- Downloads the prebuilt `generate_system_cache` from this repo's
`nightly-builds` release and builds one `<vendor>.opc` per changed
vendor.
- Packages each as
`<orca_ver>_<vendor>_<profile_version>_<UTCyyyymmddHHMM>.zip` (zip root
`<vendor>.opc`) — the asset-name contract OrcaCloud's release scanner
expects.
- Uploads them to a per-Orca-version release on the profiles repo via a
scoped
GitHub App token.
It stops there: no changelog, no R2, no OTA webhook — a maintainer still
publishes from the OTA Manager. Job is guarded to
`OrcaSlicer/OrcaSlicer`;
workflow permissions are `contents: read` (the cross-repo write uses the
App
token only).
### `.github/workflows/build_orca.yml`
Two steps on the Linux leg: upload `generate_system_cache` as a CI
artifact,
and (on `main`) deploy it to the `nightly-builds` release as
`generate_system_cache_Linux_Ubuntu2404_nightly` so the workflow above
has a
tool to download. `.opc` is 64-bit little-endian and
platform-independent, so
only the Linux binary is shipped.
### `src/dev-utils/generate_system_cache.cpp`
New `-v` / `--vendor` option to generate the cache for a single vendor
(plus the
always-loaded Orca filament library) instead of all vendors, and an
error if the
named vendor produced no `.opc` (catches typos). Reuses
`PresetBundle::set_vendor_to_validate()` from #14217.
## Operational prerequisites
- Repo secrets `PROFILES_APP_ID` / `PROFILES_APP_PRIVATE_KEY` for a
GitHub App
with `contents: write` on the target profiles repo.
- `env.PROFILES_OWNER` / `env.PROFILES_REPO` in
`post_merge_profiles.yml` must
point at the production profiles repo OrcaCloud reads.
- `generate_system_cache_Linux_Ubuntu2404_nightly` only appears after
the first
post-merge nightly `build_orca` run; a profile-only push before then
fails at
the download step.
## Screenshots/Recordings/Graphs
n/a — CI only.
## Tests
Exercised end-to-end in a staging environment: per-vendor `.opc` zips
published
in the expected `<orca_ver>_<vendor>_<profile_version>_<timestamp>.zip`
layout,
and `generate_system_cache -v <vendor>` confirmed to emit only that
vendor's
`.opc`.
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
# Description
Adds a sketch-first parametric CAD tab to the slicer: sketch → constrain
→ solid features
→ commit to plate. The feature recipe is persisted inside the 3MF, so
reopening a project
restores an editable model rather than a frozen mesh.
Opening this at @SoftFever's request, so the code is easier to read than
a fork.
**The number worth reading first:** the diff is large, but almost all of
it is new files.
Existing upstream code is touched in **23 files, +622 / -88 total**.
That is the entire
negotiable surface. The largest single one is `GLCanvas3D.cpp` at
+149/-14 (a pick path for
the CAD viewport); everything else is under 60 lines.
One thing about the raw diff: the file count includes everything new,
and the negotiable
surface is the 23 modified files above. Thanks for merging `main` in —
the branch is current
again, and I have kept building on top of it.
The regenerated i18n catalogues (`OrcaSlicer.pot`, `OrcaSlicer_it.po`,
`list.txt`) have been
kept OUT of this branch deliberately — they were 27,314 added lines of
build product standing
between you and the code. They regenerate from source with
`scripts/run_gettext.sh` whenever
you want them refreshed. A Romanian catalogue that had been riding along
was pulled out at
the same time — a translation has no business being reviewed inside a
CAD feature PR.
| | |
|---|---|
| Kernel | OCCT — already linked for STEP import. The dependency delta
is one line: `BUILD_MODULE_ModelingAlgorithms=OFF → ON`. Measured cost
in
[`docs/cad_dependency_weight.md`](https://github.com/tommasobbianchi/Orca-Cad/blob/cad-mainline/docs/CAD/cad_dependency_weight.md)
|
| Constraint solver | vendored SolveSpace `libslvs` subset, 21 files /
~10k lines under `src/libslic3r/slvs/` |
| Build gate | `SLIC3R_CAD` (default ON). With it OFF the tab is not
compiled and the deps prefix matches upstream exactly |
| Persistence | CAD recipe embedded in both the 3MF and BBS-3MF writers
|
| User docs |
[`docs/design_tab.md`](https://github.com/tommasobbianchi/Orca-Cad/blob/cad-mainline/docs/CAD/design_tab.md)
|
| Interaction model | object-driven — point at geometry, it offers the
verbs that apply:
[`docs/cad_ux_guidelines.md`](https://github.com/tommasobbianchi/Orca-Cad/blob/cad-mainline/docs/CAD/cad_ux_guidelines.md)
|
### Why it belongs in the slicer
Every round trip through an external CAD tool costs an export, a
re-import, and the design
intent both steps discard. A part changed after slicing should come back
to its feature
history, not to a mesh. Keeping the model in the slicer preserves that
loop — nozzle
diameter, build volume and material are known at design time. Longer
argument in
[`docs/design_tab_upstream_portability.md`](https://github.com/tommasobbianchi/Orca-Cad/blob/cad-mainline/docs/CAD/design_tab_upstream_portability.md).
### Two things I'd rather you hear from me than find
**Licensing.** The vendored solver is **GPL-3.0**, not LGPL
(`src/libslic3r/slvs/LICENSE`).
The combined work is distributable under AGPL-3.0 and the compatibility
argument is written
out in the portability doc, but this is a project-level decision and I
would like it
confirmed explicitly rather than assumed. If GPL-3.0 in-tree is not
acceptable, the solver
is the separable part — the timeline, features and persistence do not
depend on it.
**One CMake change is larger than it looks.** `CMakeLists.txt` is
+41/-56: it replaces a
hand-maintained list of OCCT DLLs to copy on Windows with a glob plus an
assertion that
every linked toolkit actually has a DLL. The explicit list had already
drifted from what
`libslic3r` links and shipped a portable that died at launch with `error
126`. Happy to
split that out into its own PR if you'd prefer it reviewed separately.
### Not verified
- No automated GUI test. A green kernel run says nothing about the
viewport — synthetic
clicks never drift, so the suite and the UI are two separate realities.
- Card wiring for 9 of the 16 late-wired tools has never been
click-tested.
- The click-test defect rate has not converged: one pass found nothing,
four further days
of work found five more defects. I would not present the quiet pass as
evidence of
stability.
# Screenshots/Recordings/Graphs
One part, start to finish: sketch it, feature it, print it — without
leaving the slicer.
**1. Sketch, constrained and dimensioned.** A 100 × 90 rounded rectangle
drawn straight onto
the bed, R20 corners, live dimensions, and the solver's remaining
degrees of freedom reported
in the panel. The bed is the sketch plane, so the part is sized against
the machine it will be
printed on from the first line.
**2. The feature tree is the part.** `Sketch1 → Extrude2 → Chamfer3 →
Sketch4 → Extrude5 →
Hole6 → Thread7`. Every step stays editable and re-evaluates downstream
— the modelled thread
in the boss is a real helical feature, not a texture.
**3. Committed to the plate.** The same body arrives in Prepare as
`Design Body`,
100 × 90 × 78 mm, 581,634 mm³, ready for a Sovol Zero and PETG. No
export, no re-import, no
lost design intent.
**4. Sliced.** The thread comes out as real helical toolpaths, and the
estimate is 3h26m /
134.54 g. This is the whole argument for the feature in one frame: the
geometry that was
parametric two screens ago is now G-code, and it is still parametric if
you go back.
## Tests
215 `TEST_CASE` blocks across 6 new test files, plus 2 `SCENARIO`s added
to
`tests/libslic3r/test_3mf.cpp` covering the CAD recipe's round trip
through both 3MF
writers. `scripts/kernel-test.sh` is the headless contract: it builds
only
`libslic3r_tests`, needs no display, and exit 0 means the CAD suite
passed.
Happy to slice this differently — kernel + solver first, GUI second — if
that reviews
better for you.
Merged by /bot merge on behalf of @JAYO3D-Official (id 320896770).
Grants: resources/profiles/OrcaFilamentLibrary/filament/JAYO, resources/profiles/OrcaFilamentLibrary.json
Head: fb9a6d70a0
Four files overlapped, and each resolution favours upstream where the two sides
had done the same work:
WipeTower's M104/M109 tool qualifier. Both sides bounds-checked the physical
extruder map lookup; upstream omits the T qualifier when the map cannot answer,
where this branch fell back to the logical index and so named a carriage that may
not be the one printing. Upstream's behaviour is what this branch documents
elsewhere, so its version is taken and the local helper is dropped.
get_extruders' mixed-slot switch. Upstream added the same concept to the CLI
overload as expand_mixed_slots, so the GUI overloads' parameter is renamed to
match rather than carrying two names for one idea.
GLCanvas3D's sequential-clearance branch gains upstream's
update_compacted_wipe_tower_clearance for the by-layer case.
The printer_agent re-sync in TabPrinter::reload_config was upstream's and their
preset-undo fix removed it, so it goes; the IMEX modes grid re-sync beside it
stays, since it spans three options and is not a Field.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
imex_wipe_tower_hull() open-coded "does this plate print a tower?" as one
filament and no forcing reason. normalize_fdm_2, which is what actually
clears enable_prime_tower before slicing, has a second arm the copy omitted,
and reads the mixed-filament flag project-wide where the copy read the
plate. Two cases went wrong in opposite directions.
A ByObject plate with several objects prints no tower and the scene draws
none, yet the copy validated one and could refuse the slice with "the prime
tower overlaps an area reserved for IDEX/IQEX parallel printing" - with
nothing on screen to move.
A plate using one plain slot while some other slot in the project is a blend
does print a tower, because normalize_fdm_2 keeps it for any mixed filament
in the project, and the copy skipped validation entirely, so the tower could
be placed in a carriage zone and sliced.
prime_tower_is_printed() in libslic3r states the rule once, and both the
gate and a test use it. The counts are the ones normalize_fdm_2 is handed:
filament slots as authored, so a mixed slot counts once, and distinct
objects rather than instances. filament_is_mixed is a project option, so it
is passed in rather than read from the print preset.
The test drives every combination the rule looks at and compares the verdict
against normalize_fdm_2 itself, so the two cannot drift again without
failing.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The block describing what imex_wipe_tower_hull() does and does not cover had
drifted. It named estimate_wipe_tower_size, which upstream replaced with
estimate_wipe_tower_footprint; it described the plate filament count as an
override where the rewritten estimate treats it as a floor; and it
documented a Type2 stabilization cone as unhandled when the estimate now
folds the cone's worst-axis bulge into the margin the hull is built from, so
a second allowance would double-count. The comment on the m_print arm of the
guard now says what that arm actually does, which is nothing, since the
estimate stopped reading m_print.
It also records why the gate takes "is a tower printed" from
normalize_fdm_2's rule rather than from the estimate: the estimate reports a
tower for a single filament whenever the flush matrix purges, which would
hard-block a plate whose tower normalize_fdm_2 had already cleared, with
nothing drawn on screen to move.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
With single_extruder_multi_material and purge_in_prime_tower, the flush
matrix replaces the prime volume and its average is non-zero for one
filament, so the estimate reports a tower. normalize_fdm_2 clears the tower
for that same plate, so none is printed: the estimate answers how big a
tower is, never whether there is one, and a caller reading a non-zero depth
as "a tower is printed" reserves space for, or blocks on, a phantom. The
test asserts that disagreement directly.
Type1 is the exception in the same case, because it decides from its
per-filament purge list and a lone filament is never changed to. The scene
reads the same estimate either way. Type2 is the default for every non-Bambu
printer, and so for every IDEX/IQEX one.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
nozzle_temperature is variant-expanded, so its length is columns rather than
filament slots. On the dynamic-nozzle path that makes it longer than the
slot count, and an out-of-slot index reached get_filament_config_index() and
came back as filament 0 - the clamp the bounds-checking was meant to remove.
Bound by the slot count instead; on the ordinary path the two are equal and
nothing changes. The is_extruder_used write gains the matching lower-bound
guard.
IMEXHelpers.hpp now states both halves of the rule its call sites follow.
Bound anything derived from the extruder map against the filament slot count
before using it as a filament id, not against the option about to be read.
And a miss is -1, which is a correct tool qualifier but matches no physical
head, so it cannot serve as a skip-the-primary sentinel: which head prints a
filament is answered by the filament and the map, never by a mode role,
since a primary-mode print may use any or all tools, one at a time.
The consequence is recorded there rather than left implicit. The two skip
sites skip nothing for a slot past the end of the map, so a plate with more
slots than nozzles double-writes the primary's pressure advance. It is
narrow and unreported, and a guard there would be a smaller change than
naming a head.
The header also records that RepRapFirmware sends an unqualified pressure
advance as M572 D0, naming drive 0 absolutely rather than the active tool.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
CameraPopup and StatusPanel each bind a toggle handler that captures this
and never uses it, inside an #if !BBL_RELEASE_TO_PUBLIC block. CMakeLists
defines that as $<CONFIG:Release>, so the block compiles in every
configuration except Release - and Release is the only one CI builds. Under
-Werror the two captures therefore fail RelWithDebInfo, which is what a
development build uses, while CI never sees them.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The parallel printing options shipped to testers as is_ixex and ixex_*,
became is_imex and imex_* when the feature stopped being called iXex, and
the two clearance keys were renamed once more to say what they measure:
nozzle to carriage edge on the collision side, not the carriage's full
width. Nothing translated any of it, so loading an existing printer profile
dropped every one of these values - the keys are unknown and cleared.
is_ixex is the one that matters most, because without it the others migrate
into a feature that stays switched off, leaving settings that look
configured and do nothing.
Per-plate mode is persisted twice and only one path went through
handle_legacy. Plate metadata in a 3MF is matched by exact attribute name
and written with set_key_value, so a project saved between the per-plate
mode landing and the rename loaded every plate back on Primary and sliced
single-carriage with no warning. The loader now accepts the old attribute
name.
ixex_primary_col and ixex_primary_row are dropped rather than mapped: the
primary is a role inside the mode's active-tools string now, not a grid
coordinate, and they were never in an option list, so no saved file carries
them.
The test drives the full era-1 key list and asserts the enum values rather
than non-nullness, since a forward-compatible substitution would otherwise
hide a failed deserialize behind a default. handle_legacy had no test before
this.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Two range-for loops bound const std::string& to braced lists of string
literals, so each iteration constructed a temporary to bind to. GCC 16
reports it as -Wrange-loop-construct, which upstream's blanket -Werror turns
into a build failure; clang does not report it at all. Spell the
initializers the way the loop above them already does.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Review findings on the preceding commit, plus one defect it should have
caught.
- The IDEX/IQEX pressure-advance loop fed resolve_filament_for_head()'s
result straight into enable_pressure_advance and pressure_advance. That
result is bounded by physical_extruder_map, which holds one entry per
NOZZLE, while both options are indexed per filament SLOT. On a printer
with more nozzles than the project has filaments the two spaces diverge
and get_at() clamped the overflow onto filament 0, emitting its pressure
advance on a secondary carriage. The second-layer temperature loop bounds the
same lookup, but against nozzle_temperature, which is variant-expanded and so
is not the slot count either -- it is not the precedent it looks like.
IMEXHelpers.hpp states the rule
once, and a test pins the contract that makes the bound necessary:
resolve_filament_for_head() answers in nozzle space, so a non-negative
result is not by itself safe to use as a filament id.
- The header claimed every caller renders a -1 tool qualifier as "emit
none". RepRapFirmware substitutes the historical D0 instead, deliberately
and with its own comment in GCodeWriter. Say so, rather than leaving a
contract a future author would code against.
- A cross-reference pointed at a hard-coded line number that the preceding
commit had itself shifted by nine lines. Name the function instead.
- The multi-color rejection reasons reach the user through Print::validate()
as raw English, while the returns on either side of them use L(). Wrap
them and register IMEXHelpers.cpp for extraction. They also still said
"IMEX", the internal name, so they move to IDEX/IQEX with the rest of the
user-facing strings rather than shipping the internal one to translators.
- Trim the preceding commit's comments. One block explained the same
clamping hazard six times; the canonical explanation now lives in
IMEXHelpers.hpp and the call sites point at it. The mode grid carried
twelve lines of commentary and no code, most of it archaeology already in
the commit message, and one claim about the modes editor that was not
true. The ArrangeJob threading note stays: it documents an invariant that
cannot be recovered from the code.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Review findings on the IDEX/IQEX parallel printing code, all in paths the
feature owns.
- physical_extruder_map lookups used ConfigOptionVector::get_at(), which
clamps an out-of-range index to values.front() rather than reporting a
miss. The map holds one entry per nozzle while filament ids index slots,
and nothing caps the slot count at the nozzle count, so a project authored
with more filaments than the printer has extruders silently addressed the
primary's head: pressure advance pinned to the wrong carriage, and
skip-primary loops suppressing whichever head sat at pem[0]. Bounds-check
at all four sites and treat the miss as "no mapping" (-1). Covered by a new
imex_pem_tool_for test; the header note now warns against get_at here.
- IMEXFilamentPickerPopover leaked a top-level window per ghost click:
wxPopupTransientWindow::Dismiss() only hides, and never reaches OnDismiss().
Destroy from an OnDismiss() override and dismiss the picker through
DismissAndNotify(), which is the path a successful pick takes.
- ArrangeJob read PartPlate's IMEX zone cache from the worker thread, where
a cache miss rebuilds GLModel members with no GL context current while the
GUI thread may be painting them. Snapshot the zones in prepare(), on the
main thread, already converted to plate-local coordinates.
- The mode grid anchored its row window to the Primary's gantry row. A window
as tall as the grid can only start at row 0, so this drew tiles for tools
that do not exist and hid real ones. Render the whole grid instead; a
Primary outside it is a data problem the zone layout already reports.
- Build the mode tooltip from one format string rather than two catalog
fragments concatenated around a runtime value, so translators can move the
mode name within the sentence, and register IMEXModesCtrl.cpp for string
extraction.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The remote's 3aefae016f is the same tree as local e245f5d069, and every change
in 013ac898ba (continuous smoothing, world-space bake, mirrored/scaled
placements, plate clamp, UV editor framing/HiDPI, layer scrolling) was already
carried forward and reworked by the local commits. Resolved to the local side.
The lift, speed and cached-extruder members now live in the protected section ahead of the private ones; list their initializers first so the list reads in construction order. No behaviour change.
BeltGCode is only created for belt printers, so its hooks no longer re-check belt_printer, and the BBL-machine flag is set once on whichever writer survives init_belt_writer instead of on one about to be discarded.
The bed gravity arrow, volume rendering and the painter/support gizmos each rebuilt
the tilt up-vector from build_plate_tilt_x/y; use one helper that also tolerates
presets without the keys.
Print::has_wipe_tower() is always false for belt printers, but CLI arrange, plate checks and the pre-slice tower clamp still reserved a phantom tower footprint and wrote a clamped wipe_tower_x/y into the config.
Extension layers were all created with id 0, so every one of them could be taken for the first layer by id-only checks (ooze-prevention standby temperature, cached layer ids). Renumber the support layers after inserting them.
gcode_back_transform, first_layer_plane* and belt_printer_infinite_y fell through to invalidate_all_steps(), which re-ran tool ordering, skirt/brim and G-code export on toggles that only affect G-code export.
Apron-only layers printed every band with the first tool, so objects with different brim filaments at the same apron Z shared one filament. Emit each brim filament's bands with its own toolchange.
The ordinary-layer path kept its own copy of the apron band loop. Give emit_belt_brim_bands() an optional brim filament filter and call it from the per-extruder lambda; without a filter it still prints every band, so apron-only layers are unchanged.
toggle_options() now runs on every value change and mode switch; rebuild the
brim_type choices only when the leading-edge entry has to be added or removed.
Every printer's config block lists belt_slice_rotation_angle (default 45), so the processor marked all G-code as belt G-code: imported flat G-code got the belt view on a belt printer, and the belt-only Z handling in the processor ran for non-belt prints whose config block precedes the body. Take the angle only from outside the config block, where only the belt header writes it.
A 90 degree tilt has no finite gravity drift per layer, so the option range
now stops at 89 degrees, matching the cap applied by the support generators.
The three support generators each computed lh * tan(tilt), which overflows
coord_t at 90 degrees and flips sign beyond it (belt sync can write up to
180). One helper now returns the tilt slope with the tilt capped at 89 degrees.
Painted support/seam facets, support volumes, seam occlusion, MMU and fuzzy skin painting (top/bottom
and side facets) and the adaptive infill octree used trafo_centered(), or trafo() with a centre-offset
shift, while the layers were sliced with the belt rotation, remap and Z lift; they now share
PrintObject::trafo_sliced().
The belt writer replaced the plate-offset-carrying writer mid-export, so belt G-code for any plate but the first kept the plate origin and long-travel clipping used the wrong frame. GCode now remembers the offset and hands it to the new writer, and the writer's first-layer probes use the plate-local point it emits.
build_plate_tilt_x/y are printer-preset keys; listing them in the per-object
frequent-settings and object-table bundles stored ignored values in object
configs and crashed the object table on the process config lookup.
Merge origin/main (00429da739) into belt-printer.
Conflicts resolved:
- src/CMakeLists.txt: keep both wxInspector workarounds.
- GCodeProcessor.cpp: keep the belt compare_pos / z_for_height lines.
- PrintObjectSlice.cpp: the belt bbox-Z guard also covers main's
printable_region_ids bookkeeping.
- TreeSupport.cpp: the belt-floor check runs before main's PendingNode
queueing.
- Tab.hpp: keep the belt fields, drop the removed upload description
fields.
- tests/libslic3r/CMakeLists.txt: keep both test files.
Also included:
- eSUN PLA belt presets declare their own filament_id (OFkrxQC4) and
scripts/filament_id_snapshot.json is regenerated, as main's filament_id
check requires.
- Custom.json version bumped to 02.04.00.05 so the belt entries reach
existing installs.
- Fix the ambiguous WithinRel call in the belt apron width test, which
otherwise breaks the fff_print build.
A mixed filament slot is virtual: ToolOrdering::resolve_mixed_filaments()
replaces it with its physical components before any G-code is emitted, so
the toolchanges the prism has to absorb are between those components.
ensure_belt_purge_tower() counted the slot as a filament of its own,
provisioning one island per mixed slot that no swap can ever reach -- the
"extra purge tower" on MCTEST5, where filament 5 is a 50/50 blend of 2
and 4 and the G-code reports 0.00 g of it used.
Expand the assigned set with the same expand_mixed_filaments() the
backend uses, so the GUI sizes the prism against the filament set the
slicer actually produces. No-op when nothing is mixed. Test covers the
MCTEST5 shape, a mixed slot whose components are otherwise unused, and
the no-mixing case.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SsuY8Laiyh7q2zPVVKV3HZ
Two independent leaks of filament on the belt purge prism, plus the
replan safety net the second one needs.
1. The early-truncation scan bounded itself with the prism's own
toolchanges. ToolOrdering covers the whole print and the prism is a
printed object in it, so the "last toolchange" the scan found was on
the prism's own top layers -- it runs past every model object by
design -- and the truncation cancelled nothing. Bound the scan at the
tallest non-prism object (support layers included; on a belt they can
top the object). On MCTEST5 that was 197 toolchanges over 39.4 mm of
tower that no swap ever needed.
2. On a layer with no toolchange, the prism's entire fill printed as
solid infill in its own filament. Drop the fills no toolchange
claimed, right after the purge marking and before
ensure_perimeters_infills_order() force-overrides whatever is left.
Perimeters stay so the bar keeps a continuous wall. An earlier version
of this deleted the entities and had to be reverted: psWipeTower can
rerun without regenerating infill, and a later tool ordering may claim
what this one did not. The entities are now stashed with their layer,
region and index and put back exactly, the same reversibility contract
layer truncation already had.
3. Both stashes go stale if an object step reruns: make_fills() clears
and regenerates fills over m_layers only, so a stale stash would put
old fills back next to new ones, and truncated layers would keep old
perimeters/fills. Undo the plan's edits at the top of Print::process()
whenever psWipeTower is not done. Every object-step invalidation also
invalidates psWipeTower, so that condition is exactly "some object
step may rerun"; when it is done nothing regenerates and the edits
must stay. This also covers a prism left behind after belt mode is
turned off, which previously stayed truncated forever.
WipingExtrusions::is_entity_overridden() becomes public so the prism can
tell claimed fills from unclaimed ones.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SsuY8Laiyh7q2zPVVKV3HZ
Tiles previously rendered a colored monogram (title/source initials plus an ordinal) with a per-id hue. Show the matching native SVG icon instead:
- AppAction/NativeCommand gain an `icon` field; commands get a curated key->icon table and settings inherit their group header's icon.
- Notebook tracks each page's resource icon name and reports it in tab_options(), so the tab picker can show it too.
- Searcher records the group icon so settings keep it through search.
- Web tile rendering swaps monogramFor/hue for an <img>; drop the now-unused hue/text CSS vars. Add a test asserting every non-empty icon resolves to a shipped SVG.
Added shortcut (Ctrl/Cmd+B) to pin/unpin actions.
Clear unresolved pinned actions when switching between modes.
Fixed scroll-back issue when scrolling to pin action
Fixes the report in #12998 (comment 5465250754): first-layer speed and the
slow_down_layers ramp were ignored on a belt printer. The report reads as a
per-object problem, but neither applied to *any* object -- the reporter's first
part slowed down because slow_down_for_layer_cooling was on, which is
CoolingBuffer's time-per-layer mechanism, not initial_layer_speed.
FirstLayerPlane decides first-layer-ness by perpendicular distance to a plane it
derives by composing gcode_remap_* with compute_machine_z_affine(). The plane is
therefore a function of how G-code is *addressed*, not of where the belt is:
change the output axis convention and the plane moves. On MCBELT-TYPE2 the
first layer measured 86.2 mm from the plane and got effective index 431, far
past any slow_down_layers ramp.
on_first_layer(point) and effective_layer_index_for_point() now measure height
above the belt surface, using the belt description already carried in
SlicingParameters -- belt_floor_shear_factor / belt_floor_from_axis /
belt_floor_z_shift -- the same description the support generator uses. That is a
property of how the object was sliced, so no remap or back-transform can perturb
it.
Deliberately not via BeltFloorContext: its init() folds in
belt_support_floor_offset, a support-generator diagnostic, and letting that
option steer the model's first-layer speed band would be a surprising coupling
(a negative value would switch the slowdown off outright).
Preserving the existing first-layer-plane settings:
* first_layer_plane XY/YZ/XZ keeps the FirstLayerPlane evaluator, as those are
explicit opt-outs.
* A non-zero first_layer_plane_offset also keeps it. The offset is a machine-Z
shift that FirstLayerPlane converts into a perpendicular distance in the
slicing frame; this evaluator measures along slicing Z, so there is no
faithful translation. Deferring to the evaluator that implements the setting
beats silently ignoring it.
* The two thresholds stay separate, exactly as FirstLayerPlane keeps them:
the first-layer boolean tests initial_layer_print_height, while the
effective layer index counts bands of first_layer_plane_thickness.
Brim and coincident apron bands are emitted before m_layer is switched to their
object -- for an apron band there is no Layer at all -- so both paths publish the
belt-floor owner explicitly. Without that a brim's classification would borrow
whichever object was visited previously, making it depend on plate order.
Note that first-layer-ness drives more than speed: extrusion acceleration, jerk,
the first-layer flow ratio and eligibility for overhang speed/fan analysis all
read it, so all of them are corrected on belt printers by this change.
Classification still samples only each path's first point, as it did before.
Non-belt is unaffected by construction: belt_height_above_floor() returns false
when the belt floor is inactive and both call sites fall back to the previous
path. FirstLayerPlane stays in place for its other modes and for CoolingBuffer,
whose machine-coordinate probe is a separate outstanding bug.
Measured, MCTEST4 on MCBELT-TYPE2 (initial_layer_speed=5, slow_down_layers=40):
15 distinct feedrates with no gradient and F300 absent, becomes 70 including the
full ramp 300(5) 382(6) 465(8) 630(10) 795(13) ... Two bare cubes on a belt:
0 slow extrusions becomes 2378 across Z 32.36..95.18. The same two cubes on a
Cartesian printer keep their slow extrusions confined to Z 0.20..2.00.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011jgzj1sf53KMLPweZ8yeUQ
1. Belt tree support could not slice at all.
layer_initialize() hardcodes layer 0's bottom_z to 0, encoding "below layer 0 is
the build plate at z = 0". True for a flat bed; false for a belt, whose virtual
support layers legitimately extend below zero. The bottom-most belt layer
therefore got height = print_z - 0 = -9.8, which reached Flow::with_height() and
threw FlowErrorNegativeFlow.
A 3DBenchy, a mushroom, an L-bracket and an extruded L all failed identically
with negative flow / return -100. Only a bare cube sliced, because its support
never reached that far down.
The bottom is now taken from the previous layer's z, and only a layer 0 whose
print_z is itself negative gets a synthesised bottom below it. Every
non-negative print_z -- every non-belt configuration -- keeps exactly the
previous 0, so this is behaviour-preserving off a belt by construction. An
earlier form used min(0., layer_z(0) - layer_height), which regressed flat beds
whenever the initial layer was thinner than the layer height.
3DBenchy on a 45-degree belt with organic tree support: fails to slice ->
247 support blocks / 168,596 extrusions.
2. Support generated against the belt, and against belt-tilted walls.
A plain 20mm cube on a 45-degree belt generated 86 support blocks and 46,307
support extrusions. Three causes, all gated on the belt floor being active:
a. The build-plate tilt compensation shifted the lower layer the wrong way.
tan(build_plate_tilt_*) carries a magnitude but no direction, and the sign
chosen moved the lower layer away from the newly appearing material rather
than under it, doubling the mismatch. The shift now comes from
belt_floor_shear_factor / belt_floor_from_axis, which carry sign and axis
exactly. Non-belt tilted beds keep the previous behaviour.
b. Material resting on the belt was treated as unsupported. The belt surface
is now unioned into the effective lower layer, sampled at the bottom of the
layer -- a layer meets the belt across its thickness and print_z is the
top. The half-plane is clipped to the layer's bounding box first: unioning
a +/-1000mm half-plane raw with 20mm-scale geometry put a huge dynamic
range through Clipper and left intermittent artefacts every few layers.
c. The object's first slice can be empty on a belt (the bottom vertex is a
sub-extrudable sliver), leaving the layer above with an empty predecessor
even though it rests on the belt. (b) already covers that per island. What
did need fixing is sharp-tail detection, which tests each island against
the raw lower slices; with an empty predecessor that test is trivially true
and every belt-contact island read as a sharp tail. It now tests against
the same effective lower layer.
An earlier form instead skipped the whole layer when the point of
get_extents(curr_polys) -- the bounding box of the union of every island --
nearest the belt was in contact. That was wrong in a way worth recording:
one island resting on the belt could suppress overhang and sharp-tail
detection for a separate island floating well above it. Every decision here
is per-island.
Cube on belt: 46,307 -> 0 support extrusions. Same cube non-belt: 0 before and
after. Benchy on belt still 247 blocks / 168,596 extrusions and a mushroom
111 / 82,157, so false positives are removed without suppressing true ones.
Non-belt is unchanged by measurement, not only by the belt_ovh_active gate:
the same mushroom sliced on a Cartesian printer before and after gives 65,866
support extrusions and 68,717 total extrusions both times, the two G-code
files differing in exactly one line -- the object's plate position.
3. m_anti_overhang was filled and read in different index spaces.
It is consumed in the same index space as m_layer_outlines, where object layer i
lives at num_raft_layers + i, but was filled in object-layer space. Every entry
landed num_raft_layers too low (50 for a 20mm cube at bed Y=50) and the topmost
object layers got none. The belt injection also ran before m_raft_layers was
extended, so it could not have known the offset.
The array is now shifted as a whole and the injection moved after the raft
extension. This also repairs user support blockers under a raft, which is not
belt-specific: it changes behaviour for any ordinary raft, not just the belt's
virtual one, and should be reviewed as a general fix. Measured effect on the
cube was small on its own (46,307 -> 46,334 before the other fixes) because
m_anti_overhang only feeds calculate_placable; kept as a correctness fix on its
own merits.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011jgzj1sf53KMLPweZ8yeUQ
Both change emitted G-code, which is why they were kept out of the extraction
commit. Both are wrong only where the machine mapping is non-identity, which is
the definition of each bug.
1. Suppress lifts commanded through an unknown position.
_travel_to_z() emits full XYZ whenever the mapping must emit every axis, because
the mapping can make machine Z depend on logical X/Y, and it builds that point
from m_pos. At print start, and after any custom G-code that invalidates
position, m_pos.xy is the uninitialised origin; mapping (0, 0, z) through a
non-identity remap produces a real but wrong machine point -- for a reverse
mapping, build_vol_max, i.e. the far corner of the bed. The subsequent full-XYZ
move corrects the position, but the lift has already commanded a rapid across
the whole bed at travel speed.
Belt kinematics already guarded this; the Cartesian path did not. The guard is
now applied at all three lift sites through must_skip_lift_now(), not just the
one the extraction covered: travel_to_xyz()'s pending-lift branch,
lazy_lift(spiral_vase=true), and eager_lift(). The latter two also needed the
state fix -- both recorded m_lifted = target_lift regardless, so suppressing
only the emission would leave a later unlift() descending from a height that was
never commanded.
2. Never emit a G2/G3 arc a mapping cannot represent.
extrude_arc_to_xy() emitted G2/G3 with logical X/Y and I/J and never consulted
the mapping. There is no general fix by transforming the arc: a permutation
moves it out of the XY plane that I/J describes, a negation reverses handedness,
and the belt shear maps a circle to an ellipse that G2/G3 cannot express at all.
So supports_arc_moves() gates generation through the existing
GCode::should_disable_arc_fitting() hook, and BeltGCode's special-case override
is deleted -- belt now gets the same behaviour from the general rule instead of
its own exception.
supports_arc_moves() is m_remap_x == 0 && m_remap_y == 1, not !has_axis_remap():
an arc emits only X/Y/I/J, so a mapping that merely negates or reverses Z leaves
every emitted word untouched and keeps its arcs.
The fallback for an unrepresentable arc tessellates it into linear segments at a
0.005mm chord tolerance rather than substituting a single chord, and splits dE
proportionally across the segments. The capability check is hoisted above every
extrusion mutation: an earlier form ran it after filament()->extrude(dE) and so
extruded 2*dE on the fallback path.
Known limits of that fallback, since it is worth stating rather than discovering:
emitted relative E is conserved only to per-segment rounding (a radius-5
semicircle with dE=1.5 emits 1.50012 across 36 segments); the 0.005mm bound is a
logical-frame bound, about 0.00855mm in machine space under a 45-degree belt
shear; unequal endpoint radii and non-finite inputs are unchecked. Ordinary
export takes the original polyline when the mapping rejects arcs, so this path
is a fallback rather than the normal route.
Known gap, not claimed fixed: classic wipe towers have their own
enable_arc_fitting and their own G2/G3 emitter in GCode/WipeTower.cpp, which
should_disable_arc_fitting() does not govern. Belt printers are barred from
classic wipe towers; a remapped Cartesian printer is not.
Tests in tests/fff_print/test_gcodewriter.cpp: reverse-X remap with unknown and
with known position plus an identity control; eager_lift emitting nothing and
recording nothing; the arc-capability matrix including the Z-only cases; and the
tessellated fallback. E accounting is asserted through used_filament() rather
than E(), which resets per line in relative-E mode.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011jgzj1sf53KMLPweZ8yeUQ
BeltGCodeWriter subclassed GCodeWriter and overrode seven methods, five of them
by copying the base body and changing the transform. The base writer already
carried an axis remap and already branched at each of its seven
coordinate-emission decisions; the subclass did the same branching with a
different transform, and the two copies had begun to drift.
Replace the inheritance with a strategy object owned by GCodeWriter:
CartesianKinematics to_machine = the existing apply_axis_remap; today's base
behaviour, moved rather than changed.
BeltKinematics to_machine = MachineFrameTransform o axis_remap o
BeltBackTransform, plus a world_coordinates variant for
the PA calibration generators.
New: src/libslic3r/GCode/MachineKinematics.{hpp,cpp}, GCode/BeltKinematics.{hpp,cpp}
Deleted: src/libslic3r/BeltGCodeWriter.{hpp,cpp} (341 lines)
Points worth a reviewer's attention:
* The predicate is must_emit_all_axes(), not couples_axes(). The base returns
true for any non-identity remap, including pure permutations that do not
physically couple axes, so the question is "must every axis word be
emitted", not a statement about kinematics.
* Every per-site word-omission branch is preserved. The base deliberately
emits X/Y only, or Z only, or drops Z when its quantised value is unchanged.
The strategy changes which transform applies, never whether words are
omitted.
* set_kinematics() replays the configured remap and build volume onto a newly
installed strategy, because BeltGCode::init_belt_writer runs before
GCode.cpp calls set_axis_remap/set_build_volume_max.
* uses_pointwise_travel_speed() preserves a pre-existing divergence rather
than introducing one: the base travel_to_xyz emits the raw configured travel
speed in its final branch, ignoring the first-layer value computed at the
top, whereas the belt path used the first-layer-aware value throughout. Both
are kept. Unifying them changes feedrates and belongs in its own change.
* The [BELT-DEBUG] block is deleted; it rate-limited itself with a
function-local static thread_local in the hot emission path, and this is the
commit that would otherwise have moved it into shared code.
This commit is intended to preserve existing export output. That is reviewed by
construction -- each emission site keeps its own omission branch and each policy
divergence is preserved -- and is NOT verified against a G-code diff corpus.
Building that corpus is the outstanding work here.
Two API-equivalence exceptions, neither reachable by any caller today:
* Belt kinematics with no plane pointer installed, m_is_first_layer true,
initial and normal travel speeds differing, travel_to_xyz() reaching its
final branch: the old belt writer selected the initial-layer speed, the new
writer selects the normal travel speed. The pending-lift and XY-only
branches keep their previous selection.
* Belt kinematics installed without set_force_normal_lift(true) and a
non-normal lift requested: the old belt writer forced a normal lift, the new
writer can take the slope branch.
The PA-pattern generator reaches the writer through explicit travel_to_z() /
travel_to_xy(), not travel_to_xyz() or the lazy/eager lift paths, and normal
belt export installs both the plane and the forced-normal-lift policy, so
neither exception changes output produced today. They are recorded because a
future caller could reach them.
tests/fff_print/test_gcodewriter.cpp was also not compiling before this branch:
it called writer.to_machine_coords(), a method that existed only on
BeltGCodeWriter. It never surfaced because the build targets OrcaSlicer, not
all, and BUILD_TESTS defaults to OFF, so that translation unit was outside every
compile path. Fixed here; the existing 30-degree coordinate assertions are kept
verbatim as the best available regression net.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011jgzj1sf53KMLPweZ8yeUQ
Two harness fixes, both paid for by hours of chasing product bugs that were not there.
ROUNDED RECTANGLE. It is the shape the user reported and the gate could not reach it:
the rectangle family binds R to CornerRect and leaves the other modes in the toolbar
flyout, so the three-step Width -> Height -> Radius chain was never exercised.
rung_rounded_rect() arms it over MCP the way the offer menu does. The verb id is the
OFFER id `sk_rect_rounded` — `design_rect_rounded` is the ACTION name, run_verb
throws on it, and the tool silently stays Select; a run that misses that draws
nothing and still reaches its assertions, so the rung arms AND verifies.
THE NETWORK PLUGIN MODAL. GUI_App::post_init() re-raises "Bambu Network Plug-in
Required" from an IDLE event, after any startup sweep has closed it, and
ShowModal() runs a nested event loop: the app is alive, its window is on screen,
and the MCP socket answers nothing. That is indistinguishable from a hang and was
investigated as one, with gdb, twice — the attached stack finally read
ShowModal <- show_network_plugin_download_dialog <- post_init. Seeding
`installed_networking` false stops the whole networking path, so the dialog never
exists to be swept.
Also: check() returns its verdict, so a rung can abandon itself when a precondition
fails instead of asserting into a dead end.
38 checks hold on behemoth against e5659e0f0f.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
check-gui-sketching.py located the value field by hunting for a small top-level
window and clicked into it before typing. Both halves are now wrong, and the second
was always a problem:
- field_win() cannot find a field that is not a window any more, so every check
built on it silently became one that cannot fail. Replaced by field_open(),
which asks the app: sketch_describe's `editing` is value_field_open().
- focus_field() clicked into the field first, and its own docstring said why —
"WITHOUT THIS THE TYPED VALUE IS SILENTLY DISCARDED". That workaround is exactly
what made this suite blind to the defect the user reported: a ladder that clicks
the field first can never notice that typing WITHOUT clicking is broken. Now a
documented no-op; the field is in the canvas and the canvas has the keyboard.
Measured on behemoth against e5659e0f0f: 64 checks pass, 2 fail. The two are
polygon regularity and area (sides {29.999986, 29.975164, 30.0}, area 2336.98 vs
2338.27) — geometry tolerances, nothing to do with typed values; before this change
the same run failed 5, all of them dimensions and constraints that never received
their value because focus_field() was clicking at a window that no longer exists.
No baseline exists for the remaining two, so they are reported, not claimed as
pre-existing. The suite also still stops at D3 with the app gone; tracked separately.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Rebases the in-canvas work onto cad-mainline and finishes it. The field is drawn
by ImGui inside the GL canvas instead of being a borderless top-level wxFrame.
WHY THE FLOATING FRAME COULD NOT BE FIXED. Whether a borderless top-level may hold
the keyboard is the window manager's decision, and it differs per desktop: openbox
grants it, mutter refuses it, macOS denies key status outright. Seven workarounds
fought that and one cost a macOS regression. Drawn inside the canvas there is no
second top-level for anyone to refuse, so the question is never asked. The field is
fed exactly like every other ImGui widget in the app — GLCanvas3D::on_char ->
ImGuiWrapper::update_key_data -> io.AddInputCharacter.
MEASURED, on behemoth: the click-edit ladder holds 28 checks — Line, Rectangle,
Circle, Slot, Polygon, Ellipse, Arc, and click-to-edit on a placed dimension label
— typing with NO click into the field first, committed == typed != prefill every
time, and 27 [UX] imgui_char lines showing the characters arriving.
WHAT WAS ACTUALLY WRONG. Not the field. The belief that "characters never reach the
ImGui InputText" came from the harness: the ladder was delivering keys with
`xdotool type --window` (XSendEvent), which GTK discards, so no build of any kind
could have received them. The new probe in ImGuiWrapper::update_key_data — the one
place ImGui is ever handed a character — is what separated that from a real defect,
and it stays, because a canvas-side probe provably cannot answer the question:
GLCanvas3D::on_char is bound later than any constructor-time probe, wx runs handlers
in reverse bind order, and on_char returns without Skip(), so such a probe is silent
whether or not the key arrived. A day was lost reading that silence as evidence.
Also drops DesignPanel's content-based forwarder and DesignCanvas::inline_type_char.
They were the right rule for a field that could not be focused; with the field
inside the canvas there is nothing to forward, and keeping them would have masked
whether the normal path works.
STILL UNVERIFIED: behaviour under mutter itself. Neither focus-stealing-prevention
WM available here survives long enough to judge — metacity SEGVs ~20s in and xfwm4
dies with BadWindow on SetInputFocus, both before the sketch opens and both
unrelated to this field. The design's claim is structural rather than measured: no
second top-level means no focus to refuse.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
focus= in [KEYTRACE] was never a class name. GetClassName() returns const wxChar* — wchar_t* in
this build — and the line cast that to const char* and printed it with %s, so it emitted the first
byte and stopped at the padding NUL. "wxGLCanvas" came out as "w". So did "wxWindow". Every focus
reading taken from this instrument for a whole day of diagnosis was a single character, and the
one question it existed to answer — WHICH widget has the keyboard — was the one it could not
answer. Printed through wxString now, and it says focus=wxGLCanvas: the canvas does hold wx focus
while the field is open, which removes the focus hypothesis for good.
Also here, and HONESTLY LABELLED AS INCONCLUSIVE: a wxEVT_CHAR probe on the canvas. It logged
nothing (cc=0), and the tempting reading is "the characters never reach the canvas". That reading
is not available, because the probe is bound in the DesignCanvas constructor BEFORE
GLCanvas3D::bind_event_handlers(), and wx runs the most recently bound handler first —
GLCanvas3D::on_char returns without Skip() exactly when ImGui consumes a character, which is
precisely the case under test. A silent probe is therefore consistent with ImGui consuming the
keys correctly AND with them never arriving. It measures nothing. Rebind it after
bind_event_handlers(), or instrument update_key_data itself, before believing anything about it.
Writing this down rather than acting on it: I came within one commit of "fixing" a mechanism I had
inferred from an instrument that could not see it, which is the same mistake as the focus= field
above and the same mistake that cost a whole session in September.
Deployed binary restored to 00d6c191dc (md5 0eeb9a58cef5).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Third measured step, and the last one I will take without a second pair of eyes.
set_as_dirty() BEFORE Refresh(): GLCanvas3D's paint handler returns without rendering when the
canvas is not marked dirty, so the previous commit's bare Refresh() posted paint events that drew
nothing and the frames stopped anyway. With both halves the pump sustains, and the trace shows the
field holding the keyboard frame after frame:
[UX] frame want_text=1 want_kb=1 active=1 buf=158.74
[UX] frame want_text=1 want_kb=1 active=1 buf=158.74 (repeating)
STILL OPEN, and now narrowed to one question: buf never changes. ImGui owns the keyboard and our
InputText is the active item, so what is missing is upstream of ImGui — the characters are not
reaching io.AddInputCharacter at all. The next thing to MEASURE (not to change) is whether
wxEVT_CHAR arrives at the GL canvas in the Design tab: DesignPanel's wxEVT_CHAR_HOOK Skips digits
while inline_busy(), but Skip only helps if the focused widget is the canvas, and nothing has yet
proved that it is at the moment the keys are sent.
Three attempts have now gone into this one point. Per the standing rule that is where solo
iteration stops.
Deployed binary restored to 00d6c191dc (md5 0eeb9a58cef5) — nothing from this branch is installed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Measured, not reasoned. A per-frame trace of the ImGui state is what finally named the mechanism,
and it is a deadlock, not a focus problem:
[UX] frame want_text=0 want_kb=0 active=0 buf=158.74 <- frame 1: the widget is not active yet
[UX] frame want_text=0 want_kb=0 active=1 buf=158.74 <- frame 2: now it is
(nothing further) <- and the canvas stops
This canvas repaints ON DEMAND. ImGui decides whether it wants the keyboard at the END of a frame,
from the active item; GLCanvas3D::on_char only calls render() when update_key_data() says ImGui
wants it. So: no frames -> WantTextInput never turns on -> no render on a keystroke -> still no
frames. The characters sit in ImGui's input queue and the field is exactly as deaf as the window
it replaced, for a completely different reason.
request_frame breaks the circle, and the same trace says so:
[UX] frame want_text=1 want_kb=1 active=1
That is the first time in this file's history that the value field has owned the keyboard without
asking a window manager for it.
STILL OPEN: the typed characters do not reach the buffer (buf stays at the prefill) and the frames
stop after nine. The pump is the suspect — on software GL request_repaint() calls m_canvas->render()
SYNCHRONOUSLY, so this asks for a render from inside a render; it needs to schedule one instead.
That is the next thing to measure, not to guess.
The deployed binary on behemoth is restored to 00d6c191dc (md5 0eeb9a58cef5), byte-identical to
the last good build. Nothing from this branch is installed.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
The decision (Tommaso's, put to him with the trade-offs): the field stops being a separate
top-level window, because whether such a window may receive typing is the window manager's
call and not ours. openbox grants it, mutter on his desktop refuses, and seven previous
workarounds fought that — one of them causing a macOS regression, and the test harness ending
up clicking the field before typing, which is a workaround no user can be asked to perform and
is exactly the "label value not editable" report.
DONE and proved on the rig: SketchInlineEditor is no longer a wxFrame + wxTextCtrl. It is state
plus an ImGui overlay drawn by DesignSketchTool::render(), at the same screen anchor, in the same
vocabulary as the dimension labels next to it (draw_dim_label is already an ImGui window). The
field opens where it should — [UX] open title=Length prefill=158.74 from the running app.
NOT DONE: typing does not reach it. ImGui is fed from GLCanvas3D's own key handler, so the keys
have to arrive at the canvas; giving the canvas wx focus when the field opens was not enough.
The remaining question is where a keystroke goes between DesignPanel's wxEVT_CHAR_HOOK and
GLCanvas3D::on_char in the Design tab, and whether the canvas repaints often enough for ImGui to
advance its input state. That is attempt three on this specific point, so it goes to a second
opinion rather than a third guess.
Also here: scripts/CAD/check-gui-click-edit.py, the ladder Tommaso asked for. It types into the
field WITHOUT clicking it first — the click is what check-gui-sketching.py's focus_field() does
and why that suite can never see this defect — and fails when the prefill is what gets committed.
It currently fails, correctly, on the above.
Not on cad-mainline: the deployed binary must stay the last good build until typing works.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
present_toplevel() already asked for focus with a server timestamp, but only when
the widget happened to be realized. An unrealized widget has no GdkWindow, so
there was nothing to read a timestamp from and control fell through to
wxFrame::Raise() — which asks for activation with GDK_CURRENT_TIME, i.e. 0.
Zero is exactly what focus-stealing prevention discards. metacity says it out loud
when a sketch value field opens:
Buggy client sent a _NET_ACTIVE_WINDOW message with a timestamp of 0
and mutter, same lineage, refuses it silently on the user's desktop. That refusal
is the reported defect: the field is visible, never receives the keyboard, and
Enter commits the as-drawn prefill.
So realize the widget and retry, and do NOT fall back to Raise() on X11 — a
timestamp-0 activation is refused anyway, and on some window managers it only
marks the window as demanding attention.
Not yet confirmed end to end: both window managers with focus-stealing prevention
available here abort on this frame — metacity at frames.c:1239, xfwm4 with
BadWindow on SetInputFocus as the frame is destroyed under it — so the ladder
cannot yet return a trustworthy verdict under one. Two WMs crashing on the same
borderless, repeatedly re-mapped STAY_ON_TOP frame is its own signal about this
design. Tracked in projects-1p5 and projects-40m.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
The reported defect: sketch dimension labels are "not editable" — you draw a
rectangle, its Width field opens, you type, and the as-drawn number is committed
instead. It affects every sketch tool, not just the rounded rectangle.
WHAT THIS ADDS
1. The arbiter (DesignPanel CHAR_HOOK -> DesignCanvas::inline_type_char ->
SketchInlineEditor::type_char). Routes a key by what it IS, not by who the
window manager focused: digits, sign, decimal separator and Backspace/Delete
go to the open value field, Enter/Tab commit, letters stay tool shortcuts.
This is FreeCAD Sketcher's rule (DrawSketchKeyboardManager::
detectKeyboardEventHandlingMode), and the reason its sketcher behaves the same
on every desktop: it never asks who has focus.
2. The [UX] trace (SNAPORCA_UXTRACE) in SketchInlineEditor: open/commit/refused/
cancel, with the prefill and what the control actually held at Enter. It did
not exist — the ladder below was written against a surface no build emitted,
so it could only ever report "nothing opened". typed == prefill on a commit is
the defect's signature and nothing else makes it visible.
3. A draw-then-edit trace in DesignSketchTool: four early returns can swallow the
value-field chain and from outside they are indistinguishable.
4. scripts/CAD/check-gui-click-edit.py — types WITHOUT clicking the field, as a
person does, across Line/Rectangle/Circle/Slot/Polygon/Ellipse/Arc plus label
click-to-edit, and asserts committed == typed != prefill.
5. scripts/CAD/focus-loop.sh — sync/build/assert on behemoth. NOT the orcacad-gui
rig: its image pins deps 216 non-CAD files behind cad-mainline, so today's CAD
sources cannot build there without a deps rebuild.
WHAT IS PROVEN, AND WHAT IS NOT
Green under openbox: 28 checks, every tool, committed == typed != prefill.
But openbox CANNOT adjudicate this bug and the ladder says so in place. There the
field always wins the keyboard, so the same ladder also passes against a binary
with the arbiter compiled out — measured twice. Two ways of removing the keyboard
were tried and both are recorded as dead ends: XSetInputFocus loses to the field's
own re-focus CallAfter, and XSendEvent (xdotool --window) is dropped by GTK, which
made every run red regardless of the code.
Under metacity — same focus-stealing-prevention lineage as the user's mutter — the
mechanism appears in the WM's own log:
Buggy client sent a _NET_ACTIVE_WINDOW message with a timestamp of 0
That is the activation being refused, which is exactly the reported symptom.
present_toplevel() already asks for a server timestamp, so a path is still falling
through to frame->Raise(), which sends time 0. That is the next thing to fix, and
it is tracked; the arbiter alone does not close it. metacity also aborts on this
window (frames.c:1239), so the gate needs a WM that survives before it can return
a verdict.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
"Still cannot edit labels in rounded rectangles." Reproduced on the rig in a few minutes, and it
is NOT the window-manager defect the rest of this week has been about — it happens on openbox,
where typing into a value field works perfectly. The value was never the problem. The LABEL was
not there.
render_live_quotes picks the entity to speak for like this:
else if (m_selection.size() == 1) ei = m_selection[0];
if (ei < 0 || ...) return;
A rounded rectangle is EIGHT entities — four lines and four arcs — so selecting the shape makes
m_selection.size() == 8 and the pass returns before drawing anything. Its Width, Height and fillet
Radius are live labels and nothing else, so with them gone there is no affordance at all: no
number to click, no field to open, no value to refuse. The rule hid the characteristic quotes for
precisely the shapes that have nothing but characteristic quotes.
A plain rectangle looked fine only by accident. Typing into its auto-edit chain creates a DRIVEN
dimension, which render_dimensions draws from the annotation list, so its labels survive. The
rounded rect's W/H/R go through set_rounded_rect, which rebuilds the geometry and leaves no
annotation behind. Same for slot, arc-slot and polygon: every grouped feature was in this hole.
A selection that is entirely ONE feature now speaks through any member. The switch below already
keys off feature_of(ei) rather than the entity, so nothing else had to change.
Measured on behemoth :10, before and after, same binary path:
before 8 selected -> no labels at all
after 8 selected -> R26.6 / 117.4 / 150.7 drawn; clicking R26.6 opens Radius prefilled 26.60;
typing 8 gives R8.0 mm and visibly sharper corners.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Two reports, three defects, all three measured on the running app rather than reasoned about.
"Keyboard focus in drawing tool is broken so that now they are slow and cumbersome." After a
dimensioned entity the bottom-right readout chip — 119x31, borderless, a wxFrame — held the X
input focus. Pressing r produced NO [KEYTRACE] line at all: the key never reached the panel's
CHAR_HOOK. One bare canvas click moved focus back to the main window and the identical key armed
the tool. So every shortcut was dead after every dimension, and the way to get the keyboard back
was to click somewhere harmless. That is the whole of "slow and cumbersome".
Its sibling, the status chip, is a wxPopupWindow for exactly this reason and carries a comment
warning against turning it back into a frame. The readout was left a frame on the premise that
"it appears mid-gesture and the next input is the mouse" — which the measurement falsifies: the
chip keeps the last value on screen after the gesture ends, and a frame that has the focus does
not give it back. It is now a popup too, with the placement and the iconise/deactivate lifecycle
its sibling already needed, because an override-redirect window would otherwise sit on the bare
desktop when the app is minimised.
"Planes hide the bed." Literally true, twice over. The reference planes were half-extent 0.6 *
the bed's larger side — a square 1.2x the plate — and all three are drawn with depth testing
off, so they painted over the plate grid from edge to edge. 0.3 puts them inside the bed, which
is also the Onshape look the size was reaching for: a modest square at the origin, not a
tablecloth.
And the other half was mine. 3f52166e32 muted the bed for the duration of a sketch, on the
argument that a plate grid and a sketch grid are the same visual language. The argument is right
and the call was wrong, because there IS no sketch grid to take over. Pick XY, arm Line, and the
viewport was an empty grey field: no bed, no grid, no origin, nothing to judge a length or a
direction against. The plate grid was carrying the ground reference for the whole tab. The banner
already says where you are; taking the floor away as well only made the sketch harder to draw.
The Bed checkbox is the one thing that governs the bed, in every mode.
Verified on behemoth :10 with the rebuilt binary: focus after a dimension chain is the main
window, r arms Rectangle with no click in between, and the plate grid is under the sketch.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Measured on the running app, not deduced. After a queued dimension chain the value field's frame
is left MAPPED on purpose (mutter refuses keyboard focus to a re-mapped window), so there is a
window in which m_open is already false and the frame is still on screen holding the X input
focus. GTK meanwhile reports that window inactive and routes nothing into the text control. Every
key then lands somewhere that cannot use it and will not give it back:
[KEYTRACE] key=27 ui_mode=1 inline_busy=0 <- the last key the panel ever sees
=== MARK press Delete === <- no trace line at all
xdotool getwindowfocus -> 0xe00404 86x60 <- the value field, still mapped
Delete, Esc and typing all read as dead, which is exactly the report. And nothing could recover
it: close(), cancel() and do_cancel() all return early on !m_open, so the one window still
receiving keystrokes was also the one window no code could dismiss.
is_mapped() asks the question the flag cannot answer, and dismiss() tears the frame down with no
m_open guard, since m_open is precisely what lies in this state. Esc inside the field falls back
to it — while the frame holds focus that handler is the only code the keyboard can still reach,
so if it refuses, nothing else gets a turn. Every close now hands focus back to the canvas
explicitly, because hiding a window does not move the X input focus off it. And inline_busy()
reports the union of "a value is pending" and "a frame is mapped", so Esc routes to the field
whenever one is on screen at all.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Three cues, because one is missed. A teal banner across the top of the viewport names the session
("Editing: Sketch N") and where its exits are; the printer bed is muted for the duration, since a
plate grid and a sketch grid are the same visual language and reading one as the other is how a
sketch gets drawn against the wrong reference; and N looks straight down the plane normal at the
current zoom, with the plane's own y axis as up, because no hand-orbit lands exactly square and a
sketch read at an angle is one whose right angles do not look like right angles.
The banner is an INDICATOR. Finish and Cancel stay on the single ribbon action bar — the tab had
three competing confirm surfaces once and that is not being reopened for a strip of colour. It
sits above the canvas rather than floating inside it: a child window over a wxGLCanvas is a native
window on GTK with no reliable stacking over GL, and being unmissable beats being clever.
The bed checkbox stays the stored preference and is restored on leaving the sketch; ticking it
mid-sketch still shows the bed, because that is a deliberate act and this is only a default.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Two presses used to discard a live sketch. The key was answered in four places that could not
see each other — the inline value field, a sketch branch, a feature-card branch, and the canvas
— so a press aimed at one fell through to the next, and request_exit() carried a fourth layer
that deliberately let the SECOND consecutive press through to cancel_sketch(). The warning it
showed first did not help: the two presses are never one decision, the first is aimed at a field
or a tool and the second at whatever was underneath it.
The stack is now explicit. CadLevel (DesignInteraction.hpp) is four levels deep, the enum value
IS the LIFO depth, and cad_escape_level() is a constexpr function over a POD of four booleans —
so the ordering that is the entire contract is checked by static_assert at compile time, with no
window, GL context or event loop. DesignPanel::escape() acts on the one level escape_level()
names and on no other, and every Esc in the tab routes through it.
The destructive layer is gone from request_exit() itself rather than guarded at its callers, so
the guarantee cannot be re-opened by adding a route: a session holding geometry is left only
through Finish (keep) or Cancel (discard). Cancel now asks before discarding — it used to refuse
and tell the user to press the button they had just pressed, which meant a drawn sketch could be
kept but never thrown away.
Right-click also stops rewarding navigation with a menu: the offer needs BOTH budgets, released
within 200 ms and moved no more than 3 px, and the raycast uses the press position, so the menu
describes what was pointed at rather than where the camera stopped. Two budgets because drift
alone still popped a menu at the end of a slow, careful orbit.
docs/ux/interaction-model.md carries the state machine, the routing and the transition table.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Five sites carried a hardcoded fallback for imex_tools_per_gantry that had
to match the value registered in print_config_def, with nothing enforcing
the agreement, and several explanatory comments miscounted the sites they
described or cited stale line numbers.
Add imex_cfg_int/_float/_bool/_enum<T> to IMEXHelpers, which return the
value registered for the key when it is absent from the config, so the
registration is the single source and there is nothing left to keep in
sync. Route every read of the IMEX geometry keys through them: 32 call
sites across IMEXZones, PartPlate, GCodeViewer and Tab. The only direct
lookup left is the bail in PartPlate::imex_multicolor_block_reason, which
must not default because it reports a routing conflict and a defaulted
grid would produce a false warning.
imex_cfg_enum uses dynamic_cast on both halves rather than the type()
comparison the others use: every ConfigOptionEnum<T> reports coEnum, so a
type() check cannot tell one enum type from another and would cast a
ConfigOptionEnum<OtherEnum> to the requested T. The ConfigOptionPercent :
ConfigOptionFloat inheritance that rules dynamic_cast out for the float
accessor has no analogue for enums.
Correct the comments that prompted this: the cache-key input list in
PartPlate named five inputs for a nine-part key, the ImexMarkerKey note
in GCodeViewer called imex_tool_layout an input only the preview reads
when the plate keys it too, and four file:line citations pointed at the
wrong lines. Values are unchanged at every converted site; cache key
strings keep their existing representation.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Resolves all 21 inline comments, plus six changes that altered behaviour for
users not using the feature and seven defects found alongside them.
The reviewer's central point generalised: a pressure-advance change had moved
every RepRapFirmware user onto an unverified command form. Auditing for that
class found five more — 14 config keys leaking into every exported g-code, an
ungated Moonraker sync writing to non-IMEX printers' presets, every slice
eagerly re-rendering all plate thumbnails, a preset delta-encoding regression,
and physical_extruder_map being normalised for printers that read it the other
way.
The structural asks landed as asked: the 392-line bed-zone geometry moved to
libslic3r and is now unit-tested, the preview consumes that same layout instead
of a second copy, nine open-coded mode lookups became one, and the modes editor
moved out of Tab.cpp. Making the geometry testable exposed three further
defects in it, including an aggregated gantry that raised no collision strip.
The worst bug was not in the review: the GUI computed the firmware-managed
slice offset in the plate-list world frame while both consumers subtracted the
plate origin again, so every plate after the first failed to slice with 'part
is off the plate'.
User-facing strings now read IDEX/IQEX, honouring 461c69c83e. Config keys, C++
identifiers and 3MF metadata keys keep the imex_ spelling as on-disk format.
815/815 tests pass in Release.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
No IMEX code upstream, so nothing in this merge touches the feature. All 21 overlapping
files auto-resolved; verified every upstream addition is present in the merged tree.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 10.
`461c69c83e` settled this in April — IMEX internally, IDEX/IQEX as the user-facing label — but the
UI strings were never converted. Every translated string naming the feature now reads IDEX/IQEX:
41 occurrences across the printer and process option labels and tooltips, the modes editor, the
plate mode indicator, the pre-slice warnings, the placement refusals and the slicing errors. The
reviewer listed eight; the rest were in the same class.
Nothing else moves. The config keys keep the `imex_` spelling — `is_imex`, `imex_mode_names`,
`imex_parallel_mode` and the rest are on-disk format in existing printer presets and 3MF projects,
so renaming them would break every profile and project already saved. C++ identifiers, filenames,
comments and test names keep IMEX as well: it stays the internal name of the subsystem, which is
what covers the topology space (one gantry with 2-4 tools, 2x1 and 2x2 grids) that neither acronym
names on its own. Where a tooltip quotes a key, the key spelling is preserved and only the feature
word around it changed.
The `is_imex` tooltip is reworded rather than substituted: it already named the hardware families
parenthetically, so a literal replacement would have said IDEX/IQEX twice in one sentence.
No translation impact — no IMEX string had reached OrcaSlicer.pot or any catalogue, so there is
nothing to migrate. One test asserted on the old error text and now matches the new one.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comments 13, 14 and 15, and adds the test for a shipped-profile
regression that nothing guarded.
- 14 and 15: compute_imex_slice_offset had eight tests on the calculation and none
on the result, which is the whole firmware-managed path. test_imex_slice_offset
now covers the derivation end (which config produces a non-zero offset, and that
it is plate-local rather than moving with the plate origin -- the bug that
shifted every plate after the first) and the consumption end (emitted
coordinates and first_layer_print_min/max both move by the derived amount).
The first_layer case also cross-checks the two consumers against each other: the
declared bounds must keep the same relationship to the emitted toolpaths in both
frames, which fails if exactly one of them is shifted. It deliberately does not
pin the size of that gap -- it is 2.225 mm here, set by the wall generator, the
same with no offset at all, and pinning it would fail on an unrelated change.
- 13: nothing exercised the imex_mode / imex_mode_index / imex_mode_gcode
placeholders or the {global} flow into machine_start_gcode that their ordering
exists to guarantee. Seven cases now do, including the ordering itself -- the
mode script declares a global and machine_start_gcode reads it back, so moving
the mode processing later leaves the variable undefined and fails the export --
plus the inert cases (Primary mode, and a printer with the table filled in but
is_imex off). All matching is whole-line, because the config block the exporter
appends repeats machine_start_gcode verbatim and would make substring checks
meaningless.
- New: GCodeWriter passes this->config.is_imex.value into the heater remap, and
nothing tested that it passes the flag rather than a constant. Hardcode true
there and the whole suite stays green while fdm_bbl_3dp_002_common, which ships
physical_extruder_map [1,0], starts sending filament 0's M104/M109 to heater 1.
The new case runs a two-nozzle non-IMEX printer with that map and asserts each
filament's temperature reaches only its own tool. It uses idle_temperature via
ooze prevention rather than nozzle_temperature: keys in
filament_options_with_variant are re-indexed per filament by variant slot at
apply time, and this harness pins nozzle_diameter to one value, so every filament
resolves to the same slot and the temperatures stop telling the heads apart.
Also fixes two weaknesses in tests added earlier in this branch: an assertion that
would have been prefix-satisfied by the very routing it was meant to exclude, and
a whole-file command comparison between two slices, which this slicer's output is
not stable enough to support.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 20; the other two are non-IMEX leaks found auditing the
branch.
collect_imex_warnings re-derived the active mode, its tools, the primary and the
filament routing that Print::validate also derives, so the warning and the hard
block could drift apart -- and this function had already had one index-confusion
bug, the AFC/MMU wrong-filament names fixed in fbc58d2a1d. Both now read
imex_resolve_routing() and derive nothing themselves. The function stays
file-static: what is left in it is PresetBundle lookups, bed-type resolution and
formatting, none of which can disagree with the slicer about what the plate is
doing, and the index-confusion surface is now library code with tests covering the
AFC manifold in both directions. A latent out-of-bounds read went with it -- the
primary fallback can return -1 and the bounds checks were upper-only, so
filament_presets[-1] was reachable on a profile whose roster names a head absent
from the map.
Slicing eagerly re-rendered every plate thumbnail on the main thread after
switching to Preview, up to two blocking offscreen GL renders per plate on every
slice click. The work was already redundant: select_view_3D("Preview") invalidates
the thumbnails and marks the toolbar dirty, and the next frame force-regenerates
them anyway. It could not have served its stated purpose either, since it ran
immediately after reslice(), which only starts the background slice. Both calls
dropped; the plate badge state is recomputed per frame and is unaffected. The
export_3mf thumbnail log lines are back at info, and the slice-event traces
restored.
The Moonraker device sync wrote physical_extruder_map into the edited printer
preset ungated, so any Klipper machine running a current AFC build had its preset
marked dirty with no user action, and saving persisted the map into every 3MF
after. The map is indexed by logical extruder while the device reports one entry
per lane, and nothing in the lane payload carries the logical slot, so the two
index spaces coincide only when the counts match. Now gated on is_imex, written
only when the counts agree, and only when the value actually differs.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 17.
GCodeViewer had its own copy of the flip_x/flip_y corner mapping, the active
column/row sets, the physical-to-zone index mapping and the zone pitch -- the same
derivation as the plate's, with nothing keeping the two in step. It now calls
compute_imex_zone_layout() and consumes head_zone_centers. The mirror is expressed
as a reflection about the midpoint of the two zone centres rather than about a
zone-relative strip width, which is algebraically identical for equal-sized zones
and needs no pitch, and the toolhead-box face is chosen by comparing zone centres
instead of physical columns.
The July report of a math error in the visualizer for non-primary heads was this
drift: the sec_box_offset_y else-branch hardcoded -imex_box_wy, which happened to
equal the primary's offset on the rear-* layouts and pointed the wrong way on the
front-* ones. Structurally unreachable now.
Verifying the two sides matched turned up two config defaults that disagreed, both
fixed in their own commits: imex_nozzle_clearance_x/y (the viewer's 30.0f matched
PrintConfig, the zone code's 0.0 did not, and the strip loops are gated on it) and
imex_tools_per_gantry (the library's 2 matched, both GUI paths used 1).
Consuming the shared function meant resolving it per frame, and the sequential-view
marker flag is sticky, so one drag of the slider made every subsequent frame parse
five strings and allocate a dozen containers from inputs that never change. The
resolve now sits behind a cache key mirroring PartPlate::build_imex_cache_key(),
plus the two inputs only the preview reads -- the bed extents and the tool layout.
An idle frame compares scalars and allocates nothing. The toolhead-box mesh, which
was being re-uploaded to the GPU every frame for the same reason, is rebuilt only
when the clearances change.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 9.
PartPlate::reset_imex_mode() had no callers, while the PR description said
deleting a mode resets affected plates to Primary. IMEXModesCtrl now exposes an
on_mode_removed callback that TabPrinter::build_fff handles by resetting every
plate whose mode matches the deleted row, under a single undo snapshot, followed
by the same dirty/update sequence the plate's own mode button runs.
Wired to deletion only, on purpose: the name field notifies on every keystroke, so
routing renames through the same path would orphan and reset the plate on the
first character typed. Renames stay covered by the slice-time fallback and its
warning. The callback is copied to a local before notify(), because notify()
reaches load_from_config() -> clear_rows(), which tears down the row the handler
is running inside.
The rest of this file is the modes editor moving out to its own translation unit,
leaving the include and the construction site.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comments 21, 4 and 12, and the widget half of 19.
- 21: IMEXModesCtrl was 572 lines inside Tab.cpp. It now lives in
IMEXModesCtrl.{hpp,cpp} next to IMEXFilamentPickerPopover, which was the
precedent named in the comment. The move itself is exact -- member order,
comments and every string literal unchanged -- and the class had no file-local
dependencies in Tab.cpp, only its include list, so the new source states those
explicitly.
- 4: a mode row with an empty Name was silently dropped on save, tools and G-code
with it, and matches_config() compared against that same filtered output so the
preset never went dirty and the row stayed on screen. Rows are now given a
generated unique name instead of being discarded, and add_row() pre-fills one so
the common path never produces a blank. Names are deliberately not translated:
objects store a mode name in imex_parallel_mode and GCode.cpp matches it by
string, so a localized name would break a project reopened in another language.
- 12: the editor had a third parser that read a bare token and an unknown role
suffix as Primary, while parse_imex_active_tools reads both as Copy -- so the
editor and the slicer could read one imex_mode_active_tools string two different
ways. Deleted; the editor now uses the same two helpers the slicer does.
- 19: tile state was an int shadowing ImexRole, with the role letters duplicated in
a second switch that wrote the on-disk format. The tile now holds
optional<ImexRole>, with Inactive spelled as the absence of a role rather than a
fifth integer, and the letters come from kImexRoleTable.
Four further changes, from testing rather than the review:
- Deleting a mode reported only the row's current name, so renaming a mode and then
deleting it left every plate using it stranded on a name that no longer exists.
Both the build-time and current names are now reported, minus any a surviving row
still carries.
- The instruction text and colour legend were built once in the constructor and
never rebuilt, so raising gantry count to 2 gave the tiles a Span role the legend
never explained until the preset was saved and the page reopened. Both are
rebuilt with the grid, and the per-role detail moved into legend tooltips so the
panel no longer opens with a paragraph.
- The tile holding Primary is now read-only. Primary is tool 0 and moves only via
Tool 0 Position; a click could previously demote the only Primary, leaving a mode
that parses to no primary at all, which degrades the plate to an ordinary
single-tool print with nothing in the editor showing what is wrong. A mode
arriving without a Primary keeps every tile live so it can still be repaired.
- Names and G-code were read with ToStdString() (the ANSI codepage on Windows) and
written with from_u8() (UTF-8). On a non-UTF-8 codepage a name like "Modus A"
with a diaeresis was stored as invalid UTF-8, came back blank, and was then
silently renamed by the auto-naming above. Every read is now into_u8() and every
write from_u8(); EditGCodeDialog was affected in both directions.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
calc_imex_zones() is now 69 lines: fetch the two edited configs, call
compute_imex_zone_layout(), store the result, and clip each returned rect to the
bed outline to build the GLModels. That last step is the only part that needs GUI
types, which is why it stayed. See the extraction commit for the behaviour-
preservation evidence.
refresh_imex_slice_offset() is deleted along with its call in
update_slice_context(); the offset is derived in the engine now, and it was
computing it in the plate-list world frame, which double-counted the plate origin
for every plate after the first.
Two smaller changes:
- The zone/ghost cache key omitted imex_tool_layout, which decides which physical
corner tool 0 occupies and therefore moves every zone rectangle, collision strip
and ghost offset while every other keyed field stays put. A layout change
produced an identical key. That this currently appears to work is incidental --
some other path happens to rebuild -- and not something to depend on. Found by
building the preview's own cache key against this one.
- The tools-per-gantry fallback for a missing key was 1 in two places where
PrintConfig registers 2 and the zone code uses 2. All four sites now agree; a
missing key otherwise grouped tools against a grid divided a different way.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
None of these came from the review; they were found auditing the branch for the
same class of leak review comment 6 identified.
physical_extruder_map was normalised through effective_physical_extruder_map on
every Print::apply(), so any printer with more than one nozzle and no authored map
got the identity [0,1,...,n-1] where the single-element {0} default belongs. The
key carries two readings: the IMEX paths index it by logical extruder and need one
entry per extruder, while the inherited BBL paths read it through the clamping
get_at(), for which {0} means "everything is physical 0". Deriving unconditionally
imposed the IMEX reading on profiles that mean the other one, changing the config
block line and the {first_tools} / {first_filaments} / {curr_physical_extruder_id}
placeholders for multi-nozzle non-IMEX printers. Gated on is_imex; every consumer
needing the per-extruder form is already IMEX-gated, and profiles with an authored
map of the right length are unaffected either way.
That gating unmasked a latent out-of-bounds read: WipeTower's M104/M109 emitters
index m_physical_extruder_map by tool with no bounds check, which reads past the
end of the single-element default on a multi-nozzle machine. Upstream's bug, from
the BambuStudio wipe tower sync, previously hidden because the map was being
widened for everyone. Now bounds-checked, falling back to the tool's own index --
the form GCodeProcessor already uses for the same map.
Preset::save() and get_preset_differed_for_save() carried a branch storing the
full vector whenever a child and its parent had different lengths. It was written
against a set_with_nil that threw on mismatched sizes; upstream #13035 replaced
that with a tolerant version that keeps the child vector verbatim and nil-marks
only the overlapping range, and that fix was already in the tree when this branch
was rebased. Left in, it defeated the delta encoding for every user printer preset
whose extruder count differs from its parent's: a 7-extruder profile inheriting a
single-extruder base wrote out all of its per-variant retraction keys as literals,
including ones identical to the parent, pinning them against future vendor updates
while the UI still reported the preset as inheriting. Removed; the two save loops
are now identical to upstream. Note this only affects new saves -- presets already
written keep their frozen values until re-saved.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 1.
imex_parallel_mode and imex_head_filament_map were streamed raw into XML attribute
values, while every other free-text attribute in the same writer goes through
xml_escape. Mode names are free text, so "PLA & ABS", a quote or a "<" made the
document malformed. The failure is not a bad value on reload: both load paths for
model_settings.config return false on an expat error, and m_is_bbl_3mf is set
before the second entry loop runs, so the whole project fails to open with
"Archive does not contain a valid model config".
Both attributes now use xml_escape_double_quotes_attribute_value(), which also
emits tab, CR and LF as numeric character references. That matters and plain
xml_escape would not do: XML normalises literal whitespace in attribute values on
read, so a tab in a mode name would come back as a space and silently rename the
mode. The read side needs no change -- it takes expat's already-decoded value with
no second unescape -- so this is a lossless round trip and a file written by the
new code still loads in an older build.
The round-trip test used "copy_mode", which exercised none of this; it now carries
&, <, a quote and a tab, and also pins that ' and > come back unmodified, since
both are legal raw inside a double-quoted value.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 6.
The per-tool pressure advance work changed set_pressure_advance() for users who
are not using the feature. RepRapFirmware lost its D qualifier when no tool index
was supplied: upstream emits M572 D0 S<pa> unconditionally, and a bare M572
applies to whatever tool is currently selected and errors when there is none, so
PA started depending on tool-selection state for every RRF user. The D is back,
defaulting to 0, and D<tool> is reached only from the IMEX paths.
The same rewrite had also changed the comment separator from "<value>; Override"
to "<value> ; Override" on the Klipper, RRF, Marlin 2.x and Marlin Legacy
branches, so every non-IMEX print of those flavors carried a one-byte diff.
Restored. Upstream is internally inconsistent here -- BBL and Repetier do use the
spaced form -- and the point is to match it exactly rather than to tidy it.
Emitted output for all six flavors with no tool index is now byte-identical to
upstream. Verified on a real slice: a Klipper profile emits
"SET_PRESSURE_ADVANCE ADVANCE=0.02; Override pressure advance value", an exact
string match, with no EXTRUDER= qualifier. The tests were pinning the regressed
form and are inverted.
Also records at the imex key registrations why they are kept out of the g-code
config block, matching the house convention at the other banned keys.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 2, part of 18, and one non-IMEX regression the review did
not cover.
An imex_parallel_mode naming no entry in imex_mode_names still entered the
parallel branches. get_imex_active_tools() returned empty and the else was
skipped, so no head received its 1st-to-2nd layer temperature transition, and
imex_suppresses_bare_toolchange() still dropped the initial T<n> on the
expectation that a mode script would select the tool. validate() did not catch it
because its guard is declared_primary >= 0 and an unresolved mode yields -1. You
reach it by renaming a mode after a plate is set to it, or by opening a 3MF whose
printer preset names its modes differently. m_imex_parallel_mode is no longer
assigned before the lookup; a non-Primary name that matches no row now warns and
re-resolves against the Primary row, which is the fallback the PR description
already claimed. A mode that resolves to an empty tool roster takes the same path,
since the emitted G-code is wrong in the same way.
Warned rather than blocked: opening someone else's 3MF is a legitimate way to get
here and the Primary reading prints correctly, so refusing to slice would turn a
recoverable situation into a dead end. Silent was not an option either, because
the plate keeps showing the stale mode name while drawing no zones.
Separately, the 14 imex config keys all register non-nil defaults, so
append_full_config was emitting "; imex_* = <default>" into every exported
G-code, including on single-nozzle printers with nothing to do with the feature.
They are banned from the dump, matching the treatment already given to the
fast-purge, extruder-change and timelapse keys, so the config block is
byte-identical to the pre-IMEX baseline for the whole shipping fleet. Nothing
reads them back: GCodeProcessor has no imex reference, and the two per-plate keys
round-trip through the 3MF's model_settings.config on an independent path.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 11, and fixes a worse bug found while doing so.
The offset was only ever pushed from PartPlate::refresh_imex_slice_offset(),
reachable from update_slice_context() and the plater -- both GUI-only, and it
dereferences wxGetApp(). A headless slice therefore kept Vec2d::Zero(), so
orca-slicer --slice on a plate with imex_firmware_managed_zones emitted
slicer-managed coordinates while the firmware applied its own offsets on top.
Print::update_imex_slice_offset() now derives it from the applied config and runs
from process() and export_gcode(), so a CLI slice gets the value a GUI slice does.
It reads m_full_print_config rather than m_config because imex_tool_layout and
imex_carriage_margin are printer-preset options with no member in the static
PrintConfig, and it takes the mode from the same place GCode.cpp resolves it, so
the shift cannot disagree with the mode that is emitted.
The GUI push and Print::set_imex_slice_offset() are deleted rather than kept as an
override, because the two did not agree. calc_imex_zones() divides
get_extents(m_shape), and set_shape translates m_shape by the plate position, so
the pushed offset carried the plate origin -- which translate_to_print_space() and
the writer offset already subtract. Plate 1 sits at the origin and agreed by
accident; every later plate had the origin subtracted twice and was shifted by a
full plate stride. Not silent, either: the displaced geometry fell outside the
printable area, so slicing plate 2 failed validation with "part is off the plate".
Confirmed fixed on hardware profiles -- the same model on plates 1 and 2 now emits
identical extents.
Deleting the push also removes the post-apply ordering constraint that forced the
duplicate call in Plater::priv::update_background_process: the value is computed
at the point of use, and process()/export_gcode() are structurally after apply().
Also routes validate()'s primary-routing check through imex_resolve_routing() so
the hard block and the plater's warning cannot describe a plate differently
(review comment 20), and through find_imex_mode() for the mode lookup (18).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comment 16, and fixes three defects the move exposed.
PartPlate::calc_imex_zones() was 392 lines deciding where every zone, collision
strip and safety margin sits, in the GUI layer, with no test coverage. The three
wxGetApp() calls that kept it there were all in its first 25 lines, fetching two
configs. The geometry now lives in compute_imex_zone_layout(); the wrapper fetches
the configs, calls it, and clips the returned rects to the bed outline for the
GLModels, which is the only part needing GUI types. libslic3r gained no wx
dependency: it takes DynamicPrintConfig directly, so the option lookups moved
verbatim rather than through a hand-written value struct that could drift.
The move is otherwise exact -- verified by a line-for-line diff of every
arithmetic expression against the original, and by running 15 scenarios through
the extracted code against hand-derived values. The only deletion is a lambda that
was never called.
Three fixes on top, each of which needed the code to be testable:
- An off-grid or absent Primary left pri_col/pri_row at their (0,0) initialisers
and built a layout from them, reporting the whole bed as the clear primary zone
and the whole bed as a blocked mirror zone at once; under
imex_firmware_managed_zones it shifted the slice by the bed centre. Guarding on
the resolved primary head covers both routes. Reachable only from a hand-edited
preset or a 3MF authored against another printer -- the editor pins Primary to
tool 0 -- but that is the same class the unresolved-mode fallback handles.
- imex_nozzle_clearance_x/y fell back to 0.0 where PrintConfig registers 30.0.
Both strip loops are gated on the value being positive, so the fallback emitted
no collision strips at all while the preview still drew 30 mm toolhead boxes.
- The collision-strip loop asked each mirror head for its own grid cell, but an
aggregated gantry's cell is pinned to the primary's column and expanded into a
full-width row strip. Where the representative's column differed from the
primary's, no boundary matched and the plate came back with no strips and no
margin bands -- an object flush against the shared boundary sliced without a
warning while the far carriage occupied it. Present since Span aggregation was
added in 4966d0fae8 and carried out of PartPlate verbatim. The flags now come
from the painted cells; an exhaustive sweep of the reachable grid, role and
layout space (1,630,720 configurations) shows the only behaviour change is the
missing strips appearing.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Closes review comments 3, 5, 7, 8, 18 and 19, plus the library half of 20.
These share a file, so they share a commit; each is independent of the others.
- 3: ::isspace(char) is undefined for bytes above 0x7F because char is signed on
our targets. Three call sites now go through one strip_whitespace() using an
unsigned char cast. Line 308 parses imex_head_filament_map straight out of 3MF
metadata, so a non-ASCII byte reached it without passing through the UI.
- 5: an imex_head_filament_map override past the end of physical_extruder_map now
falls through to the printer's own routing instead of resolving to a wrong
filament. Bounded in resolve_filament_for_head, where the slot count is known,
rather than at the parse site, which has no count to check against; the parse
site also gains the absolute MAXIMUM_EXTRUDER_NUMBER cap its sibling already had.
- 7: imex_physical_heater_for's !is_imex early return is what keeps a stock BBL
profile (physical_extruder_map [1,0]) out of the heater remap, and had no test.
Six cases now cover it, pinning pass-through rather than get_at()'s clamp.
- 8: ImexRole::Span was missing from the imex_head_transform switch, so it warned
under -Wswitch. Identity is correct, not merely convenient: a Span tool prints
the primary's own zone through mid-print toolchanges and has no zone to be
translated into.
- 18: three positionally coupled string vectors were resolved by nine open-coded
lookups using three incompatible bounds idioms. None read out of bounds, but six
folded the guard into the match condition, so a ragged row did not stop the scan
and a later duplicate name could win. struct ImexMode + find_imex_mode() is now
the only resolution rule: the names array is the roster, first match wins, a
short sibling pads to empty and sets ragged, not-found is an explicit -1.
- 19: the letters P/C/M/S existed in three independent copies, one of which was the
writer of the on-disk format. kImexRoleTable is now the single source, read by
both parsers and the serializer. Adding a role was 14 edit sites with one
compiler-enforced; it is now the enum, the table entry, and four -Wswitch
switches. Verified by adding a fifth enumerator and recompiling: exactly four
warnings, nothing else.
- 20: imex_resolve_routing() extracts the mode/primary/routing chain that
Print::validate and the plater's warning collector each derived separately.
The three config keys keep their names, types and on-disk representation. This is
a read-side view only; presets and 3MF files are unaffected.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Revolve failed on a sketch whose profile was closed. Decoding the reported 3mf: four
entities forming a proper closed loop (joints open by 4.44e-06 mm, well inside
tolerance) plus one stray 1.82 mm Line at (-24.2, 80.3), inside the shaded region,
touching nothing.
The viewport's region_loops discards open chains ON PURPOSE — it exists to find
EXTRUDABLE regions — so the user saw one clean closed region. entities_to_wires kept
the stray as its own one-edge loop, so it returned two wires, and Revolve goes through
entities_to_wire which demands exactly one. Extrude would have failed one step later in
wires_to_face, because a one-edge open wire bounds no face. Same class as the tolerance
split fixed in 8b568b7b: the viewport and the kernel disagreeing about the sketch — this
time about what BELONGS to the profile.
entities_to_wires/entities_to_wire/build_sketch_wire take closed_only. It is not a
blanket rule: a SurfaceExtrude builds a sheet FROM an open profile and a Sweep PATH is
normally open, so all ten call sites are classified individually — true for the face
fallback, Extrude-taper, Revolve, the Sweep PROFILE and Loft profiles; false for
SurfaceExtrude/Revolve/Loft/Fill and the Sweep path.
A component counts as open when some welded node has DEGREE 1. The first attempt used
"the traversal did not return to its starting node", which regressed the bridged C
profile: a closed loop that also carries a second edge across the same two nodes has no
free endpoint, but its Eulerian walk ends elsewhere. Degree-1 is the property that
actually distinguishes a stray segment from a closed profile; the suite caught the
difference.
Behaviour change decided by Tommaso: a stray is IGNORED, not refused. The test that
required refusal dates from when ignoring meant falling through to a default rectangle —
geometry nobody drew. That fallback is gone, so ignoring now builds the circle the user
actually drew. Its assertion is updated with the reason.
The bridge round-trip test extruded an ENTIRELY open chain and "worked" only because
OCCT will make a face from an open wire. It gets a genuinely closed profile: the test is
about serialization, and deserialize_recipe recomputes, so the document has to be one
that legitimately builds.
Failures now say WHERE. sketch_open_ends reports free endpoints under the same weld
tolerance the wire build uses, and open_loop_message is shared by both throws, because
Extrude fails through build_sketch_face and Revolve through build_sketch_wire — enriching
only one would have left the commoner path the less informative one.
Kernel 66115 assertions / 608 cases green.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Tommaso asked the question that names the real defect: if the sketch was open, why was
the same sketch shaded closed and offered for extrude? Because the two halves used
different tolerances. region_loops shades a region closed at 1e-3 mm; connected_loop
chained at 1e-3; the kernel welded at 1e-4 and OCCT matched vertices at 1e-7. The
2.28e-5 mm gap in the reported sketch did not cause that disagreement, it only made it
visible — and fixing the gap alone would have left the contradiction in place, ready to
reappear anywhere in (1e-4, 1e-3].
kSketchJoinTol now lives in SketchEngine.hpp and is the only place the number exists.
region_loops, loop_report, connected_loop and entities_to_wires all read it through
sketch_join_tol(). The viewport cannot promise a region the kernel refuses to build.
The welding is optional, because a kernel that silently closes loops should let you say
no: "Auto-close sketch loops" in Preferences, default ON, no restart. OFF means only
exactly coincident endpoints join — and since both halves read the same value, the
viewport simply stops shading the region closed, so an open loop is visible rather than
welded behind your back. No separate UI needed for that; it falls out of sharing one
number.
Details that matter. The kernel defaults to auto-close ON independently of the GUI, so
headless and MCP callers behave like the viewport instead of inheriting an unset
preference. With the tolerance at 0 the comparisons become <=, because OFF must mean
exact, not broken. OCCT never receives a zero vertex tolerance — it is clamped to
Precision::Confusion.
The preference is pushed from EVERY entry that starts a sketch session, not just
begin(): a Constrain session enters through begin_constrain / begin_constrain_entities
and uses region_loops and connected_loop, so a single push site would have left those
sessions running on whatever the previous one set. begin_imported_transform is excluded
deliberately — it works on imported regions, not chained entities.
Tests: a loop with one joint open by 9e-4 mm, given out of traversal order, builds a
closed four-edge wire; with auto-close off the same loop yields no wire; and an exactly
closed loop still builds with auto-close off, proving OFF means exact. Kernel 66104
assertions / 606 cases green.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
An extrude built a solid the user never drew: three sides of the handle plus the arc
that bulges outside the outline, with the bowl's second arc and the left edge missing.
Two faults met. entities_to_wires added edges in ENTITY-CREATION order, so a partial
wire rejects the next edge even when the sketch closes perfectly; and one joint of the
reported sketch is open by 2.28e-5 mm, wider than OCCT's 1e-7 vertex tolerance and
wider than this function's own EPS of 1e-6, so that edge was refused on geometry too.
Neither showed up, because BRepLib_MakeWire::Add DROPS a disconnected edge
(BRepLib_DisconnectedWire + NotDone) while every successful Add ends with
BRepLib_WireDone + Done() — overwriting the failure. `if (!wm.IsDone()) return {}`
was therefore asking only whether the LAST edge connected. Six edges in, four out,
IsDone() true.
Endpoints now weld into shared nodes at one tolerance (kSketchWeldTol) used by BOTH
the union-find grouping and the wire build — they disagreed before, which is how a
joint gets united into a loop and then refused by the builder. Each node becomes ONE
TopoDS_Vertex, so the builder matches on identity instead of proximity, with the
vertex tolerance widened because BRepLib_MakeEdge::Init projects a vertex onto the
curve within that tolerance and a welded node sits up to the weld gap off its
neighbour's curve. Members are then walked in traversal order. Finally the result is
counted: IsDone() alone is not evidence, edge_count == members.size() is.
Arc geometry is untouched — the midpoint from (start_angle+end_angle)/2 and the
solver's angle reflow both measured correct and were never part of this.
The regression case carries the reported sketch verbatim, open joint included. It
fails 4 == 6 without the fix, which was measured, not assumed. Kernel 66092
assertions / 604 cases green.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
The rows, their ✗ buttons and the click-to-highlight were all built against
m_doc.features[m_constrain_feat].entity_constraints — a COMMITTED feature. A live sketch
has no committed feature, so the card was hidden for the whole session and the list it
would have shown was empty by construction. Every constraint applied while drawing was
nameless: the badge said one existed, nothing said which.
rebuild_constraint_list now picks its source by scope. live_constraint_scope() is the same
discriminator apply_constraint already used to route to apply_live_constraint — both
Constrain modes set m_active, so is_sketching() alone would claim the live scope while the
committed manager is open. delete_constraint and highlight_constraint_entities branch on
it too, and the card shows in Sketch mode as well as Constrain.
Keeping the rows in step needed a signal that did not exist: on_solve_state fires on every
frame of a drag, so rebuilding from it would rebuild the list continuously. The tool now
fires on_constraints_changed only when the constraint SET changes — one added by
try_add_constraints, one removed by remove_constraint (the indexed form the badge click and
the ✗ row now share).
The rebuild is deferred through CallAfter. One of its callers is the ✗ button's own click
handler, and rebuild_constraint_list destroys those buttons: deleting the window whose
handler is still on the stack is a use-after-free.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
A glyph reads as decoration until something says otherwise, so the badges shipped
last commit were discoverable only by accident. Two places now say it, chosen because
they are where the eye already is:
- the moment of applying, which is the one the user is watching ("Applied constraint ·
its badge is on the sketch — click the badge to remove it"). The hint line could not
carry this alone: it only refreshes when the (mode, step, picks) tuple changes, and
applying a constraint changes none of them.
- the Select-mode hint line, appended only while the live sketch actually holds a
constraint, so it never advertises a badge that is not on screen.
Also fixes what the previous commit got wrong: remove_constraint_near solved through
solve_sketch_entities directly, which relaxes the geometry but leaves m_dof and the
per-entity conflict flags untouched and never fires on_solve_state. Deleting a
constraint therefore left the DoF readout describing the system as it was BEFORE the
deletion, and any red over-constrained tint stranded on screen. It goes through
resolve_live() now, the same path every other live edit uses.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
The constraint badges existed and had never once been drawn where they were needed.
build_constraint_glyphs read m_constrain_cons, a vector only the COMMITTED-feature
Constrain mode fills, and the draw call sat inside `if (m_mode == Mode::Constrain)`.
Every constraint applied during a live sketch — which is the path the Constrain buttons
take while drawing, the one added in "Constrain while you sketch" — went into
m_constraints and was rendered by nothing. You could not see that Parallel had applied,
so "nothing happens" was indistinguishable from "applied and invisible".
The glyph builder now takes its constraint list as a parameter: Constrain mode passes
m_constrain_cons as before, the live session passes its own m_constraints. Same glyphs,
same teal.
Seeing them is half of it. A constraint's entire state is exists / does not exist, so the
toggle is a delete, and there was no way to reach one during a session — the ✗ rows in
the panel list are bound to the committed feature. Each badge now records where it landed
(m_glyph_hits) and a plain left click in Select mode within its cell drops that constraint
and re-solves. Shift/Ctrl clicks are left alone so multi-select still works.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
Two reports, one root: the sketch tab could not show what was selected.
The bed grid landed last commit, and selection was painted pure white — a freshly
drawn line is auto-selected by the creation tool, so the first thing a new line did
was disappear into the grid. Selection now wears design_selection_color(), the same
cyan a picked solid already wears. White is kept for the hover handle alone.
That invisibility is also why "I apply Parallel and NOTHING HAPPENS": drawing two
lines leaves exactly ONE selected (the last), Parallel needs two, so the planner
correctly refused — but the status text named only the requirement, never the current
pick, which reads as a dead button. It now reports how many are selected and how to
pick the second.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_011FbJKJAJxxkhDTs9XdZzKA
In CAD the centre IS the sketch origin -- GLCanvas3D already moves the axis
triad there for exactly that reason. The grid under it did not agree: it comes
from PartPlate::calc_gridlines, generated from m_origin, the plate's front-left
corner, with an adaptive step meant for a print bed.
Measured on a screenshot from the user's machine: the nearest grid line was 10 px
from the origin in a 23 px pitch. The origin floated mid-cell, in both axes.
Corner-origin is CORRECT for Prepare -- a print bed starts at a corner -- and the
plate list is SHARED with the plater, so re-centring it there would change the
bed for every user of the app to fix one tab. The seam used instead already
existed: _render_platelist takes show_grid, and m_axes_at_bed_center is already
the "this is the Design canvas" flag. The Design canvas suppresses the plate's
grid and draws its own.
Minor every 10 mm, major every 50 mm, both generated from bed_center() so a line
passes exactly THROUGH the origin in each axis. Two GLModels, rebuilt only when
the bed shape changes, not per frame.
White majors, grey minors, the SAME in both themes. There is no white bed to
vanish against: the plate is dark grey either way (DEFAULT_MODEL_COLOR
{0.326,0.337,0.337} light, DEFAULT_MODEL_COLOR_DARK {0.255,0.255,0.283} dark),
a difference of 0.07. An earlier draft inverted the palette on the light theme;
that was a branch buying nothing. For contrast with what this replaces: the
plate's grid draws BOTH its thin and bold families in one 0.43 grey, which is
most of why the stock grid reads as a flat mesh with no scale to it.
z = -0.26, the same value as PartPlate::GROUND_Z_GRIDLINE -- below the bed fill
at -0.03, above the bed model at -0.41, so no z-fighting. Matched by
construction, since that constant is file-static in another TU.
Known limit, commented: the grid is clipped to the bed's BOUNDING BOX, not its
polygon. Identical on a rectangular bed; on a circular one it would spill past
the round edge. The target printers are rectangular.
Verified on the rig, not just compiled: white majors over a fine grey mesh, and
a white line through the origin in both axes.
snaporca-kha0
Twenty buttons on the CONSTRAIN bar, nine of them exercised. The other eleven
were "implemented" in the sense that the kernel builds the right def for them --
which is exactly what was true of Parallel yesterday morning, right up until a
user pressed it and got nothing.
What a kernel test cannot see is whether the BUTTON is wired to the index its
name claims. CON_BTN is 449 + 42*i over a hand-written name list, and it has
drifted once already: six buttons were inserted, everything from index 6 on
pointed at the wrong control, and nothing caught it for months because no rung
pressed past index 5. D13 presses index 6; D14 through D22 press 8 to 19. The
map turned out to be intact, which is worth knowing rather than assuming.
D12 vertical, D13 equal_radius (its own button, not Equal's promotion),
D14 concentric, D15 tangent, D16 midpoint, D17 symmetric (the three-pick form),
D18 sym_h, D19 radius, D20 diameter, D21 fix, D22 dist_y.
Three are shaped around a specific way the code could be wrong rather than
around "does something happen":
D20 exists for a factor of two. Diameter wired to the Radius handler gives
r = 30 for a typed 30, and nothing on screen looks wrong.
D21 -- Fix alone is unfalsifiable: nothing moved, so nothing proves the
constraint exists. It only becomes observable when a SECOND constraint would
otherwise move the fixed point, so the rung drives the pair to a 70 mm gap and
checks which end travels.
D14 asserts the radii did NOT change. Concentric is about centres; a solve
that also equalised the radii would pass a naive check.
D17 failed on its first run and the rung was wrong, not the app: all three
entities are free, so the solver is entitled to satisfy the mirror by moving the
AXIS instead of the points -- and it did, landing the pair symmetric about
x = -8.18. It now measures signed perpendicular distance to the axis where the
axis actually is, which is the stronger property anyway.
Coverage: 20/20 buttons pressed, up from 9. Ladder 135 -> 177 properties across
43 rungs, all holding.
snaporca-l2vm
Five defects from ten minutes of real use, and the mode split behind the worst
of them. One commit because the changes overlap in the same functions; the
pieces are separable in the diff, not in the file.
CONSTRAINING NO LONGER NEEDS A COMMITTED SKETCH. A constraint could only be
applied by committing the sketch, selecting it in the feature tree, pressing the
padlock, and only then picking. While drawing, the CONSTRAIN toolbar was not even
on screen (set_ui_mode showed it in UiMode::Constrain alone) and apply_constraint
answered "Press Constrain on a sketch first" -- in a status line nobody looks at.
Draw two lines, press Parallel, get nothing: that is what a user reported as "the
UX is a mess", and they were right.
apply_constraint now takes the live session first, reading the picks from the
selection model the sketch tool already had (click, ctrl-click to extend,
double-click for the loop) and applying through try_add_constraints, which
already did append -> solve -> keep-or-rollback. The committed Constrain path
stays for editing an old sketch; it is no longer the only way in. The twenty
constraint buttons now show in Sketch mode as well.
The discriminator is is_sketching() && !is_constraining() &&
!is_constraining_entities(). Both begin_constrain and begin_constrain_entities
set m_active, so is_sketching() alone is true DURING a constrain session and the
new path would hijack the old one -- compiling perfectly and failing in
behaviour.
ONE PLANNER, NOT TWO. A second caller meant duplicating the logic that decides
whether a constraint is legal, which roles it binds and whether it needs a typed
value. That duplication is how today's Coincident bug survived: fixed in one
branch, alive in the next one down. plan_entity_constraint() now lives in the
kernel -- pure, no wx, no translation -- and both UI paths call it. DesignPanel
loses 304 lines and gains 155.
Being in the kernel makes it TESTABLE. The Parallel defect existed because a
constraint type met an entity type nobody had tried, and the only instrument was
a 13-minute GUI ladder. 19 new kernel cases cover the matrix: 264 -> 283 cases,
7648 -> 7867 assertions.
Parallel, Perpendicular and EqualLength gain the two-line guard they never had.
On non-lines they used to emit a def the solver silently dropped -- the sketch
reported itself constrained when it was not, the same class as Horizontal on a
Point. EqualLength on two rounds still promotes to EqualRadius first.
Symmetric is planned completely, including its axis pick: the plan carries a
VECTOR of defs because Symmetric on two lines is two constraints (P0/P0 and
P1/P1). A single def would have half-applied it -- one end pinned, one free,
looking correct until something moves.
A PLACED POINT SURVIVES THE COMMIT. Type::Point was created correctly and never
drawn once committed: both renderers skip it, correctly, since entity_polyline
gives a point nothing. What was missing is the vertex-marker path the live
session already used. rung_point passed throughout because it asserts the
document, and the point was always in the document -- the pixels lied.
ESC STOPS EATING AN UNSAVED SKETCH. The third press reached cancel_sketch(),
clearing m_entities with no warning and nothing to undo. live_sketch_has_work()
existed and was never consulted. The exit layer refuses once when there is work
and lets a second consecutive Esc through; the refusal re-arms on a button press,
never on mouse motion, or Esc could never exit while the hand moves.
TWO NEW RUNGS. D10 drives Parallel through the committed path -- it passes on
the PRE-fix binary, which is how we know the user's failure was the mode and not
the constraint. D11 is the acceptance for the collapse: draw, pick both, press
Parallel, no commit and no padlock. Ladder 126 -> 135 properties, all holding.
CON_BTN_SKETCH is measured, not derived: in Sketch mode the group renders after
the sketch toolbar, so the first button is at 677, not 449. Pitch 42, twenty
buttons, read off a screenshot. Deriving it by offset is how that table drifted
the last time.
Known limit, commented at the call site: a constraint added to a LIVE sketch is
not on the document undo stack, so Ctrl+Z will not take it back until the sketch
is committed.
snaporca-itp4, snaporca-oyhx, snaporca-l2vm
Port of snaporca da8d011b87; parity holds (DesignPanel.cpp still exactly 32 divergent
lines). Verified independently on this fork's own rig: full ladder 126/126 against
BuildID 96c697a3, built from this tree.
Reviewing the DistanceX/Y fix for OTHER members of its class found three more live defects
on the constrain toolbar. All four share one root: a branch assumes every picked entity has
two endpoints, and the solver's refusal to resolve a role it cannot find is silent.
COINCIDENT had the identical closest-pair walk over {P0,p0},{P1,p1}. For two Points the
phantom (0,0) pair sits at distance 0, which is the smallest distance there is, so it
ALWAYS won: ptOf(Point,P1) -> 0, ref_ok fails (SketchSolver.cpp:185), constraint dropped.
Not sometimes -- every press.
HORIZONTAL/VERTICAL hardcoded ra=P0, rb=P1 with no type check. With a Point picked the
constraint is dropped by the same mechanism but still STORED: constraints goes 0 -> 1 after
the commit and nothing moves, so the Constraints list shows a dimension that can never do
anything. Worse than refusing -- the panel claims the sketch is constrained when it is not.
ANGLE computed p1-p0 on whatever was picked. On a circle that is (0,0)-centre, so two
circles pre-filled the field with the angle between their centre POSITION VECTORS (178.83
deg for two on the x axis), and accepting it emits SLVS_C_ANGLE on two circle prims.
Both branches now refuse with a message. entity_ends()/closest_ends() are file-scope and
shared by Coincident and DistanceX/Y, so there is one implementation instead of two that
drift.
Two smaller findings from the same review: infer_auto_constraints' roles_of omitted
EllipseArc while heal_coincidences' identical copy has it; and set_point(Circle, Center)
wrote e.center and not e.p0, breaking the "p0 mirrors centre" invariant for the duration of
a live drag.
New rungs D8 and D9, both RED against the shipped binary and green here.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
Format/3mf.cpp also saved and loaded it, but nothing calls its store_3mf and
its load_3mf only sees files fingerprinted as PrusaSlicer's, which never carry
a recipe. Its round-trip test only exercised that dead loop. The BBS backend,
which every save and load goes through, is untouched.
The Design tab's viewport is the fourth GLCanvas3D on the shared GL context and
the only one the plater does not own, so unbind_canvas_event_handlers() and
reset_canvas_volumes() never reached it — the macOS Command+Q and Debian cases
those calls exist for. Its frame-level handlers become members so they can be
unbound.
The bind sat above GLCanvas3D::bind_event_handlers(), and on_size never Skips,
so wx's reverse-order dispatch stopped before it and the handler never ran.
Port of snaporca b3221f8a12; this is the fork the fault surfaced on.
The perpendicular rung drew its two lines 93.5 degrees apart and then asserted they did
NOT start perpendicular. On this rig, whose camera maps the same click a pixel differently,
they arrived at exactly 90.000000 -- inference had already done the job the rung exists to
test, so the precondition failed while every later check passed. Held on the other fork and
failed here from identical source: the rung depended on where a click happened to land, not
on the app.
The second point now starts the pair 56 degrees off, well outside any snap tolerance, so
the button has real work to do.
That immediately exposed a second, milder fault in the same rung. From a 51 degree start
the LIVE solve converges to its own tolerance and lands at 89.999999991; the old 1e-9
assertion held only because the correction used to be tiny -- it was measuring how little
work the solver had to do, not whether the lines came out perpendicular. It is 1e-6 degrees
now, which is 1.7e-8 radians. The round-trip check still demands exactly 90 and gets it,
because the committed feature re-solves from scratch.
Full ladder 118/118 on BOTH rigs after this, each driving its own fork's binary. This fork
had never had a green gesture ladder before today (snaporca-eoj1).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
Four of this fork's five absolute chrome coordinates were shifted by CHROME_DY when the
ladder was first brought up here (DESIGN_TAB, CONSTRUCTION_CHECKBOX, CON_BTN_Y,
CONFIRM_BTN). TREE_ROW0 was not, because it is declared above the CHROME_DY block and was
simply never in view.
The unshifted click lands 26 px below the first tree row, just past its 23 px height, so
the row is never selected and Delete does nothing. reset_document then spends 40 rounds on
it and dies with "could not empty the feature tree" — a message that names the feature
tree, which is not the fault. The same 26 px is why confirm_and_reopen's double-click did
not reopen the sketch, which surfaced as "sketch_describe: no sketch is open" three frames
away from the cause.
Measured, not inferred: the Sketch1 row centre reads y=241 on the rig at 1920x1080 with
the window at (0,0), against the constant's 215. CONFIRM_BTN was checked in the same pass
from a screenshot taken in CONSTRAIN mode and is correct at (1751, 101).
With this, the four new constraint rungs hold 20/20 on this fork's rig, driving the binary
built from 648b930e75 (BuildID 56417445) — so the DistanceX/Y fix ported here is now
exercised, not merely parity-checked.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
Port of snaporca 2951e3c60b; parity holds (17 identical, DesignPanel.cpp still exactly 32
divergent lines, so the hunk landed on the right side of the DropDown divergence).
The sketch-constraint epic added six toolbar buttons and covered all six with kernel
tests. Not one of them was ever clicked. The gesture ladder only pressed 'perpendicular'
and 'equal' -- and CON_BTN, which locates buttons by index, was silently wrong for every
entry past index 5 for the whole epic. The untested half of the toolbar was exactly the
broken half (snaporca-rqsy).
Four rungs now drive them: D4 Equal on two circles (must mean equal RADIUS, not the
equal-length no-op the epic fixed), D5 Collinear on two oblique lines, D6 a horizontal
distance, D7 symmetric about the implicit vertical axis. Full ladder 118/118 on snaporca;
this fork's rig has not been rebuilt against the change yet, so here it is reviewed,
parity-checked and NOT exercised.
D6 found a shipped defect. apply_entity_constraint enumerated {P0,p0},{P1,p1} for BOTH
entities regardless of type, but a Point's p1 is unused and reads (0,0), as does a
Circle's. The closest-pair search then picked those two phantom origins, distance 0: the
field opened pre-filled 0.00 and the solver dropped the constraint, because ptOf(Point,P1)
resolves to no handle. Nothing errored -- the dimension simply did nothing. ends_of() now
enumerates only the roles an entity actually exposes, and the pair with no point at all is
refused with a message instead of a silent no-op.
Five kernel tests, a 7/7 ladder, a review and a fork port all passed over this, because
every one of them exercises the kernel, where the geometry was always right.
Two rig faults fixed in the same pass, both of which produce a green-looking session that
tests nothing: start-headless-gui.sh never exported SNAPORCA_MCP or SNAPORCA_KEYTRACE, so
a freshly launched rig comes up healthy and every ladder dies on "Connection refused"; and
it never dismissed the "Restore" dialog a killed session leaves behind, which grabs every
synthetic click afterwards.
The value field also takes no keyboard focus from the WM -- typed digits go to the canvas
and Return commits the pre-filled number (typed 40, got 54.94). focus_field() finds it as
its own top-level window and clicks it first. The no-op tolerance is now 5e-3, the field's
own two-decimal display resolution, not 1e-6.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
This fork had never been gated end to end. Five separate things stopped it, none
of them a defect in the CAD code itself, and each failure named the wrong
subsystem — which is why they survived.
1. run-all-checks.sh invoked docs/ux/mockups/gen_offer_table.py. The design docs
moved to docs/CAD/ (bbd1989e1e) and this path did not follow, so the rung
failed on a missing file.
2. gen_offer_table.py then resolved REPO one dirname short, because it now sits a
level deeper. OUT pointed at docs/src/.../DesignOffer.hpp, which does not
exist, so --check diffed the generated table against an EMPTY file and
reported all 189 lines as a difference.
3. DISPLAY was never passed into the container. The check scripts fall back to
":10", which is the other fork's rig; this one's Xvfb is :11. Symptom:
"FATAL no app window on :10", which reads like a dead app.
4. Every absolute chrome coordinate was written in the Snapmaker fork's layout.
This fork keeps mainline's top row (File / save / undo / redo / Calibration
with the title), so the whole chrome sits 26 px lower. At the unshifted y the
Design-tab click landed in the toolbar and the app stayed on the Home page,
reported as "no sketch opened after plane click + Shift+S"; the unshifted
Construction checkbox reported "0 construction axis". The three constants now
derive from CHROME_DY. Canvas coordinates were never affected -- clickmm()
computes them from live canvas geometry -- which is why dozens of geometric
properties passed exactly on a GUI that had never been driven.
5. CON_BTN was stale for every index after 5, from the sketch epic's six new
buttons. Fixed in both forks; see the companion commit on snaporca.
pdftocairo was also missing from the rig image (installed there, not a repo
change), without which every corpus sheet threw.
RESULT: offer-table, kernel, engine, corpus, corpus-scale and offer all hold.
The gesture ladder now reaches D2 and applies Equal length through the Constrain
toolbar, both sides landing at 99.928133658; it then fails re-entering the sketch
after commit, filed as snaporca-eoj1 with the evidence and the next measurement
to take. Kernel here is 7701 assertions / 270 cases.
update_values_from_multi_to_multi_2 iterates the destination PRINTER's variant
list while writing into a row taken from the destination PRINT preset. Those two
lengths are maintained independently -- print_extruder_variant against
printer_extruder_variant -- and Tab::load_current_preset() runs the migration
before the print preset is re-selected for the new printer. Opening a project
saved on a single-variant printer and switching to a seven-variant one therefore
wrote six elements past the end of a one-element vector. The corruption stays
silent until the next allocation, so the abort surfaces somewhere unrelated and
the backtrace points at innocent code.
Size the row to the variant count before indexing it. Every write is then in
range, and the result carries one value per destination variant, which is what
the callers consume. Pad with nil rather than a copied value: set_to_index()
skips nil entries, so a variant the object has no opinion about keeps tracking
the print preset instead of being pinned to another variant's number.
The same shape -- a count from one array indexing another -- appears twice more
in this file. update_values_from_multi_to_multi has three of these writes
protected only by assert(idx < old_count), and NDEBUG is defined for every
non-Debug configuration, so those guards are absent from shipping builds.
update_values_from_single_to_multi has the read half. Both are bounded here;
leaving them would fix one third of one defect.
Source reads are bounded too. is_nil(size_t) indexes values[idx] without
checking, so an index past the end was undefined behaviour on that side as well.
Where the row already matches the variant list -- every case that was not
corrupting the heap -- the resize is a no-op and the output is unchanged.
Fixes#15455
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This fork's kernel suite has never run. scripts/CAD/run-kernel-tests.sh died at
CMake CONFIGURE time on find_package(assimp REQUIRED), before a single source
file compiled, so every kernel change ported here was parity-checked against
snaporca and never independently tested (snaporca-w80c).
WHY IT WAS MISSING. OrcaSlicer mainline gained assimp (glTF/GLB/FBX import for
texture-to-colour) after the orcacad-deps image was baked: the image's deps/ tree
has no Assimp directory at all and nothing named assimp anywhere in it. On the
host, deps/build/dep_Assimp-prefix carries only `patch` and `update` stamps -- no
build, no install -- so the dependency was fetched and then never built, inside
the image or out of it. The Snapmaker fork never hit this because its base
requires neither assimp nor OpenCV.
WHY A LAYER. A full deps rebuild is hours and would rewrite artifacts that
currently work; this adds the one missing package on top. It is a Dockerfile
rather than a `docker commit` so that what was done stays reviewable and
repeatable instead of being an undocumented image mutation.
The flags are the project's own recipe (deps/Assimp/Assimp.cmake) plus the
standard superbuild arguments from orcaslicer_add_cmake_project
(deps/CMakeLists.txt:158) and DEP_CMAKE_OPTS (deps/deps-linux.cmake). The file
says to keep them in step with that recipe: it stands in for the superbuild, it
is not a separate opinion about how to build assimp.
The tarball's SHA256 was checked against the recipe's URL_HASH before this was
written and is re-checked inside the build, and the build asserts the installed
cmake config exists rather than trusting an exit code. OpenCV was confirmed
already present, so it is not a second wall behind this one.
RESULT, and it is the point: orca_cad kernel now runs and is GREEN at 7701
assertions / 270 cases. snaporca is 7700 / 270 -- same cases, one more assertion
here, which is the tolerated test_caddocument.cpp divergence. The "this fork's
kernel suite cannot run" caveat carried by the five commits of the sketch
constraint epic no longer applies.
Port of snaporca e635627b81. Parity OK: 17 files identical, 8 diverging at their
expected counts.
The last reference an industrial sketcher offers that this one did not: Onshape's
Use, SolidWorks' Convert Entities. Project already turned a body's 3D edges into
2D Lines, Circles and Arcs on a plane, but the result landed in its OWN feature,
so while drawing in one sketch you could not borrow an existing body's edge and
constrain to it.
No new geometry code: Project's per-edge conversion loop is factored into
project_edges_to_entities() and called from a second entry point that appends
into an EXISTING sketch with construction = true. The loop appears once now,
not twice.
Construction is what makes it cheap and safe: SketchEngine already skips
construction entities when building wires, so the references guide without being
built, and being otherwise ordinary entities every constraint from this epic --
Collinear, EqualRadius, the axis-projected distances, PointOnLine, Symmetric --
works against them for free.
Part of this refactors working code, so Project's behaviour identity is the
invariant; the existing [CadDocument][project] cases guard it and a new case
asserts a Project feature still emits construction == false.
project_edges_into_sketch returns the number of entities appended, or -1 on a bad
reference rather than throwing.
VERIFICATION LIMIT, as with the previous four commits: this fork's kernel suite
still cannot run (find_package(assimp) at configure time, snaporca-w80c). Shared
sources are byte-identical to snaporca's, where kernel is 7700 assertions / 270
cases and ALL LADDERS HELD 7/7.
Port of snaporca 2d36d28770. Parity OK: 17 files identical, 8 diverging at their
expected counts.
infer_axis_constraint returned only Horizontal or Vertical. On the
CAD-1000-hours corpus the top two transitions are sketch_dim -> sketch_draw
(5896) and back (5756): the signature of geometry that does not self-constrain
as it is drawn.
Every rule requires the relation to be ALREADY TRUE within tolerance, so nothing
the user drew is moved; parallel/perpendicular and tangent additionally require a
shared endpoint.
TWO LIMITS THE CORPUS RUNG FORCED, neither visible to the unit tests:
1. At most ONE constraint per rule per new entity, not one per PAIR, and no
one-at-a-time fallback for the relations batch. EqualRadius has no locality
restriction, so 200 equal holes produced ~20000 candidates; the rejected batch
then cost a solve per constraint and pinned the app at 95% of a core with the
MCP socket unresponsive.
2. Relations only for gesture-sized batches. "A scripted add is not a drawn
gesture" is already this file's rule at its bulk call site (snaporca-8xg1), and
EqualRadius also couples geometrically distant entities, merging independent
connected components and defeating the partitioning that makes large sketches
solvable (snaporca-yww4). With the cap alone geometry stayed correct (32/32
sheets clean) but seven of the largest timed out, including MPD681 -- the sheet
that call site's own comment names.
Also fixes the tolerance leak behind 2: the bulk path asks for exact inference
with ang_tol_rad = 0 but len_tol_frac kept its 0.01 default.
ALSO independent of this feature: run-kernel-tests.sh defaulted to
TAGS=[CadDocument] while four CAD test files carry their own tags and nothing
selected them (2624 assertions / 206 cases reported, 7648 / 264 actual). All 58
dark cases were passing; the coverage was never exercised.
VERIFICATION LIMIT, as with the previous three commits: this fork's kernel suite
still cannot run (find_package(assimp) at configure time, snaporca-w80c). Shared
sources are byte-identical to snaporca's, where kernel is 7651 assertions / 265
cases and ALL LADDERS HELD 7/7.
Port of snaporca 5b1294de59. Parity OK: 17 files identical, 8 diverging at their
expected counts.
Every industrial sketcher gives you the origin and the axes as references. Here
the origin was only a SNAP target and the axes did not exist, so Symmetric needed
a third picked ENTITY as its mirror axis: symmetry about the sketch's vertical
axis first required drawing a construction line.
Every constraint reference resolves through four lambdas in the solver
(valid/ptOf/primOf/coordOf), so teaching those about three negative sentinel
indices makes the origin and both axes available to EVERY constraint type at
once. No new SketchEntity type, no serialization change; -1 still means "unset".
The references live in G_FIXED and add no degrees of freedom, which a test
asserts via the reported DoF.
SymmetricAboutY / SymmetricAboutX are two buttons that need no third pick and no
construction line. Making the axes clickable in the viewport is deliberately left
out: that is canvas hit-testing work with its own risks.
Recorded in the tests because it will catch the next person: sys.dragged[] is
populated only during a drag, so a plain sketch_solve of an UNDER-constrained
system may move any free parameter -- solvespace runs Newton, it does not
minimise movement. PointOnLine onto an axis is one equation in two unknowns and
the point legitimately slides along it. Those tests pin the free direction
instead of asserting the other coordinate is untouched.
VERIFICATION LIMIT, as with the previous two commits: this fork's kernel suite
still cannot run (find_package(assimp) fails at configure, snaporca-w80c). The
shared sources are byte-identical to snaporca's, where kernel is 2624 assertions
/ 206 cases and ALL LADDERS HELD across all seven rungs.
Port of snaporca 9fa304c77a. Parity OK: 17 files identical, 8 diverging at their
expected counts.
The everyday dimension in SolidWorks and Onshape, and this kernel had no form of
it. Distance constrains the straight-line gap; LockX/LockY pin one point's
ABSOLUTE coordinate. Neither relates two points along an axis.
DistanceX/DistanceY emit SLVS_C_PROJ_PT_DISTANCE against two unit direction
lines built in the solver's G_FIXED group, so they add no degrees of freedom.
THE DIRECTION IS SIGNED, and getting it backwards is silent. libslvs defines a
LINE_SEGMENT's direction as point[0] - point[1] (entity.cpp) and
PROJ_PT_DISTANCE constrains (pB - pA).dot(dir) (constrainteq.cpp:234), so the
reference lines are built head-first to mean +X and +Y.
The same signedness was a real defect in the GUI: the inline editor was
pre-filled with |delta|, so when the closest endpoint pair ran right-to-left,
opening the dimension and accepting the number shown would flip the point to the
other side of its anchor. Opening a dimension and accepting its own value must
be a no-op. The refs are now ordered so the shown value is positive.
On the CAD-1000-hours corpus, dimensioning and constraining is 31.9% of all
observed CAD time -- the largest single class, 7.6x feature operations. This is
the item in the constraint epic that lands most directly on it.
VERIFICATION LIMIT, as with the previous commit: this fork's kernel suite still
cannot run (find_package(assimp) fails at configure time, snaporca-w80c). The
shared sources are byte-identical to snaporca's, where kernel is 2603 assertions
/ 200 cases and ALL LADDERS HELD across all seven rungs.
Port of snaporca 9ec6405e2d. Parity OK: 17 files identical, 8 diverging at their
expected counts (DesignPanel.cpp 32, test_slvs_constraints.cpp 3).
Measured on the CAD-1000-hours corpus: 51.5% of observed CAD time is 2D sketch
work, and dimensioning/constraining alone is 31.9% -- the largest single class.
Two constraints every industrial sketcher has were missing here.
EqualRadius fixes a dead end rather than adding a feature. Picking two circles
and pressing Equal emitted EqualLength, which maps to SLVS_C_EQUAL_LENGTH_LINES
and constrains nothing on a curve: a silent no-op with no error. Equal is now
one button with two meanings, as in Onshape and SolidWorks.
Collinear emits PARALLEL plus PT_LINE_DISTANCE=0 rather than PT_ON_LINE, whose
internal valP param this libslvs port leaves at 0, drifting an already-collinear
pair.
Both types are appended at the END of SketchConstraintType: cereal serializes it
positionally, so inserting elsewhere reinterprets every saved recipe.
VERIFICATION LIMIT, stated rather than implied: this fork's kernel suite could
NOT be run. scripts/CAD/run-kernel-tests.sh fails at CMake configure time on
find_package(assimp), before any source compiles -- a pre-existing deps gap
(snaporca-w80c), not this change. The shared sources are byte-identical to
snaporca's, where the full gate passed: kernel 2588/195 and ALL LADDERS HELD
across all seven rungs.
Also fixes two defects in this fork's scripts/CAD/run-all-checks.sh:
- `cd $(dirname $0)/..` landed in scripts/ instead of the repo root, so every
rung looked for itself under scripts/scripts/. Broken since the script moved
into scripts/CAD/; the three sibling scripts were fixed then and this was
missed, so the gate has not run since.
- C defaulted to snaporca-gui, the OTHER fork's rig container, so this fork's
gate would drive snaporca's app and report green about the wrong binary.
run-kernel-tests.sh:31 documents the identical defect being fixed once
already for the build volume; this is the third instance.
It was already public and used elsewhere, so the new get_background_process()
and its duplicate forward declaration were redundant; Plater.hpp is now untouched.
exussum12 on PR #15238: "Esc hardly ever works". He was right, and the word
that matters is "hardly" — it works while you are mid-entity and stops working
the moment you finish one.
The branch asked whether a DRAW TOOL was armed, not whether a SKETCH SESSION
was open:
if (key == WXK_ESCAPE && m_viewport && m_viewport->is_sketching())
is_sketching() is DesignSketchTool::is_active(), true only between arming a
tool and finishing with it. Commit an entity and you are left at
ui_mode=Sketch with no tool armed, and from there Esc did nothing at all, no
matter how many times you pressed it — the Cancel button was the only way out.
Reproduced on the headless rig with SNAPORCA_KEYTRACE, which is what settled
it rather than reading: the key ARRIVES and focus is fine,
[KEYTRACE] key=27 ui_mode=1 is_sketching=0 in_text=0 inline_busy=0 focus=w
so this was never the focus problem it looks like from the outside.
Gate on the session instead. request_exit() is already layered — abort the
in-progress entity, else drop the tool to Select, else exit to Feature — so
widening the gate adds no new behaviour, it just lets the ladder be reached
from its own last rung. is_sketching() stays in the condition as an OR, so
nothing about the armed-tool path changes.
This is the same mistake snaporca-0ud fixed thirty lines above in the same
handler, where gating the sketch key MAP on is_sketching() made all 17 keys
read as dead. The comment there now has a sibling.
VERIFIED ON THE RIG, before and after, same sequence (Design > XY > Shift+S >
click canvas): before, two Escapes left the toolbar on SKETCH with zero pixels
changed; after, the toolbar reads FEATURES — one Esc drops the armed tool to
Select, the second exits the session.
Kernel suite green, 2568 assertions in 191 test cases. libslic3r_gui builds
clean on both forks. Parity 17 identical / 8 diverging as expected.
The recipe's 3MF entry moved from Metadata/SnapOrca_cad.bin to
Metadata/orca_cad.bin, so projects saved by earlier builds opened with an
empty Design tab. Both 3MF backends now read either name and write only the
new one; the recipe version advances to 6 to mark the move.
The tab is now gated on a new enable_cad_feature app-config key, off by
default, exposed in Preferences under General > Features. When off, the
page is never created, so the Design-only camera option is hidden too.
Takes effect on restart, like the other feature toggles.
Requested by SoftFever on PR #15238: ten of these had accumulated loose in
scripts/ next to ~20 unrelated upstream ones, with names that only meant
something to whoever wrote them. They now sit in scripts/CAD/, mirroring the
src/libslic3r/CAD/ and src/slic3r/GUI/CAD/ split, and the verb in the name is
the role: build- produces a binary, start- brings something up, run- runs a
suite, check- asserts one thing against a live app.
kernel-test.sh -> CAD/run-kernel-tests.sh
ladder-all.sh -> CAD/run-all-checks.sh
sketch-ladder.py -> CAD/check-sketch-engine.py
ladder-corpus.py -> CAD/check-sketch-engine-corpus.py
gui-ladder.py -> CAD/check-gui-sketching.py
offer-ladder.py -> CAD/check-gui-context-menu.py
mcp-sketch-smoke.py -> CAD/check-mcp-sketch.py
rig-build.sh -> CAD/build-gui.sh
docker-iter-build.sh -> CAD/build-gui-incremental.sh
gui-session.sh -> CAD/start-headless-gui.sh
"Ladder" was the worst of them: it named the shape of the test (rungs of
increasing difficulty) rather than what the test proves, so nothing in the
directory listing told you which one needed a GPU and which was pure kernel.
Every reference rewritten -- the docs, the cross-calls between the scripts,
Dockerfile.deps, and the container-side /OrcaSlicer/scripts paths. The three
shell scripts resolve REPO relative to themselves and now sit one level
deeper, so that walk went from /.. to /../.. . The copies these push into a
container's /tmp were renamed to match, or the container would have kept the
old names alive.
Two runtime paths deliberately NOT renamed. /tmp/orca-rig-build.lock is a
cross-fork contract -- both forks take the same lock so two concurrent builds
serialise instead of OOMing the box, and renaming it on one side silently
removes that guard. /tmp/gui-session.log is a runtime artefact, not a script.
Added scripts/CAD/README.md: what each script proves, what it needs, and the
two constraints that have each cost a session (never build inside the GUI
container; a window manager is required or synthetic keys are ignored).
On CI, which was the other half of the request: the kernel suite is already
there and always has been. The cases are registered in
tests/libslic3r/CMakeLists.txt under if (SLIC3R_CAD), which defaults ON and no
workflow turns off, so they build into libslic3r_tests and run under ctest on
every platform via unit_tests.yml -- like any other unit test, needing no new
job. They have simply never been seen to run, because the workflows on this PR
are still awaiting maintainer approval. run-kernel-tests.sh is the local loop
over the same cases, and it is the only script here CI could run: the other
six need an OpenGL canvas and synthetic input.
Verified: scripts/CAD/run-kernel-tests.sh from its new location, all tests
passed, 2562 assertions in 190 test cases.
Building the Design tab on first use (1750c52211) leaves m_design_panel null
until a human selects the tab. McpControl::handle_on_main refused every verb
while it was null, so start_mcp_control_if_enabled() opened the socket and
then answered "Design panel not ready" to everything — for the whole session
if nobody clicked. That is precisely the headless case the socket exists for:
the click-test rig drives this app over it with no window manager and no user.
MainFrame::ensure_design_panel() now builds the panel on demand and returns
it; the tab activation and the MCP dispatcher both go through it, so there is
one construction site rather than two. Safe to build wx controls there: that
handler is dispatched on the main thread by CallAfter, as its own comment
says.
The startup saving is untouched — a launch that never opens the tab and never
speaks MCP still builds nothing.
Verified: libslic3r_gui builds clean 745/745; fork parity green with snaporca,
which carries the same fix (McpControl.cpp is a byte-identical shared source).
SLIC3R_CAD=OFF now builds without SolveSpace or OCCT's ModelingAlgorithms
module, leaving the dependency set identical to upstream's, and the Design
tab's mate connector preference no longer appears in builds without the tab.
When the deps prefix and the project disagree about the option, the configure
fails naming the cause, rather than failing at link time or at first launch
on Windows. The Windows packaging step stages exactly the toolkits libslic3r
links.
The Camera is Plater-owned and shared by every canvas, and Design sits
outside the panel switch that saves and restores it for Prepare,
Preview and Assemble — so orbiting in Design moved what the editor
tabs showed, and Design lost its own view on every switch. Trade the
live camera for a parked one on the way in and back on the way out, so
Design keeps its view the way Assemble already does.
New Project and Open Project went through Plater::priv::reset, which
drops ModelObjects but not the Model-level recipe, and never touched
the Design panel's document at all — so the previous design stayed
loaded and its next edit wrote itself into the new project.
A design that has not been committed to the plate has no ModelObjects,
so the project also read as clean: no autosave, and no unsaved-changes
prompt before the reset threw the design away.
DesignPanel's constructor creates several hundred controls and its own GL
canvas, which every launch paid for whether or not the user ever opened the
tab. The notebook page is now an empty placeholder and the panel is built
into it the first time the tab is selected, so m_design_panel stays null
until then.
This also lifts the ordering constraint that had pushed the plater's
background colour and Hide() below the panel construction; they go back
where they were.
append_tcr2 scanned the tower's G-code for a "preheat T<n>" comment and rewrote
its S value to the interface temperature. Nothing it could match was ever there.
That comment has exactly one producer, GCodeProcessor's backtrace injector, and
that runs inside run_post_process() -- a pass over the finished, exported file.
append_tcr2 runs while the file is still being generated, so the text it looked
for did not exist yet and could not.
The loop therefore walked every line, matched none, and swapped the string for
an identical copy. Delete it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The badge is meant to predict whether slicing will be refused, and it delegates
to the same helper for that reason. It was feeding that helper a different
filament list. get_extruders(true) resolves a mixed slot into its physical
components -- right for AMS mapping, which has to know what is actually loaded
-- while Print::validate counts the slot itself.
So a plate holding one two-component blend reads as two filaments to the badge
and one to validate. The badge sees two, decides the plate is fine, and stays
silent; the slice is then refused. It also runs the other way: a plate the user
sees as a single colour draws a multi-material warning, because the expansion
made it look like two.
Give get_extruders an expand_mixed flag, defaulted so every existing caller
keeps the resolved list, and have the badge ask for the authored one.
The badge was also only mirroring validate's multi-color rule, not its first
one -- a mixed filament is unsupported in a parallel mode outright. Without it
the badge stays quiet on exactly the plate validate refuses first.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The rule's comment claimed the plate "is not printable as configured: the
primary tool executes the toolpaths while the flow and temperatures were
computed for a filament it cannot load". That is not what the emitter does. It
never uses the declared primary -- it re-derives an effective one from the
filament actually in use -- so a plate whose only filament sits on a Span tool
sharing the primary's gantry produces coherent G-code and would print.
The refusal is still correct, but it rests on intent rather than physics: a
parallel mode exists to run carriages in parallel, and a single-colour plate
riding one span lane is not that. Left as a physical-impossibility claim, the
rule reads as a false positive to anyone who checks it against the emitter --
a review already flagged it as one -- and the obvious "fix" is to relax it.
Say which it is, and keep the genuinely-broken case distinct: a filament routed
to a head outside the mode's active tools still yields a stuck-hot nozzle, and
that one is not a matter of taste.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A mixed filament is unsupported in a parallel mode outright, but the rule saying
so ran third. A plate carrying a blend plus any second filament tripped the
multi-color rule's used > 1 gate first and was told its active tools all sit on
one gantry -- a diagnosis of a multi-color print the user never configured,
whose remedy is to go rework the mode's tool roster. The blend was never
mentioned. Move the check ahead of both rules below it; being unsupported
regardless of routing or topology, it dominates them.
Nothing is masked that leads anywhere else: every branch of
imex_multicolor_block_reason is itself confined to non-primary modes, so the
mixed message's remedy -- switch this plate to Primary -- silences those too.
Say "Mixed filaments", not "Blended". Every other string in the app calls these
mixed, including the button that creates one and the sibling refusal for the
wipe tower filament, so the user had no way to connect the message to the
feature it names.
Three comments in the block were wrong, and two of them were newly wrong. The
routing rule's bounds-check note still said "Blended slots are out of range by
construction, but they never reach here -- the rule above returns first": the
rule above is now the multi-color one, which does not return first for a single
mixed filament, and out-of-range is not guaranteed at all. Mixed slots are kept
at the tail of the filament arrays by convention, not by enforcement --
PresetBundle::set_num_filaments grows filament_is_mixed with resize(), so
raising a printer's extruder count with a blend present lands physical slots
after the mixed one. The scan is position-agnostic and stays correct; only the
stated reason was wrong.
The same discovery makes the empty-routed_list guard live rather than the dead
code it was described as. Print::apply() normalises physical_extruder_map before
validate() runs, so an unauthored map is never the cause -- but a printer with
more filaments than logical extruders leaves the tail slots outside the map, and
raising the extruder count does exactly that.
The new test validates the plate twice. The first pass, with no blend, asserts
the multi-color rule is armed at all; without it the second proves nothing,
because the rule only fires here thanks to a degenerate fixture mode whose two
tools share a gantry. Give that mode a Span tool and the whole test would pass
under either ordering while appearing to guard it. It also pins err.object,
which the mixed path sets and the multi-color path leaves null -- a discriminator
that survives the next wording change. Verified by reverting the order: the test
fails on both the message and the object.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
M104/M109 address a heater, but every caller of the instance
GCodeWriter::set_temperature overload addresses filaments by logical id, so on a
printer whose physical_extruder_map is not the identity the emitted T named the
wrong head -- or, where the logical id exceeds the head count, no head at all.
With a map of 0,0,0,0,1,2,3 a toolchange to filament 5 emitted "M109 S265 T4"
and "M104 S190 T4 ;cooldown" while the head it meant was T1.
Upstream already treats these commands as physical: the preheat it injects in
GCodeProcessor maps through the same map before emitting, and BBS's own wipe
tower does likewise. Emitting logical is the half that never got the memo.
That mismatch also disabled the cooldown suppression beside the preheat, which
compares the line's T against pem[tool_number] and so never matched a logical
one -- 171 cooldowns survived in a two-head print where none should have. Worse,
it could match the wrong line: a cooldown for filament 1 emitted T1, and a
toolchange to filament 5 gives pem[4] == 1, so a legitimate cooldown for head 0
was deleted because the incoming head happened to be numbered 1.
Translate once, in the instance overload every logical-space caller passes
through. The static overload is already physical-in and is left alone.
Gated on is_imex. physical_extruder_map carries two readings in this tree: the
BBS paths index it by extruder id, the IMEX paths by filament id, and the two
coincide only when the filament and nozzle counts match. Mapping unconditionally
would impose the IMEX reading on profiles that mean the other one --
fdm_bbl_3dp_002_common ships a non-identity [1,0], spared today only because
single_extruder_multi_material suppresses the T qualifier entirely.
The wipe tower's interface-temperature pass has to move with it. It strips the
M109 that post_toolchange emits by searching for that filament's tool index, so
it now searches for the mapped one; left alone it would have stopped matching,
and the surviving blocking M109 would have silently defeated the interface
temperature. Its sibling pass reads WipeTower2 output, which emits no T at all,
and is deliberately unchanged.
The bare T<n> toolchange stays logical -- it selects an AFC lane, not a heater.
Test slices two objects across a head boundary, the only case that reaches this
emission: the same-physical short-circuit in set_extruder suppresses the
cooldown entirely for lane swaps within one head. It scans every M104/M109
rather than matching fixed strings, so it catches any unmapped emission and not
just the two sites changed here. With the mapping neutered it reports 101
offending lines; with it in place, none.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The routing error ran to roughly 450 characters and explained the mechanism
before it got to the remedy. It also offered to "edit the mode in Printer
Settings so its Primary tool is one of %3%", which on a plate whose filaments
resolve to no head at all rendered as "one of no configured extruder". Cut it
to the mode, the tool it prints with, where the plate's filaments actually are,
and the two things the user can do about it.
The second msgid that named candidate modes went with it. It could only suggest
a mode whose primary is among the routed heads, and every mode on the printers
this fires for declares 0:P, so it had nothing to offer.
Blended filaments now return before that check rather than falling through it.
A blend is mixed at the nozzle by its component toolheads, and a parallel mode
is already using those toolheads to print copies or mirrors, so the two cannot
run at once regardless of where the components route -- including when a
component sits on the declared primary. Reaching the routing rule would also
have described them wrongly: mixed slots sit past the end of
physical_extruder_map, so they resolve to no head and read as merely unrouted.
Keeps the empty-list guard the shortening first dropped. validate() reads the
raw physical_extruder_map, whose registered default is a single entry, so a
profile that declares IMEX modes without authoring a map leaves every slot past
the first outside it -- and the sentence ended in a dangling "on .".
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Brings the upstream color-mixing feature and its follow-ups onto the branch so the
IMEX placement and primary-routing checks are built and tested against them for the
first time.
Merged clean, no conflicts. Not yet exercised together: a mixed filament is a virtual
slot no nozzle carries, while physical_extruder_map routes logical slots to physical
heads, so the IMEX pem lookups have no defined answer for one. Print::extruders()
lists mixed slots under their own id while tool_ordering.all_extruders() lists them
post-expansion, and the IMEX code reads both.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Plater::validate_current_plate() runs the same background_process.validate() as
update_background_process(), and on success clears update_apply_result_invalid(false)
and closes the ValidateError notification -- but it never consulted
imex_placement_violation(). Any event reaching it wiped a live IMEX placement error
and re-enabled the Slice button on a plate the slicer still refused; pressing Slice
then hit the check in reslice() and returned early, so the job simply never started.
Reproduce by clicking the bed with the prime tower overlapping a reserved area.
Plater::select_plate_by_hover_id() -> select_plate() calls validate_current_plate()
unconditionally, even when the clicked plate is already current, and deselects as a
side effect -- which makes it look as though deselecting the tower cleared the error.
Escape and clicks off the bed go through selection_changed(), which only renders and
clears nothing.
Extract the violation-to-message mapping into one helper and call it from both paths.
model_fits is set alongside err.string in validate_current_plate, mirroring the
missing-plugin block below it: the slice is already gated by m_apply_invalid, but
leaving m_ready_for_slice true would trap a future consumer that reads it alone.
These are the only two sites that matter. NotificationType::ValidateError has exactly
three references in the tree and update_apply_result_invalid exactly four; the other
slice-ready writers can only touch m_ready_for_slice, never m_apply_invalid, so they
cannot re-enable Slice on their own.
Three adjacent gaps are left alone, all pre-existing: "Slice all" is hard-coded
always-enabled regardless of plate state; a slice-all batch halts silently at a
violating plate because reslice() returns above the line that queues the advance; and
object_list_changed() computes its own can_slice from geometry, harmless only because
the result is ANDed with PartPlate::can_slice().
This is a hole in the shipped tower-zone check rather than a regression from rotating
the tower hull -- it was simply invisible until a tower could be placed in violation.
No automated gate: Plater is GUI-only, the helper is file-local and unlinkable, and
imex_placement_violation() needs a live wxApp and preset bundle. "Both call sites
consult it" is a call-graph property no unit test can express. The grep for a single
imex_placement_violation reference is a future regression tripwire, not evidence this
refactor happened -- it already returned 1 beforehand.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Brings in 54 upstream commits, the bulk of them the BambuStudio-ported color
mixing / mixed filament subsystem (#15347) plus its follow-ups, along with the
Assimp-backed colored OBJ import, warning-policy build changes, and assorted
profile and localization updates.
Two conflicts, both "each side added at the same point", resolved by keeping
both:
- Print::validate() -- our IMEX multi-color block and upstream's new gradient
mixed filament warning were inserted at the same spot after the empty
extruders check. They test unrelated conditions, so both are kept, each with
its own closing brace.
- tests/libslic3r/test_3mf.cpp -- our three IMEX per-plate round-trip scenarios
and upstream's mixed-filament round-trip scenario both append to the end of
the file, and each side added one include. All four scenarios and both
includes are kept.
Everything else merged cleanly, including GCode.cpp, ToolOrdering.cpp,
PartPlate.cpp and PrintConfig.cpp. Upstream left the is_extruder_used block
untouched, so the IMEX supplement still applies, and estimate_wipe_tower_polygon
is unchanged, so the prime tower hull work is unaffected.
Not addressed here, and worth its own change: a mixed filament is a virtual slot
that no nozzle carries, while physical_extruder_map routes logical slots to
physical heads. Print::extruders() lists mixed slots under their own id whereas
tool_ordering.all_extruders() lists them post-expansion, so the IMEX pem lookups
have no defined answer for a mixed slot. Upstream's own guards reject a mixed
filament where a physical slot is required; IMEX likely wants the same.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The IMEX primary tool prints the sliced paths directly, so it can only load a
filament that physical_extruder_map routes to it. The ghost filament picker
enforces that for the secondary tools -- it offers only lanes whose pem entry
equals that head -- but the primary's filament comes from the ordinary object
filament selector, which has no IMEX awareness. Nothing detected the mismatch:
collect_imex_warnings() computes the same condition and discards it into a
display fallback, and the multi-color rule never examines it.
Block it in Print::validate() via the existing imex_primary_tool_for_mode and
imex_primary_logical_from_objects helpers. The message names the declared
primary, the heads the plate's filaments actually live on, and any configured
modes whose primary would work, and carries the object so the notification can
offer a jump to it.
Blocks rather than warns, matching the multi-color rule: the plate is not
printable as configured, and where the routed head is also absent from the
mode's active tools the 1st->2nd layer temperature branch skips it too, leaving
that head at its initial-layer temperature for the whole job.
The multi-color check now runs first. Its constraints -- an MMU manifold sharing
one head, a single-gantry mode -- cannot be fixed by switching mode, so the more
specific error should win rather than be masked by routing advice that leads
straight back to it. The extruders().size() > 1 gate moved onto that call, since
the routing check must also see single-filament plates, which is its common case.
The copy-mode guard-rail test printed on a filament routed off the primary, so
it asserted a plate this rule now refuses; retargeted to a well-formed plate.
Its replacement pins the object's own extruder, because ModelVolume reports its
extruder_id and would otherwise put a primary-routed slot on the plate.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Port of snaporca 1bb9825db0. Four defects from an independent 20-agent audit,
each verified in the code first; two further findings from the same report were
verified OUT and are not in this commit.
remove_feature()/move_feature() remapped Extrude::sketch_ref and a Mate's two
connectors and nothing else, leaving seven of the nine index-bearing fields —
sweep_path_ref, loft_profile_refs[], pattern_curve_sketch, rib_sketch_ref, and
sketch_ref on Revolve, Sweep, Rib and the Surface* family — pointing at whatever
slid into the slot. Quiet by construction: the shifted index still names a real
feature, recompute() succeeds, the solid is built from the wrong profile. The
comment above the loop already required "EVERY field holding a feature index"; the
code under it handled two, because a type switch is only correct on the day it is
written. for_each_feature_ref() visits the FIELDS instead, so a feature type added
later is covered the moment it reuses one. plane_base and axis_plane_a/b are
excluded on purpose and documented at the helper — they encode an ordinal into the
datum-plane list, not an index into features[], and are filed separately. The
delete cascade got the same field-based treatment.
The regression test was run against the pre-fix code to prove it bites: all three
sections fail there, and move_feature returns TRUE while leaving sketch_ref == 1
where it must be 0 — success with the wrong answer, which is what makes this class
expensive.
apply_constraint, commit_entity_constraints and delete_constraint mutated the
recipe with no checkpoint() and no sync_recipe_to_model(), alone among seventeen
mutation sites in that file: Ctrl+Z reached past the constraint edit and discarded
unrelated work, and saving persisted the pre-constraint blob. A rejected constraint
now calls abandon_checkpoint() rather than leaving an undo step that does nothing.
MCP: params["generation"].get<uint64_t>() sat outside the try inside a bare
CallAfter lambda, so one malformed string terminated the process through the wx
event loop; it is type-checked now and the lambda lets nothing escape. The socket
bound with no mode of its own in a world-writable directory — umask around bind()
plus chmod, and it refuses to listen rather than listen wide. The reply write is no
longer a bare write(), which could SIGPIPE the app when a client hung up.
Kernel suite on this fork: 190 cases / 2562 assertions, green. GUI target compiles.
The full ladder gate ran on snaporca (ALL LADDERS HELD — gestures 98/98, offer
108/108, corpus and corpus-scale green) and fork-check parity holds at 17 identical
/ 8 diverging as expected, which is what makes that gate transferable here.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
User report 2026-08-23, and it is right: "you have renamed the feature extrusion,
not the body. i clicked rename on the body feature tree and the feature extrude
changed name. this means that you consider the extrusion = the body. this is very
far from truth as a body can contain several extrusions."
That is exactly what the previous commit did. It resolved a selected body to
CadBody::source_feature and renamed THAT feature, on the reasoning that a body has
no name of its own. The reasoning described an implementation detail — CadBody::
name is derived and restamped on every recompute — and mistook it for the user's
model. An Extrude, a Cut and a Fillet all land on the same body: the maker is one
operation in its history, and renaming it renames the wrong object.
A body now has a name of its own. CadBody::user_name, set only by a rename, is:
- carried across recompute() by body index, next to the per-body colour override
and under the same index contract the GUI already relies on for visibility and
Move — without which a name would survive exactly until the next feature;
- written into the recipe, because bodies are recomputed and never serialised, so
a name has nowhere else to live and would otherwise vanish on reopen;
- shown on the Bodies row ahead of the derived maker name, as "Body N — name",
with the number still leading because every status line, the interference
report and the mate errors identify a body that way.
The recipe block is APPENDED after the variables block rather than given a version
bump. A build that predates it reads features and variables, returns, and never
looks at the trailing bytes — so yesterday's projects open here and today's
projects still open there. A bump would have cost every project written today its
readability by the previous build, for one optional field.
Renaming a FEATURE is unchanged. The feature tree renames features; the Bodies
list renames bodies; neither reaches into the other.
WHAT THIS COST, and why it is written down. Getting here took two wrong turns
inside one fix, both mine:
1. UnselectAll() -> Unselect(). Both trees are wxTR_SINGLE, where UnselectAll()
— the MULTI-selection call — does nothing. That was the one-word reason the
rename had been vetoed everywhere (BEGIN_LABEL_EDIT refuses while
tree_body_selection() >= 0). Fixing it turned a pair of harmless no-ops into
a real loop: the two lists clear each other so "the target" is unambiguous,
so clicking a body row ran apply_body_row -> m_tree->Unselect() -> the feature
tree's SEL_CHANGED -> m_parts->Unselect(), which cleared the row just clicked.
The handler now clears the other list only when it actually holds a selection.
2. Trusting a screenshot taken after a polluted run. A leftover Rib card had
shifted the whole panel, so a click measured against it landed nowhere near
the row. Relaunch, then measure.
VERIFIED, on the rig and in the kernel:
body renamed ('Extrude', user_name=False) -> ('Bracket', user_name=True)
features untouched ['Sketch', 'Extrude'] before and after
survives a recompute add a Hole to the same body: features become
['Sketch', 'Extrude', 'Hole'], body stays 'Bracket'
survives the recipe serialize -> deserialize -> 'Bracket'
New kernel test "a body carries its own name, through recompute and the recipe"
pins all three properties; the suite is 189 cases / 2547 assertions.
Gate green: ALL LADDERS HELD — offer table matches the atlas, kernel suite,
engine rungs 1-8, 977-sheet corpus + the heaviest sheets, gesture ladder 98/98,
offer ladder 108/108.
User report 2026-08-23: "clicking on a body row in feature tree, I cannot find
rename on right click", then "still i cannot rename body1 in custom name".
TWO SEPARATE CAUSES, one in data and one in a single method call.
THE MISSING ROW WAS DATA. The `rename` verb's accepts list in the tool atlas was
["sk_loop"] alone, so the offer built for a selected BODY carried no Rename row.
Right-clicking a body row already opens the offer — that is the designed gesture,
bound on m_parts as wxEVT_TREE_ITEM_MENU — so the menu the user was looking at
was the right menu, and it was simply missing the verb. accepts is now
["sk_loop", "body_solid"], the generated table regenerated with it (the accept
mask moves 0x00004000 -> 0x00004080), and the offer trace confirms the row:
"[OFFER] row=7 Modify > rename".
THE RENAME ITSELF WAS BLOCKED BY UnselectAll(). Both trees are wxTR_SINGLE, and
wxTreeCtrl::UnselectAll() is the MULTI-selection call: on a single-selection tree
it leaves the row selected. So every path that tried to open the label editor
while a body row was selected hit the BEGIN_LABEL_EDIT guard — which vetoes while
tree_body_selection() >= 0, the rule that stops a body taking a name it cannot
keep across a recompute — and the editor never opened. Unselect() is the call
that works, and it fixes every route at once.
That took five attempts, four of them wrong, and the reason they were wrong is
worth more than the fix: each one addressed a plausible cause that the evidence
did not actually support — the popup's nested event loop, keyboard focus,
deferring with CallAfter, dispatching through a different verb. What settled it
was a DISCRIMINATOR rather than another fix: pressing F2 on a selected body row
takes the same handler with no menu and no nested loop. F2 failed identically,
which ruled out every menu-shaped theory in one measurement and left only the
state the veto reads.
A body still has no name of its own — it is recomputed from the recipe on every
change and CadBody::name is derived from the feature that builds it — so the verb
resolves the body to CadBody::source_feature and renames THAT, saying so on the
status line: "A body takes its name from the feature that makes it — renaming
'Extrude'". The Bodies row now reads "Body 1 — Extrude" so the rename is visible
where it was made; the positional "Body N" leads, because every status message,
the interference report and the mate errors identify bodies that way. Confirmed
as the wanted format by the user.
Also here, from the same report: the Bodies card keeps its own action row (Move,
Show / hide, Delete, Colour), and the competing context menu an earlier pass had
added to body rows is REMOVED — right-clicking a body belongs to the offer, and
two menus on one gesture is how the offer ended up being blamed for a veto.
Verified on the rig, both routes, with a body selected:
F2 on the body row -> ['Sketch', 'Extrude'] became ['Sketch', 'Base block']
the offer's rename verb -> {'applies': True, 'dispatched': True,
'selection_kind': 7, 'ok': True} and the same rename
Gate green: ALL LADDERS HELD — offer table matches the atlas, kernel 188 cases /
2532 assertions, engine rungs 1-8, 977-sheet corpus + the heaviest sheets,
gesture ladder 98/98, offer ladder 108/108.
RIG DISCIPLINE, repeated twice in one session and now written down: ladder-all.sh
does not relaunch the app, so hand-driving the rig immediately before a gate
leaves state its reset_document() cannot clear — both times the first rung drew
nothing and reported "sides []", which reads exactly like a broken rectangle
tool. Relaunch before gating.
Two decisions from the user, 2026-08-23, after the first pass at making rename
reachable.
ONE SURFACE CARRIES THE VOCABULARY, and it is the row's own menu: Rename (F2),
Edit, then Move up, Move down, Show / hide, then Delete. Offered as a choice
between completing the menu or completing the header icons; the menu won because
the element you click answering with what applies to it is this fork's charter,
and because a menu grows without spending an icon nobody recognises. The header
icons stay exactly as they are — a quick bar for the common three — so nothing
that worked yesterday moved.
The menu is grouped rather than listed: what the row IS (name, contents), where
it SITS (order, visibility), and what removes it. Right-click SELECTS what it
points at before opening, so the menu can never act on a row other than the one
under the cursor.
MOVE BODY LEAVES THE FEATURE-TREE HEADER. It never belonged there: that header
sits over the FEATURE tree, and the button had to guess its subject from
whatever happened to be selected — a feature row got answered with "Select a
body to move it", an instruction about a different kind of object in a list that
does not contain one. A body now has its own action row on the Bodies card:
Move, Show / hide, Delete, Colour. The last three are deliberate COPIES of
feature-tree actions, not a reorganisation: a body row is a different subject,
and someone working in the Bodies list should not travel to another card to hide
or recolour what they have just selected. Each handler already resolves the body
row itself (on_toggle_visibility, on_delete_body, on_set_body_color), so the
card gives them a home and decides nothing.
The offer already carried Move for a selected body (verb `transform`, accepts
body_solid), so that half of the requirement was in place and is untouched.
One piece of the old button was function, not clutter: it also scaled imported
Text/SVG artwork, which IS a feature-row action. That moved to the row's own
menu as "Scale artwork", shown only on a row that has imported regions — so the
capability survives the split instead of disappearing with the button.
Verified on the rig, by the gestures themselves: right-click a row -> the six
items in their three groups; choosing Move down turned ['Sketch1', 'Sketch2']
into ['Sketch2', 'Sketch1']; and with a body present the Bodies card shows its
four actions while the feature header no longer shows Move.
Gate green: ALL LADDERS HELD — offer table OK, kernel 188 cases / 2532
assertions, engine rungs 1-8, 977-sheet corpus + the heaviest sheets, gesture
ladder 98/98, offer ladder 108/108.
RIG DISCIPLINE, learned the expensive way in this session: ladder-all.sh does
NOT relaunch the app, so hand-driving the rig immediately before it leaves state
the ladder's reset_document() does not clear — here the first rung drew nothing
at all and reported "sides []", which reads exactly like a regression in the
rectangle tool. Relaunch the app before a gate, and re-run before believing a
failure that appears in rung one.
User report 2026-08-23: "on feature tree, i cannot rename sketch name".
The rename was not broken. Verified on the rig before changing anything: select
the row, press F2, type, Enter — Sketch1 becomes Base, the name reaches
m_doc.features[idx].name and sync_recipe_to_model() persists it. The offer's
btn:rename verb does the same. What was missing was any way to find that out.
Everything a person would try did something else:
- the pencil in the section header is EDIT (on_edit_feature)
- a double-click fires wxEVT_TREE_ITEM_ACTIVATED, which is also Edit
- none of the seven header icons renames
- right-clicking a row did nothing at all
and the comment above the label-edit handlers claimed a "slow double-click"
renames, which does not survive wxGTK: the activation wins and the sketch opens
for editing instead. So the only route was an undocumented function key, and the
report is exactly right from where the user stands.
The row now answers the gesture people actually use on a named row: right-click
gives Rename (F2) / Edit / Delete, with Rename opening the in-place editor on
that row. Right-click SELECTS what it points at first, so the menu can never act
on a different row than the one under the cursor.
And selecting a row now says what the row can do: "Sketch1 selected — F2 or
right-click renames it, double-click edits it". Cheaper than a tooltip nobody
hovers, and it uses the status line that already exists for exactly this.
A note on the surface, since it is a design call: the CANVAS right-click is the
offer, this fork's single adaptive menu, and this is not that. A tree row is a
different surface, and its menu is three items about the row. Routing tree rows
through the offer would mean teaching the offer a selection kind that is not
geometry, which is a larger change and not what this report needed.
Verified on the rig by the gesture it names: right-click the row, choose Rename,
type, Enter -> ['Sketch1'] becomes ['Base profile'] in describe_scene.
Gate green: ALL LADDERS HELD — offer table OK, kernel 188 cases / 2532
assertions, engine rungs 1-8, 977-sheet corpus + the heaviest sheets, gesture
ladder 98/98, offer ladder 108/108.
Two user reports from the same session on the deployed build, 2026-08-23.
FIRST: "if I select a shape (es a circle draw in construction lines) and then I
try to toggle contruction to obtain a full line, does not work". Reproduced: Q
converts the selection and so does the offer's Reference > Construction row —
both run m_keys_sketch['Q'] — but the CHECKBOX, the one control actually
labelled Construction, only ever called set_sketch_construction(), which arms
the mode for the NEXT entity. So the obvious control was the one route that
could not convert existing geometry, and it failed silently while also flipping
the draw mode behind the user's back. It now carries Q's meaning.
Scoped to Select mode, and that scoping is not cosmetic: drawing AUTO-SELECTS
what was just drawn (draw-then-edit), so with a draw tool armed "there is a
selection" does not mean the user picked anything — it means they finished a
line. The first version converted there and turned the box into a trap: arm
construction, draw the axis, click the box to go back to real geometry, and
instead of disarming the mode it converted the axis just drawn. The gesture
ladder's C4 rung does exactly that and reported three construction entities
where it wanted one. In Select mode the intent is unambiguous.
SECOND, and this one destroyed work: "after creation of a circle, a round angled
rectangle and a slot, and mirror of those shapes on a vertical line inside a
outer rectangle, preview is ok but application creates errors: the circle is
mirrored, but rectangle and slot are redrawn as pieces of circles screwing both
the original shapes and the copies." A screenshot came with it, and it showed
more than the words did: the ORIGINALS were wrecked too — the rounded rectangle
was drawn as a four-lobed cloud, each corner fillet having gone the long way
round, and the slot had ballooned into two near-full circles.
Measured on the rig, a slot mirrored about a vertical line:
rails 62.873 / 62.873 -> 2.082 / 62.913
caps r=21.554 sweep=-180.00 -> r=32.214 sweep=-237.66
and all four sources moved, the axis line with them
Cause: confirm_op's Mirror branch bound an Arc copy to its source with a
Symmetric constraint on the CENTRE ALONE. An arc has five degrees of freedom;
pinning two of them leaves the endpoints and the sweep free while the shape's
own coincidences still pull on them, and the solver answers with a different,
internally consistent sketch — which is what a reflex cap and a 2 mm rail are.
A circle came through the same code untouched because a circle HAS no endpoints
to leave free, which is exactly why the failure reads as "circles fine, rounded
rectangles and slots destroyed".
Three parts, and each one is here because the measurement caught the previous
one being half a fix:
1. Arcs are bound by BOTH ENDPOINTS. Endpoints before centre in the ladder:
{p0, p1} is four equations against five DoF and pins the sweep, while
{centre, p0, p1} is six and is refused — the refusal is what silently
degraded the batch to a set that left the sweep free.
2. Every copy is reflected from the PRE-BATCH source, so a batch that disturbs
the sketch cannot hand the next copy already-moved geometry.
3. THE APPLIED RESULT IS THE PREVIEW — checked on the sources AND the copies,
and on violation the whole constraint web is dropped and both halves are
restored to the reflection the preview drew. try_add_constraints rolls back
only when a solve FAILS, and every failure here came from a solve that
succeeded at something else. Guarding only the sources fixed the slot and
left the rounded rectangle's copies at a 13.8 mm rail and a 308 degree cap:
the original was safe and the copy was still wrong, which is half a fix.
The parametric link is kept whenever it provably holds the geometry, and dropped
when it does not. A wrong shape is worse than an unlinked one.
WHY NOTHING CAUGHT THIS: the gesture ladder's mirror rung reflects three
straight LINES. It sat green through the whole defect. C4b now mirrors a slot,
so the reflection has arcs in it, and grades the property the user actually
stated: the copy is the source reflected, the source does not move, and no cap
comes back reflex.
VERIFIED against the user's own scene, rebuilt gesture by gesture on the rig —
outer rectangle, circle, rounded rectangle and slot, a vertical CONSTRUCTION
line as the axis, all 17 entities mirrored in one gesture:
ok the mirror axis is a construction line
ok picked the axis and all 17 entities
ok originals unchanged (moved: [])
ok every copy is the exact reflection (worst 0.000000000)
ok no source arc turned reflex — the 'cloud' failure
ok no copied arc turned reflex (6 arcs checked)
One grader correction worth recording, because it cost a round and would cost
the next one too: a reflection REVERSES ORIENTATION, so a copy legitimately
stores p0/p1 the other way round. Comparing p0 to p0 grades the storage order,
not the geometry, and reported a perfect mirror as an 8.98 mm error. Endpoints
are compared as an unordered pair.
User report, 2026-08-23, after using the freshly deployed build: "selection and
removal of existing elements of the 2d sketch is not intuitive, and the bottom ui
text does not illustrate what the user has to do to properly use the selected
tools. Mirror, for example, does not indicate: first select mirror line then
entities to be selected, and there is no UI indication of what is being selected.
normally, costruction lines are dotted." Four defects, all of them in the 2D
vocabulary this fork's charter puts at the centre, and all four fixed here.
snaporca-1c0c (P1) — the prompt was written ONCE, when the tool was armed.
DesignPanel's select_tool lambda set a sentence and nothing ever revised it, so
every step after the first was unguided: Mirror said "pick axis, then entities"
and then never said which of the two you were on; Escape silently downgraded an
armed tool to Select (request_exit is layered: anchors, then tool, then session)
while the line still named the tool you had left; and nothing ever mentioned that
Del removes a selection. The fix moves the line off the arm event and onto the
tool's LIVE state. DesignSketchTool::emit_step_hint() reports (mode, step, picks)
whenever that triple moves, from render() — the one place every state change in
this tool passes through. Putting it there instead of in the thirty-odd branches
of on_mouse is the whole point: a per-call-site notification is a thing the next
tool forgets to add, and it costs three int comparisons a frame. DesignPanel owns
the words, in ONE table (sketch_step_prompt), whose step numbers are the same ones
render() previews and on_mouse consumes, so the description cannot drift from the
code that reads the clicks. Every tool now names the gesture that ENDS it, because
none of them was discoverable: an empty click applies an edit-op or a transform,
right-click cancels it, Esc goes back to Select.
snaporca-vd6v (P2) — Mirror mirrors its axis pick and its target picks into
m_selection, so both painted white and the picture could not answer "what did I
select as what". The edit-op's first pick — Mirror's axis, Fillet/Chamfer's first
line — now paints violet. Violet and not cyan: cyan means SELECTED in this canvas
and nothing else may wear it, a rule this file already carries in writing.
snaporca-imlq (P2) — construction geometry drew as a solid grey line. Every CAD
dashes it, and grey alone does not read as "reference" against the
under-constrained orange. dash_polyline() chops the polyline before it reaches
draw_quad_strip, with the dash and gap in world units scaled by units-per-pixel,
so a dash keeps its size on screen instead of becoming a solid line when you zoom
out and three dashes when you zoom in.
snaporca-oql1 (P2) — Backspace now deletes as Del does. On every laptop this runs
on, Del is a chord and Backspace is what a hand reaches for. The Select-mode
prompt states the rest (Shift-click adds, double-click takes the loop, Del
removes), and the first step of every armed tool names the Esc route back to
Select, which was the invisible half of "selection is not intuitive".
Also, on the same report: the sketch stroke half-width goes 0.6 -> 0.3 mm. At 1.2
mm wide the orange line swallowed a short segment and hid which of two near
parallel lines the cursor was on. One constant, because all twenty call sites of
draw_quad_strip are sketch strokes.
Retired on the way: the on_sketch_selection_changed status writer. It said "N
selected — Delete removes them" while an edit-op mirrored its picks into the
selection, i.e. in the middle of a Mirror gesture, where Delete does nothing of
the sort. on_sketch_step says the true thing for Select and says nothing false
anywhere else. And the live length/angle readout is now APPENDED to the step
guidance rather than replacing it: it fires on every mouse move, so it used to
erase the instruction for the step in progress one move after the click that
started it.
Two false trails, recorded so the next session does not walk them again:
- DesignPanel.hpp deliberately does not include DesignSketchTool.hpp, so the
panel's handler takes the mode as an int and the .cpp casts it back. The
first attempt put Mode in the header signature and the build said only
"expected ',' or '...' before 'mode'".
- The offer ladder failed six properties against a perfectly good binary
because I had relaunched the rig myself without SNAPORCA_KEYTRACE=1, and its
[OFFER] trace lines ARE its instrument. A ladder with no instrument reports
"None", which reads exactly like a regression in the offer. ladder-all.sh
launches it correctly; a hand relaunch must too.
VERIFIED, not merely compiled. Driven on the headless rig with synthetic mouse
and keyboard, and photographed at each step: "Mirror — first click the LINE to
mirror about (a construction line works) · Esc goes back to Select" ->
"Mirror — axis set · now click the entities to mirror · right-click cancels"
-> "Mirror — axis set · 1 to mirror · click another to add or remove it ·
click empty space to apply", with the axis violet, its target white, and the
construction lines dashed while real geometry stays solid.
Full gate green afterwards (scripts/ladder-all.sh, ALL LADDERS HELD):
offer table vs the atlas OK
kernel suite 188 test cases, 2532 assertions
engine ladder rungs 1-8, ALL RUNGS HELD
corpus rung 977-sheet drawing corpus, every 20th -> 49
sampled, 39 gradeable, 39 clean
corpus scale rung the 6 heaviest sheets, all clean
gesture ladder 93/93 properties, real mouse and keyboard
offer ladder 108/108 properties, through the right-click
menu and the verbs behind it
That harness is the reason a UX change of this size can be made in one pass and
believed: 93 + 108 properties are driven the way a person drives the app, and the
977-sheet corpus keeps the engine underneath them honest against real drawings
rather than against my own arithmetic.
It passed alone and failed inside scripts/ladder-all.sh, twice, on the same property: "right-click
with two picked -> None". The diagnostic it now prints said what a screenshot could not — editing=True
with an empty selection after two clicks that should have picked two entities. Three rig facts, all
about the DRIVER, none of them a product defect:
A LEFT-CLICK ON A LINE'S MIDDLE OPENS ITS LENGTH FIELD. The Select branch tests m_live_quotes before
it picks, with a ~24 px label tolerance, and a line's Length quote sits at its midpoint — so the
click that was meant to select it promoted a dimension and froze the canvas instead. Everything
after it landed on nothing. It bit only when the previous step had left that line selected, because
live quotes are drawn for the SELECTION: hence passing alone and failing in the gate. Lines are now
picked at 0.3 along, clear of the label.
A FIELD THAT HAS NOT OPENED YET READS LIKE ONE THAT NEVER WILL. The queue opens each field from a
CallAfter, so a driver that looks once, sees nothing and moves on gets frozen by the field that
arrives a moment later. keep_as_drawn() now waits for QUIET — two consecutive clear readings — and
draw_line_at drains stragglers before handing back.
A KEY CANNOT CANCEL A SESSION WHOSE CANVAS IS FROZEN. gui-ladder's enter_sketch dismisses the old
sketch with Escapes, which an open field swallows, so the session survived and the four calibration
probes landed in it on top of what was already there: "calibration expected 4 points, got 7". Every
rung now enters through fresh_sketch(), which cancels through the socket first — that cannot be
swallowed.
Measured after the fix, in the sequence that failed: gesture ladder 93/93 then offer ladder 108/108,
back to back on the same app.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
imex_wipe_tower_hull() took the axis-aligned box estimate_wipe_tower_polygon()
returns and compared it to the IMEX collision zones as-is. The real tower is
rotated about its anchor corner before placement -- first_layer_wipe_tower_corners
builds the box in tower-local coordinates, rotates about the local origin, then
translates by wipe_tower_x/y -- so a rotated tower's true footprint fell outside
the hull and the placement check passed on a tower that intrudes into a
carriage's reserved space.
The gap was documented in place and previously harmless, because
wipe_tower_rotation_angle was read from the project config after it had been
moved to the print preset, so the setting did nothing. Upstream repaired that
read ("Fix prime tower rotation angle setting not working"), which makes the
angle reachable and the stale hull wrong.
Rotation is applied to the hull alone. estimate_wipe_tower_polygon() still
returns an unrotated box and still leaves ArrangePolygon::rotation unset, since
the arranger consumes that field separately.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
snaporca-ziam said gen_offer_table.py would silently delete the model-mode "Constrain sketch"
row, because that row lived in the generated header and not in tool_atlas.json. Running it found
more than that: FOUR rows existed only in the header — constrain, rename, and the three typed-
value rows sk_length / sk_radius / sk_angdist — and sk_delete's action had drifted, pointing the
sketch row at btn:delete, the FEATURE delete.
All five are now in the atlas, so the header regenerates byte-identically from it. Verbs may carry
a `note`, emitted as a C++ comment above the row: a rationale written into a generated file is
deleted by the next regeneration, which is how this started.
snaporca-z8rs (P1), found by making that true: after the atlas held all 92 verbs, the regenerated
header differed from the checked-in one by EXACTLY ONE LINE — kOfferVerbCount, 91 against 92.
Every consumer loops i < kOfferVerbCount, so the last row of the table was invisible: never listed
by show_offer_menu, never findable by mcp_run_verb. The verb that fell off the end is sk_angdist,
"Angle / distance…" — the typed-value row for a two-entity selection. On the one selection where
you would ask for the angle between two lines, the row that types it was not in the menu.
It survived because nothing compared the Sk2Ent menu against the table: sk_angdist accepts Sk2Ent
and nothing else, so an off-by-one that dropped the LAST verb was invisible from every other
selection. The vocabulary rung now covers Sk2Ent too, and picking the pair taught it one more
rig fact — shift-clicking a circle at its +X point grabs the RADIUS GRIP, which replaces the
selection with that one entity, so the pair silently collapsed to one and the offer answered
SkLine. Correctly, for the selection that actually existed.
gen_offer_table.py --check proves header == atlas and changes nothing; it is now the first step of
scripts/ladder-all.sh, and the only one that needs no rig.
Offer ladder 107/107, gesture ladder 93/93, both on the rig.
snaporca-ziam snaporca-z8rs
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
The 2D vocabulary is 46 verbs: 22 have a shortcut and the gesture ladder drives them, 24 have
none and nothing had ever exercised those. They are reachable only from the right-click offer, so
a key-driven ladder could not have touched them whatever it did. Four new rungs drive all 24, and
the coverage claim itself is now arithmetic against DesignOffer.hpp (rung O8) rather than a
sentence in a comment that rots when a verb is added.
The assertions are CONSTRUCTION invariants wherever a click cannot be exact — a regular polygon's
sides are equal to 1e-9 and its vertices lie on one circle; a tangent arc's radius at the shared
endpoint is perpendicular to the line to 1e-9 (measured cos 5.97e-17); the three clicks of a
3-point circle all lie on it; a circumscribed pentagon's circumradius is the inscribed one's over
cos(pi/5), 1.236067977 against 1.236067977. Where a value field opens, the typed value is graded
exactly: a moved line travels +25.000000000 in X and 0 in Y, a rotation turns 30.000000000 deg
and leaves the length alone, a scale multiplies it by exactly 3, a linear array's pitch is
[20.0, 20.0, 20.0] and a polar one's spokes are 60 deg apart all the way round.
Three defects found doing it, all fixed here:
snaporca-ua9g (P1) — delete_selected left three things behind. The AUTO-EDIT QUEUE, so a queued
field opened on a deleted entity and its commit went nowhere: draw a rounded rectangle, delete
everything, draw a 2-point circle, type 30 — the field opens, the digits are accepted, and the
radius stays 32.992020763. reset_autoedit() exists for exactly this and its own comment says so;
it was simply never called from here. The FEATURE GROUPS, whose [begin,end) ranges all shift on a
delete, so feature_of() answered with a group the user never drew — survivors are now remapped
and any group that lost a member is dropped, the rule the placed quotes already followed. And the
SOLVER STATE: no re-solve, so sketch_describe reported dof=16 for a document holding one circle.
snaporca-ekt9 (P2) — the read-back could not see three of its seven entity types. Ellipse,
EllipseArc and BSpline serialised as a bare type name: no centre, no semi-axes, no rotation, no
sweep, no poles. gui-ladder's ellipse rung had to grade the faceted area of the loop at 2e-2 —
that tolerance IS the faceting error — and its spline rung could only count entities. Now they
carry their parameters, and the ellipse arc's ends are asserted to satisfy (x/a)^2+(y/b)^2 = 1 to
1e-9.
Also read-only, and the reason the other two were found at all: sketch_describe now reports the
armed TOOL, the count of PENDING anchors, and whether a value field is EDITING. A menu walk that
lands one row off arms a neighbouring tool and then draws something plausible — the first run of
the authoring rung drew a circle of area 45238.93 and graded it as a rectangle. Every menu pick
now asserts which tool it armed, and the polyline rung (a per-segment Length field freezes the
canvas after every click) could only be written once the driver could ask whether a field was open.
Offer ladder 102/102 -> 105/105 with coverage. Gesture ladder 93/93 and the kernel suite
188 cases / 2532 assertions, both unchanged.
snaporca-ua9g snaporca-ekt9
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
The gesture ladder proved the TARGET — a complex closed profile, exact in vertices, lengths, arcs
and symmetry, voids correctly attributed. It proved it by arming every tool with a letter key,
which leaves the goal's own MECHANISM untested: the design logic pivots on right-click, and the
verbs offered are supposed to adapt to the element under the cursor. 47 of 86 Design-tab verbs
have a GUI action and no shortcut, so a key-driven ladder cannot reach more than half of them.
scripts/offer-ladder.py drives the menu. It asserts nothing from pixels: show_offer_menu emits an
[OFFER] trace from the same loop that builds the rows (behind the existing SNAPORCA_KEYTRACE), so
what the ladder reads cannot drift from what the user is shown, and the expected row set is
predicted by parsing DesignOffer.hpp rather than transcribed by hand. 25 properties, four rungs:
what each element type offers, that the menu equals the table for four selections AND that the
four differ, a 120 x 80 profile authored entirely through the menu, and a tool with no keyboard
route at all driven from the only door it has.
Two real defects, both found by it, both fixed here:
snaporca-ghcz (P1) — right-click was a black hole while any draw tool was armed. Every draw case
ended with `if (evt.RightDown()) { m_points.clear(); return true; }` and returned true even with
nothing to abandon; on_mouse records that in m_right_consumed and DesignCanvas suppresses the
offer whenever it is set. Measured: with Line armed, two right-clicks in a row produced no menu
and no tool change; only Escape freed it. Same rule snaporca-xmh6 wrote for the selection —
clearing nothing is not a gesture terminator. One shared right_abandon() now consumes the click
only when an anchor was really down; 16 sites, plus Polyline/BSpline (which end a chain, correct
only when there IS one) and Point (which has no anchor at all).
snaporca-lnri (P2) — right-clicking a sketch point offered the empty vocabulary. select_at_screen
tests hit_test_point first and records the hit in m_point_sel, but the offer counts m_selection
only, so a Point entity could never reach the entity branch and SkPoint was unreachable by
construction. A Point IS its own handle, so it is selected as an entity; other entities keep the
handle pick, since a line's endpoint is a drag target, not a vocabulary.
Offer ladder 25/25, gesture ladder 93/93 (no regression), both on the rig. The offer ladder joins
scripts/ladder-all.sh as the fifth rung.
snaporca-ghcz snaporca-lnri
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
Draw-then-edit is armed from a jump in the entity count: render() sees n > m_autoedit_seen,
selects the last entity and schedules open_primary_autoedit. A scripted add made while a
creation tool was armed looked exactly like a drawn gesture, so it opened that tool's value
field — and an open field freezes the canvas (on_mouse_impl returns early on m_awaiting_length)
and swallows every letter (in_text includes inline_busy()). Measured on the rig: after
sketch_add, 'p' + click added nothing (4 entities before, 4 after); one Escape and the identical
sequence gave 5. It also explains the selection = [last index] that sketch_describe reported
although action_sketch_add never selects anything — the render pass wrote it.
Escape worked because it sequences two set_tool calls: the pending CallAfter fires between them,
so the second one commits the field it finds open. Arming a tool directly is one call, and the
CallAfter fires after it.
Fix: resync m_autoedit_seen at the end of add_entities_scripted, so a scripted add is not read as
something the user just drew. An already-open field is left alone. Covers sketch_add,
sketch_mirror and sketch_offset — the three callers.
The scale rung's Escape workaround is deleted, which is the issue's acceptance criterion; it is
now the regression test. Gesture ladder 93/93 on the rig.
snaporca-j7gc
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
Over the whole 977-sheet corpus: 767 graded, 767 fully clean, 0 failures. The 210 skips are sheets whose part outline is not a closed stroked path at all (largest loop 5 to 132 mm2, measured), and one of them — MPD133 — is password-protected, which was being reported as an engine ERROR. Both now say what they are.
snaporca-j6sr
Two commits carried across (snaporca 579a9a9162, f68613cfc5).
Past about 480 entities a sketch had NO constraints at all and said nothing: libslvs
declares MAX_UNKNOWNS = 1024 and is handed every entity in the sketch at two params
per point, so the whole system came back TOO_MANY_UNKNOWNS and try_add_constraints
rolled the entire inferred batch back. From there no dimension could ever be applied.
Constraints only couple entities that share a point, so the solver now falls back —
only on TOO_MANY_UNKNOWNS — to solving connected components separately and committing
all-or-nothing. The auto-constraint pass batches its Horizontal/Vertical constraints
instead of one solve each, which is what kept the bulk path fast once solves started
succeeding: a 1204-entity load went 1585 ms -> 562 ms.
Plus the scale rungs (a thousand-entity plate drawn on by hand; the heaviest real
drawings graded and timed), the --step 1 fix that used to select nothing while
reporting a clean run, and scripts/ladder-all.sh as the one-command gate.
Parity 17 identical / 8 diverging as expected. Kernel suite here: 188 cases /
2532 assertions, including "a sketch past the solver's unknown limit still solves".
snaporca-yww4, snaporca-x6v7, snaporca-j6sr
and a ladder that draws with the mouse
Three commits carried across (snaporca 4ffd60eacb, 421055c2ec, b71216ce0b):
1. A design made only of sketches must survive being saved. CadDocument::recompute
returned false with "no solid-producing features" for a document that has no
solid, and two callers read that as "unusable": the GUI syncs the 3MF recipe
only after a successful recompute, so a sketch-only design was saved with no
recipe at all, and deserialize_recipe ends with `return recompute()`, so even a
project that carried one was refused on load. Having nothing to build is now a
success; a feature that MEANT to build a solid and produced none still fails.
DesignPanel::refresh_tree syncs the recipe too, for the paths that call
m_doc.recompute() directly.
2. Scripted geometry arrives exact. The Horizontal/Vertical inference window and
the endpoint weld window both close to zero for add_entities_scripted; void
attribution probes from a point strictly inside each loop instead of from its
first vertex. Corpus rung 39 graded / 39 fully clean, was 35 with 6 failures.
3. scripts/gui-ladder.py — 17 rungs, 84 properties, all driven by synthetic clicks
and typed values rather than through the socket.
Parity 17 identical / 8 diverging as expected. Kernel suite here: 188 cases /
2532 assertions.
snaporca-mtav, snaporca-8xg1, snaporca-5hvl, snaporca-730j
Upstream replaced MainFrame's fixed-index TabPosition enum with string-based
page ids ("feat: refactor notebook/tabs to be string based instead of fixed
index based"). The IMEX toolhead-visibility menu item was the only consumer of
that enum left on this branch, so its enable check now compares
m_tabpanel->GetSelectedPageName() against TAB_ID_PREVIEW -- the same form the
neighbouring upstream menu items use.
That mismatch is what broke CI: the branch built on its own, but the merge
commit CI builds no longer had TabPosition declared. No other conflicts.
666/666 tests pass in Release.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
- parallel modes now give the printing head its second-layer temperature
transition; it previously held nozzle_temperature_initial_layer all print
- single-tool primary mode no longer marks a phantom filament slot used, which
made machine_start_gcode heat an extruder that never prints
- the IMEX mode icon's bed raycaster is re-registered after the icon is
rebuilt, fixing a use-after-free in the picking pass
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
SceneRaycasterItem keeps the MeshRaycaster it was registered with as a raw
pointer, while PickingModel::reset() destroys it through a unique_ptr. Rebuilding
an icon therefore invalidates any registration still referring to it.
refresh_imex_icon(), reached only from Plater::on_config_change when is_imex or
the bed shape changes, rebuilt the IMEX mode icon without touching the
SceneRaycaster. The stale entry survived, and the next picking pass dereferenced
freed memory inside AABBMesh::intersect_ray.
Swap that one registration in place, matching how calc_vertex_for_plate_name()
handles the name-edit icon. Only the mode icon is registered for picking -- the
warning badge beside it is a plain GLModel -- so a single id is affected and the
blanket remove/re-register reload_scene() performs is not needed here.
Crashes were delayed and looked unrelated to the config change, because bed
raycasters are only tested when the camera looks down (SceneRaycaster::hit). The
reported dump landed on File > New Project, whose render ran a picking pass with
a registration that had gone stale earlier.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
get_imex_active_tools returned every physical head named by the active mode's
tool string, including the one carrying the Primary role. The pressure-advance
and nozzle-temperature sites are already gated on the mode not being primary, so
only the is_extruder_used supplement was exposed.
In primary mode that supplement treated the mode's single declared tool as a
secondary carriage and marked its filament slot used, so machine_start_gcode
emitted a heat command for an extruder that never prints. The phantom slot
appears when the mode's declared tool differs from the head the initial tool
routes to through physical_extruder_map -- on an AFC/MMU layout, printing with a
filament that lives on any head other than the declared one.
Return an empty roster for primary mode, where there are no parallel carriages by
definition. This lives in the enumerator rather than at the call site because the
mode is already resolved and normalized there, and the two guarded callers cannot
reach it in that mode, so their behaviour is unchanged.
Scope: this closes the primary-mode instance. The same phantom slot still occurs
in a parallel mode when the initial tool's head is not the mode's declared
Primary, which turns on which of the two notions of "primary" the three emission
sites should skip. That question is unresolved and deliberately left alone here.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The IMEX branch of the 1st->2nd layer temperature transition is mutually
exclusive with the standard per-extruder path in its `else`, but it skipped the
head carrying the print's own toolpaths on the premise that "the standard
per-extruder temp path already addresses it". That path is the `else` branch and
never runs for a parallel mode, so the printing head received no transition at
all and held nozzle_temperature_initial_layer for the entire job.
Emit for every carriage the mode drives, the printing one included. The printing
head takes this layer's own filament; the parallel carriages, which carry no
toolpaths of their own, keep resolving through the per-plate head map with pem
inversion as the fallback. The lookup now goes through get_filament_config_index()
like the standard path, since a variant-expanded printer gives a filament its own
column and a raw index would read the wrong one.
Reproduced on a 4-carriage IQEX in copy mode: the only temperature command in the
whole file set the idle secondary carriage to the value it already had, while the
head doing the printing never left its first-layer temperature. The defect is
invisible whenever initial and regular temperatures match, which is why earlier
per-tool validation passed.
Tests cover both gantry counts, since the active set comes from the mode's tool
roster: an IDEX copy mode drives two carriages, an IQEX mode drives four, and the
IQEX case asserts a first-layer filament and a second-layer transition for each of
the four.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Carries snaporca 9f0a6656bc.
Mouse3DController::apply DRAINS the input queue and every BOUND canvas idles and calls it, but a
hidden canvas's render() early-returns on _is_shown_on_screen() — so it swallows motion, applies
it to the shared plater camera, and draws nothing. The next visible frame jumps by more than one
state change. The plater keeps exactly one of its three views bound; the Design canvas binds once
at construction and never unbinds, so two canvases drain the same queue.
Guarded at the apply site so only the canvas actually on screen takes motion off the queue,
whatever happens to be bound, and without touching the plater's view-switching state machine.
Reported by exussum12 on PR #15238 as lag and jerkiness in the Design tab against "really smooth
on the other tab"; he guessed the mechanism correctly. NOT verified with a device — there is no
SpaceMouse here and the rig has no HID, so this is a mechanism traced in source and matched to a
user's description, not a measurement.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Carries snaporca 0231bd5b68. Parity holds: 17 files identical, 8 diverging as expected.
A reflection reverses orientation, so mirror_entities now hands the reflected half back reversed
in ORDER and flipped per ENTITY — an arc swapping its angles as well as its ends, a spline
reversing its control points. Appending it to the source then yields one walkable chain instead
of two halves meeting head-to-head, and a mirrored CCW loop stays CCW.
This is the producer half of the confusion that cost three defects; the consumers (offset, and
the exact loop area) keep their defensive handling, because that is what makes them correct for
hand-built and imported sketches rather than only for geometry this function produced.
Contract change, carried with the reason: a mirrored line's p0 is the reflection of the SOURCE's
p1, and a mirrored CCW arc keeps a POSITIVE sweep — the reflection negates it, walking it the
other way negates it again. Both [SketchEdit] cases updated, and a new [SketchProfile] case
"a mirrored half continues the original chain" pins the property directly.
Kernel here: all tests passed, 2687 assertions in 232 test cases. GUI target builds and links.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Carries snaporca 572f794c84, d0f9a0052a, 9f7e4e3627 and 3974f8a170. Parity holds: 17 files
identical, 8 diverging by their expected counts.
EXACT AREA. A loop's area is now integrated entity by entity in traversal order — Green's
theorem — instead of being shoelaced over the render polyline, which faceted every arc into 24
chords and lost 2.02 mm2 on a 3706.86 mm2 stadium. 0.054%, invisible on screen, and wrong in a
number reported as "the area".
OFFSET FOLLOWS THE TRAVERSAL. Offsetting a mirrored profile put one half on the wrong side and
split the loop in two, because the chainer only followed p1->p0 links and each entity's offset
side was taken from its stored direction. Chains are now orientation-aware, seeded at a free end,
offset by `reversed ? -d : d`, and normalised head-to-tail on the way out — so offset is correct
for any input ordering and its own output cannot reintroduce the problem.
Both are the same underlying lesson, which has now cost three separate defects: an entity's
STORED direction is not its direction of TRAVEL around the loop.
THE LADDER. scripts/sketch-ladder.py is a graded suite of 2D sketches judged the way a person
judges them — VERTEX, LENGTH, ARC, TANGENT, SYMMETRY, CLOSED — with area only as a cross-check,
because area is derived and nobody can confirm it by eye. Eight rungs from a rectangle up to
MPD5 from the StudyCadCam corpus, a dia 27 x 95 pin reproduced as its revolve half-profile with
the R5 fillet tangency solved exactly. Entirely 2D: no extrude or any solid feature.
Kernel here: all tests passed, 2681 assertions in 231 test cases, including the new
"profile: a mirrored half offsets as one loop, not two". GUI target builds and links.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Carries snaporca 95e59289f9, faec177d42 and 20df726ecb. Parity re-verified: 17 files identical,
8 diverging by their expected counts — DesignCanvas.cpp back to 16 and DesignPanel.cpp back to
32, which is the proof each hunk landed on the right side of the FeatFlyout and TAB_ID_PREPARE
divergences rather than on top of them.
All three answer exussum12's review on OrcaSlicer PR #15238.
ENTER/ESC IN THE VALUE FIELD. The field is a borderless always-on-top frame, and whether it may
hold keyboard focus is the platform's decision — a borderless NSWindow can never be key, and
mutter refuses a re-mapped window. When focus is denied the keys reach the panel instead and the
queued-dimension chain (a line queues Length then Angle) cannot be walked. The CHAR_HOOK now
forwards Enter/Numpad-Enter/Tab/Esc to the field when it is open and unfocused, and stays out of
the way when it is focused.
ESC FROM ANYWHERE. Separately and more simply: `dismissable` is false throughout sketch mode
because m_active is the FEATURE tool, so Esc fell through to whatever widget had focus. Click
any toolbar button or the Construction checkbox first and Esc did nothing at all — the likelier
reading of "Esc hardly ever works", and platform-independent. DesignCanvas exposes
request_sketch_exit() and the hook calls it whenever a sketch is live, after the inline-field
forwarding so an open field still takes Esc first.
RENAME. wxTR_EDIT_LABELS plus the two label-edit events write through to CadFeature::name and the
recipe, with a Rename verb in the offer and F2. The rebuild is deferred with CallAfter because
refresh_tree() destroys the very wxTreeItemId wx is holding during END_LABEL_EDIT — inline, it
killed the process.
STALE PICKS. set_tool now drops the Dimension tool's first pick, the Constrain picks and
m_point_sel, and delete_selected clears the pending dimension reference that could otherwise
dereference a renumbered entity.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Carries snaporca 39fac9b725. Parity re-verified: 17 files identical, 8 diverging by their
expected counts, DesignPanel.cpp still at 32 — the mirrored files were copied and the two
divergent ones patched hunk by hunk, so the counts returning to their expected values is the
proof each landed on the right side.
All 90 offer verbs are now firable by name over the socket, which matters because a deck key
can only send a keystroke and 49 of them have no shortcut at all. sketch_set_value calls the
same apply_dimension the in-canvas value field calls, so a typed dimension can be asserted with
no window manager in the way.
Three guards came with it, each confirmed against the source: on_mass_properties bounds-checks
m_sel_solid_body (it defaults to -1, and run_verb bypasses the menu grey-out that used to hide
that); sketch_set_value validates its value at the boundary because apply_dimension records a
driving constraint even for values it refused to apply; and run_verb refuses btn:/fly: verbs
that do not apply to the selection while leaving key: verbs alone, so the socket offers exactly
what the GUI offers. Dispatch is deferred through CallAfter so no modal verb can wedge the
socket thread.
GUI target builds and links against the rebuilt deps image.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Carries snaporca 971320e129, 6b049f0dc6, 4aae782029, 444d59f212, 74cf3d7e54 and the build
guards from 597557a6e4. Parity re-verified after every hunk: 17 files identical, 8 diverging by
their expected counts — DesignPanel.cpp still 32, DesignCanvas.cpp still 16, which is the proof
each hunk landed on the right side rather than being copied over a real divergence.
OFFSET OFFSETS THE CHAIN. Per-entity offsetting returned a closed rectangle as four parallel
segments that no longer touch, so entities_to_wires gave four OPEN wires and nothing could be
extruded. offset_entities now chains by shared endpoints and repairs each seam by mitering the
neighbours to their intersection. Second bug, invisible to any single-entity test: +d meant
"left of travel" for a line but "radius + d" for an arc regardless of sweep, so a slot outline
offset with its straights going one way and its caps the other. The convention is now written on
the declaration and pinned by a test.
tests/libslic3r/test_sketchprofile.cpp is new and asserts the LOOP rather than coordinates —
the property that decides whether a profile can be built, and the one the existing single-entity
[SketchEdit] cases cannot see. Its include is catch2/catch_all.hpp here: this fork ships Catch2
v3 while snaporca is on v2, which is why the test files are a tolerated divergence.
RIGHT-CLICK PICKS WHAT YOU POINTED AT, so a line's own verbs are offered instead of the
empty-selection vocabulary; sk_delete stops sharing btn:delete with the feature tree; and an
element's defining number (length / radius / diameter / angle / distance) can be typed, from the
menu or from V.
TWELVE MCP SKETCH VERBS. The socket had ~40 verbs and none touched a sketch, so the 2D layer
could only be exercised by driving a GUI with synthetic clicks. sketch_describe reports each
closed loop, the loops it encloses as voids, exact areas, and where a chain is still open;
sketch_validate/sketch_heal are FreeCAD's ValidateSketch — find vertices that overlap within a
tolerance but carry no coincidence, then weld them AND record the constraint, so a loop closed
by floating-point luck becomes one closed by construction. scripts/mcp-sketch-smoke.py is the
loop that asserts all of it.
Kernel suite on this fork: all tests passed, 2677 assertions in 230 test cases. The GUI target
links against the rebuilt deps image (the wxInspector blockage is gone) and the binary carries
the new verbs.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Building orcacad-deps from this Dockerfile removes the snaporca-deps base, which is the point
(Trap 1: the old image's baked tree was project(Snapmaker_Orca)) — but that base was also where
Xvfb, openbox, xdotool and scrot came from. Without them gui-session.sh reports display DOWN and
orca-slicer dies with a trace trap on no display. Verified: the session now comes up with
display up, wm up, and the Untitled - OrcaSlicer window present.
Raises the bound from -j8 to -j12: the incident dump measured 1.17 GB average per cc1plus,
so 12 in flight is ~14 GB typical, well inside the 40 GB cgroup ceiling that is the real
guarantee. Keeps the file byte-identical to snaporca's copy, which the header requires.
Dockerfile.deps builds the deps with -j 12 for the same reason, and symlinks
deps/build/destdir -> deps/build/OrcaSlicer_dep: this tree installs under the latter name
while the orcacad_buildcache volume has the former baked as absolute paths in its CMakeCache.
Same tree, two names — without the link, one directory rename costs a full cold rebuild.
Both forks ran this script at once on 2026-08-21, each with ninja -j$(nproc)=16.
36 cc1plus held 42 GB of a 62 GB box, the kernel OOM-killed for 2h28m, ssh went
unreachable, lightdm was destroyed (2946 session kill events), and neither build
produced a single object file.
Three bounds, weakest to strongest: a flock on a path SHARED by both forks so
they serialise instead of summing; -j8 so the box stays usable while it
compiles; and --memory on the container, which is the actual guarantee — a
runaway build now dies inside its own cgroup instead of taking the host with it.
--memory-swap is pinned equal to --memory because swap thrash is what made ssh
hang rather than fail.
Kept byte-identical to the copy in the snaporca fork, as the header requires.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Both were mouse-only: Commit to Plate is a toolbar button bound to wxEVT_BUTTON,
the bed is a CheckBox, and neither had an accelerator. That put them out of
reach of anything driving the keyboard, and out of reach of a hand that had not
already left the model to find them.
Ctrl+Shift, because the Shift+letter space is full to the last letter and
because the char hook deliberately ignores every Ctrl-combo -- which is exactly
what leaves this layer free to claim. P is Plate and B is Bed; neither collides
with OrcaSlicer own Ctrl+Shift+S (Save as) or Ctrl+Shift+G (Print plate), and
nothing else in the tree binds either.
The lookup goes ahead of the guard that drops Ctrl-combos, and nothing already
bound changes meaning: a plain Shift+letter still resolves as before, because
the new layer only answers when Ctrl is held as well.
The bed toggle drives the checkbox rather than the viewport alone, so the
control and the view cannot disagree about what is shown, and it says which it
did in the status line.
Both verified on the running build: Ctrl+Shift+B toggles the grid and the
checkbox together, Ctrl+Shift+P commits to Prepare.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Reported on PR #15238: on macOS, drawing a corner rectangle and pressing Enter
on the first dimension opens the second field and then wedges the entire
application — no typing, no Escape, dead menu bar, no tab switching, and the
field stays composited over the desktop after minimising. That last detail is
what identifies it: an already-drawn window keeps being composited by the
WindowServer once the process stops answering. The main thread is stuck.
4c8c93b512 is the only commit that ever changed the second-queued-field path.
It stopped unmapping the value field between two queued dimensions, and its
whole justification is mutter: focus-stealing prevention refuses keyboard focus
to a window that was just re-mapped, which left the second field visible but
dead on GNOME (snaporca-p8uw). It was applied with no platform guard, so macOS
re-activates an already-visible borderless NSWindow at NSModalPanelWindowLevel
from inside wxOSX's pending-event drain — which on macOS runs from a
CFRunLoopObserver at kCFRunLoopBeforeTimers (evtloop_cf.cpp) over an unbounded
`while (!m_handlersWithPendingEvents.IsEmpty())` (appbase.cpp).
One constexpr now carries that choice, and both halves of the contract read it,
because the failure mode of this fix is the two halves drifting apart: open()'s
reuse-if-shown and do_commit()'s deferred hide are one decision, not two.
Not a macOS guess I could not check: the non-GTK branch was exercised on the
Linux rig by forcing the constant to false and rebuilding. A corner rectangle
drew, its first field committed at 60 mm, the SECOND field opened, committed at
40 mm, and the sketch ended at 60.0 x 40.0 mm with the field closed and no
freeze — so the map-afresh path is functionally complete, not merely different.
On GTK the constant is true and every generated instruction is unchanged.
What is still unproven is the exact line where macOS wedges; the reporter has
been asked for a `sample` of the hung process. This fixes the cause the evidence
points at without waiting for that, and cannot regress the GTK behaviour.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Review request on PR #15238: "Place CAD-related files (e.g. CadDocument/
GeometryEngine) into a separate folder."
src/libslic3r/CAD/ the kernel — CadDocument, GeometryEngine, the four
Sketch* units, SketchSolver, ThreadStandards
src/slic3r/GUI/CAD/ the tab — DesignPanel, DesignCanvas, DesignSketchTool,
SketchInlineEditor, McpControl, generated DesignOffer
Pure relocation: no line of logic changes. Two include rewrites follow from it —
files that moved re-spell their own neighbours against src/ (already on the
include path), and files that did not move pick up the new folder. docs and
docs/ux/mockups/gen_offer_table.py follow the same paths.
Verified: libslic3r, libslic3r_gui and libslic3r_tests all build, CAD suite green
at 2518 assertions in 194 test cases, and the sibling fork builds identically —
17 shared sources still byte-identical, 8 diverging by their expected counts.
Review request on PR #15238: "Move the SLVS to deps if the source code remains
unmodified. If any changes were made to the source code, move it to deps_src."
It is unmodified — all 20 files under src/libslic3r/slvs were byte-identical to
JacobStoren/SolveSpaceLib@4d87045, the extraction of solvespace.com's libslvs.
So deps/ it is, fetched by hash like every other dependency.
Only the CMakeLists is ours: upstream's builds a demo executable and installs
nothing, so deps/SLVS/CMakeLists.txt.in replaces it via PATCH_COMMAND — the same
shape deps/OpenCSG already uses. The public header keeps its spelling, so
SketchSolver.cpp still says `#include <slvs.h>` and needs no edit.
The CI deps cache is keyed on hashFiles('deps/**'), so it rebuilds itself.
Verified: dep_SLVS builds and installs, libslic3r links against SLVS::slvs, and
the CAD suite is unchanged at 2518 assertions in 194 test cases.
The last unbuilt element of snaporca-wgsc. Two connectors a mate binds are one
object with a gap still in it; drawn as two separate frames they read as
unrelated, and "which two of these five frames are the mate?" had no answer on
screen at all.
Dashed, grey, drawn IN WORLD along the segment joining the origins -- so it
foreshortens with the model and its length is the gap the mate has left to
close. A Fastened mate therefore draws nothing, which is correct: the gap is
zero. Screen-constant dash pitch (6 px dash, 4 px gap) like every other gizmo
here, with the pitch opening up beyond 400 dashes so a mate across a large
assembly cannot emit thousands of segments. Depth test off: the line's job is
to say "these two belong together", and it has to say it even when a part sits
between the camera and one end.
Fed from BOTH sources of polarity truth, the same two the role colours already
use: every committed Mate feature (connectors named by feature index), and the
live pick of an open Mate card (named by combo ordinal). Only resolved frames
are eligible, so an unresolved end draws no line rather than a line to the
origin of the world.
RIG-VERIFIED on :10 with the fresh binary (/OrcaSlicer/build/src/Release,
2026-08-17 12:43): two imported bodies, a face-and-direction connector on each,
Planar mate offset 40. Sampling the segment between the two origins gives a
regular dash/gap alternation of 13:10 sample units -- the 6:4 px pitch -- in the
stroke grey (107,117,133), over both solids. The grey end keeps its open collar
head and the blue end its filled one, so polarity and pair now read together.
Refs snaporca-wgsc
Tommaso's decision (snaporca-x0kd): face orientation is hardwired perception -- a toddler reads
a face's roll and verse with no instruction -- so the connector is a face by default and the
conventional disc + roll quadrant stays, selectable, for users who expect it.
Preferences > Control > Camera > "Draw mate connectors as a face", default ON, key
design_connector_face_glyph. Read every frame rather than latched, so toggling takes effect on
the next repaint -- a look you cannot A/B without restarting will not get compared. Verified on
the rig: unchecking it switches the viewport to the disc live, no restart.
WHY A RELIEF AND NOT A DRAWING. A flat face in the connector's plane foreshortens by
sin(elevation) and collapses at a grazing view exactly like the quadrant it replaces -- measured,
the quadrant falls 89 -> 20 -> 3 -> 0 lit pixels from 47 degrees to edge-on. The relief does not:
its silhouette carries the information. So the glyph is a small shaded solid, painter-sorted,
lambert-shaded against a light fixed in CAMERA space so orbiting does not swing the shading.
THE MUZZLE, AND THE MISTAKE THAT NEARLY LOST IT. It is the only feature standing along +Z, so it
says which way the connector points and it is all that survives edge-on. Two errors on the way:
1. I built its footprint from height*tan(draft) and got a needle. The real base OVERHANGS the
crest at both ends (0.062 nose, 0.034 tail) and that overhang is what makes it a wedge. Base
now lifted straight off the mesh.
2. Worse, I chased fidelity. Scaled honestly the ridge is 11.3 mm on an 83.3 mm face -- 13.6 %
of the width -- and at 22-48 px that is a scratch. Tommaso looked at it and could not find
the muzzle at all, which is the only test that counts. A glyph is a symbol, not a scale
model, so it now gets two deliberate exaggerations, and COLOUR does most of the work:
muzzle share of lit pixels at 90/16/6 deg -- body tone 14.8/11.3/17.5 %, accent gold
18.3/19.2/23.9 %, accent gold at 1.8x width 23.5/25.2/31.2 %.
The accent is the same gold the disc spends on its roll quadrant, so it stays this tab's "here
is the direction that matters" colour. Polarity is still on the Z arrow's head; nothing collides.
A connector whose ROLL COULD NOT BE DERIVED keeps the disc treatment whatever the preference says.
A face asserts a definite orientation, and asserting one for a roll that was never derived is the
same confident lie that got billboarding rejected.
Geometry is emitted from the part by docs/design/mate-connectors/emit_glyph_table.py, not
hand-drawn, so glyph and printed connector cannot drift: 12-vertex outline, two eyes, chin bar,
cheek dot, and the snout wedge. Crest 29.0 mm / 6.58 mm drop / 13.1 deg against the review's
28.3 / 6.61 / 13.1 on the B-rep.
Also fixes extract_outline.py, which walked w.Edges: OCC returns them in storage order, not ring
order, ignoring per-edge orientation, so the outline was scrambled -- 45 points and perimeter
6.380 where a clean ring gives 31 and 3.335. Every measurement in the design notes was re-run.
The correction reversed one earlier finding: handedness does NOT read on its own (5.4/8.0/9.1 %
different from its mirror, not the 32-35 % the scrambled ring produced), so the cheek dot is
required rather than merely nice.
RIG-VERIFIED on Xvfb :12 against a 60x40x10 box with a face+edge connector: the face renders with
both eyes, ears, chin bar, cheek dot and a gold muzzle standing proud; the Z arrow degenerates to
its ring when viewed down the axis; and the preference switches to the disc live.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The connector work has lived outside the code since 2026-08-05, in a workspace repo with no
remote. It is the basis of a decision that now shapes the Design tab, so it belongs here.
docs/design/mate-connectors/
DESIGN_MATE_CONNECTORS.md seven CAD systems surveyed; the frame-pair model this kernel
already matches; sections 8b/8c on the glyph, and section 9's
four open decisions (D1-D4) still awaiting Tommaso.
bear.step the male, Onshape 2026-08-05T08:27Z, md5 faf228326ee3f971
BearConnector_Female*.step/.stl, BearConnector_Cutter.step
built by make_female.py FROM the real male B-rep rather than
re-modelled, so the pocket is complementary by construction
including every deliberate asymmetry. Fit measured at exactly
0.2000 mm, zero interference, mated hosts proven coplanar.
BEAR_CONNECTOR_REVIEW.md the symmetry-group result: identity 81/81 edges, mirror-x 0/81,
mirror-y 0/81, rot180Z 0/81, rot90Z 0/81, diagonal 0/81 at
0.1 mm. Trivial group, so every PARTIAL view fixes orientation.
extract_outline.py, simplify_study.py, relief_sheet.py, handedness.py, make_female.py,
trim_female.py, fit_check.py, verify_trimmed.py, coplanar_test.py + their sheets
THE DECISION THIS SUPPORTS (snaporca-x0kd): the mate connector is drawn as a simplified BEAR
FACE by default, with the standard disc + roll quadrant + Z arrow kept behind a preference.
Face orientation is hardwired perception -- a toddler reads a face's roll and verse with no
instruction -- and no abstract glyph earns that. Measured against the alternative: the disc's
gold quadrant+tick falls 89 -> 66 -> 37 -> 20 -> 3 -> 0 lit pixels as the camera drops from
47 deg to edge-on, and is a shapeless blob by 16 deg.
WHAT THE SIMPLIFICATION STUDY SETTLED (snaporca-wi3z), all measured off the real B-rep:
The eyes are load-bearing. Same outline and muzzle with the eyes removed stops reading as
a face at every size. Whatever else goes, they stay.
45 -> 22 outline vertices with no loss of read at 22 / 32 / 48 px; the muzzle reduces to
one filled triangle. Three marks plus a cheek dot.
Drawn FLAT the face fails exactly where the disc fails: in the connector's plane everything
foreshortens by sin(elevation). Rendered as its real relief instead, lit pixels at 32 px go
164 -> 210 at 16 deg and 69 -> 120 at 6 deg, and the snout ridge stands proud as a profile
rather than smearing. The glyph must be a shaded relief, not an outline.
Handedness already reads without any added mark -- 32 to 35 % of lit pixels differ from the
mirror, and re-registering by best whole-pixel translation returns offset (0,0), so it is
real shape asymmetry. But it reads only BY COMPARISON. A dot on one cheek makes it local:
34.5 / 37.0 / 36.4 %, and unlike uneven eyes (42 %) it does not read as a defect.
Tommaso's calls: it stays a bear, and handedness must read.
The scripts were repointed at the co-located male and extract_outline.py re-run from here to
prove it -- same 45 outline points, same three inner wires, same 3829.5 mm2 back plate.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
snaporca-wm4s. Thickening the 4-walled open box (60x60 in plan, 40 tall, no caps)
by 5 produced volume 29648.15 where the geometry requires (60^2-50^2)*40 = 44000
— about 67% of it. The corner material at the four vertical edges was simply
absent.
CAUSE. MakeThickSolidBySimple offsets each face along its own normal and sews;
it never extends neighbours to meet, so wherever two faces join at an angle the
corner is empty. A flat sheet has no such join and was always exact (18000.000),
which is why the defect looked like a measurement artefact.
WHY THE TWO EARLIER ATTEMPTS COULD NOT HAVE WORKED. Both switched to ByJoin —
plain, then with Intersection/GeomAbs_Intersection — and both returned a shell,
not a solid, so the body lost its volume entirely and both were reverted. That is
not a parameter problem: in OCCT, BRepOffset_MakeOffset::MakeThickSolid builds a
solid only inside `if (!myFaces.IsEmpty())` (BRepOffset_MakeOffset.cxx:1115).
Handed an open sheet with no closing faces, it stops after the offset shell and
returns it, reporting IsDone() with a non-null shape containing no TopAbs_SOLID.
ByJoin hollows a CLOSED solid by removing faces; an open sheet is outside its
contract.
FIX. Close the sheet, then use the call that mitres: cap the free rims
(ShapeAnalysis_FreeBounds -> MakeFace), sew shell+caps into a closed shell, make
a solid, and hollow it inward passing the caps as the faces to remove — the caps
come back off and leave the wall. Two details, each found by measurement rather
than reasoning:
* A shell sewn from an extruded sheet carries no guarantee of outward
orientation, and MakeSolid does not fix it. Inside-out, the inward offset goes
OUTWARD: measured bbox 70x70x40 and volume 339141.59, larger than its own
bounding box because the result overlaps itself. A negative GProp mass is
exactly that inversion, so it is the test; Reverse() on it.
* A SINGLE face has no neighbour to mitre and must keep the BySimple path. It
does have a free boundary, so "has free wires" is the wrong question — capping
a lone face with its own rim sews a zero-thickness shell and measures 6000
against 18000.
Also: IsDone() is not a success test here, since both failed attempts had it
true. The code now explores for TopAbs_SOLID and refuses a shell.
Tests: new case asserts 44000 with the wall's bbox at 60x60x40 (catching the
inverted-orientation shape, which has the right volume nowhere near the right
place), plus the flat-sheet control at 18000 that must not regress. Full kernel
suite green: 2502 assertions in 187 test cases.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Follow-up to the previous commit, found by driving the app: clicking Login /
Register on a fresh install pops "You are currently in Stealth Mode. To log
into the Cloud, you need to disable Stealth Mode first." — to a user who has
never touched the toggle, whose config says stealth_mode: false.
It is the same pre-wizard latch one step earlier. handle_web_request() gates the
login commands on get_stealth_mode(), which reports stealth while
firstguide/finish is unset, so the app blocks sign-in because the wizard is
unfinished — backwards, since signing in is how a user leaves that state.
Worse, the escape it offers does not work. "Quit Stealth Mode" writes
stealth_mode = false, which was ALREADY false, and never touches the latch: the
config is byte-identical afterwards and get_stealth_mode() still returns true.
The user clicks the button, believes stealth is off, signs in, and finds every
cloud feature still dead. That is the state the reporter of #15239 described.
So the login guard now reads the user's OWN setting via the new
get_stealth_mode_setting(), not the pre-wizard default. A user who deliberately
enabled Stealth mode still gets the dialog and the working Quit button; a user
who merely closed the wizard goes straight to the login page.
Measured on Xvfb with a fresh datadir (firstguide absent, stealth_mode false):
before, clicking Login produced a "Stealth Mode" window; after, it opens the
"Login" window directly. And with the previous commit's latch release, a real
Orca Cloud sign-in on that same unfinished-wizard profile now runs the whole
post-login flow — the sync prompt fires (sync_user_preset lands in the config),
the per-user preset folder is created, and Sync Presets syncs with no refusal.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Reported on #15239: after signing in to Orca Cloud on a fresh install, no sync
prompt appears and File > Sync Presets is greyed out with nothing to explain why.
CAUSE. AppConfig::get_stealth_mode() returns true whenever `firstguide/finish` is
unset, and that flag is written in exactly one place — GuideFrame::SaveProfile(),
i.e. only when the setup wizard is COMPLETED. Closing the wizard is what people do
today to reach the login (the wizard never offers it, which is #15239 itself), so a
new user ends up permanently in a stealth mode they never chose. Every cloud gate
keyed on get_stealth_mode() then switches off silently, including:
* GUI_App::on_user_login_handle(), which returns EARLY on stealth — so the whole
post-login flow is skipped: preset migration, plugin fetch, user-preset load and
show_sync_dialog(). That is the missing sync prompt.
* the Sync Presets item in both the top menu and the File menu, whose enable
lambda was `is_user_login() && !get_stealth_mode()`. That is the greyed item.
The result is indistinguishable from real Stealth mode, and nothing in the UI says
so, because the one place that DOES explain it — the "Quit Stealth Mode" dialog in
handle_web_request() — only covers the homepage login commands.
FIX, two parts.
1. The pre-wizard value is a DEFAULT for "the user has not been asked yet", not a
setting, so it must not survive the user answering. Signing in to a cloud account
is that answer. AppConfig now carries a session-only `m_cloud_logged_in` mirrored
from the network agent (on login, on logout, and at agent start so a restored
session counts), and get_stealth_mode() consults it before falling back to the
pre-wizard default. An explicit Stealth mode setting is untouched and still wins:
a user who turned it on deliberately stays offline whether or not they sign in.
2. Sync Presets no longer greys itself out. Both refusal paths already had a message
to show — "You must be logged in…" and now one for Stealth mode naming the
Preferences toggle — and the enable lambda was making both unreachable. A disabled
item that cannot say why is the reason this took a bug report to find.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
snaporca-x1k7 filed three Transform defects. Two were real and are fixed here;
the third does not reproduce and is withdrawn with its measurement.
(1) The feature tree had no ITEM_ACTIVATED binding at all, so double-clicking any
row only highlighted it. Double-click is the documented edit gesture elsewhere
(a committed sketch opens that way on the canvas), which made every feature look
dead until the user found the Edit button in the section header. Bound to
on_edit_feature(), so the gesture now works for every feature type, not just
Transform.
(3) The Transform card's typed fields called refresh_preview(), but preview_fields
returns {} for Transform and the solid-preview path has no ghost to build for a
feature that moves an existing body — so typing a distance changed nothing on
screen until Confirm. The fields now drive the same channel the gizmo drag already
uses: the body's display transform. In EDIT mode the committed transform is
already baked into the kernel geometry, so the preview undoes it first; without
that term, re-opening a committed Z=20 and typing 40 would show the body at 60.
(2) NOT REPRODUCED. Dragging a rotation ring does fill the field: a tangential
drag on the red ring gave Rotate axis = X, Angle = 27.44 deg, plus the translation
that rotating about the card's pivot implies (Y 17.60, Z -62.11). The original
reading came from a drag that never grabbed the 7 px ring; this run took its
candidate points from the rendered ring pixels themselves and 6 of 6 answered.
Rig-measured (docker snaporca-gui, Xvfb :10, llvmpipe), vertical screen shift of
the body by image correlation:
commit Translate Z 0 -> 20 : +140 px
double-click the Transform row : +0 px, and the card re-opens showing 20.00
step the re-opened card 20 -> 40 : +142 px (not +280 -> the undo term is right)
Cancel : +0 px vs the committed frame, residual 0.46
Add mode: stepping Z moves the body immediately (viewport diff bbox
200,143-1181,999); Cancel puts it back with only the status strip differing.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Review follow-up. Removing pem from the ghost cache key also removed its
accidental role as the key's only printer-identity signal: the resolved
active-tools string (roster + primary) and imex_firmware_managed_zones both
shape the baked zone/ghost set but were never keyed directly, so a printer
swap between presets with matching mode names and topology could leave a
stale ghost set. Key all three in build_imex_cache_key, which also hardens
the zone cache against the same pre-existing gap.
Also from review: the tooltip swatch reuses the pem/map its label already
hoisted (one resolution, not two); the bake constructs ghosts with no color
at all, making update_imex_ghost_colors the sole color author; the headless
!m_plater guard is documented as the wxGetApp sentinel it is.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Follow-up to the live ghost-color fix, addressing review findings: with the
render loop restamping ghost colors every frame, the bake-time resolution in
calc_imex_ghosts was dead code (its RGB was displayed for zero frames), and
the pem + head-filament-map entries in the ghost cache key had become
color-only inputs that forced a full mesh re-bake — including a visible
hitch on every ghost-picker selection — for what is now a pure recolor.
- calc_imex_ghosts bakes an alpha-only placeholder; IMEX_GHOST_ALPHA is
hoisted to file scope as the single opacity authority (no more reading
alpha back out of the field the restamp overwrites).
- build_imex_ghost_cache_key drops pem and the head-filament map; the
forced invalidations in set/reset_imex_head_filament_map go with them.
Picker selections now recolor live with no rebuild.
- The restamp moves into PartPlate::update_imex_ghost_colors(), beside the
transform refresh, so PartPlate owns its volumes' colors and the canvas
calls one hook. Plate-level inputs (pem, override map) are hoisted once
per frame via a new get_imex_head_filament_color overload that the
single-head form delegates to, keeping tooltip parity by construction.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
On a Wayland session the in-canvas value field never took the keyboard focus, so
after drawing a rectangle the first keystrokes went nowhere and the field had to
be clicked before a number could be typed. open() already did Show, Raise,
SetFocus on both frame and control, SelectAll, and re-asserted all of it in a
CallAfter — none of it works here, and no amount of re-asserting would: under
Wayland a client cannot focus itself, and mutter ignores gtk_window_present()
without an activation token as focus-stealing prevention. The earlier fix
recorded in this file (dropping wxFRAME_FLOAT_ON_PARENT, whose GTK _UTILITY_
hint made an xrdp session refuse focus) addressed a different compositor.
Stop needing WM focus. The canvas keeps the focus and feeds the field:
SketchInlineEditor::feed_key() types into the control directly — Enter commits,
Esc cancels, Backspace/Delete edit, digits and '-' '.' ',' are accepted, and
anything else is handed back so a stray letter cannot vanish into a numeric
field. A m_fresh flag reproduces the SelectAll semantics the field already had,
so the first digit replaces the prefill. It returns false when the control
genuinely holds the focus, so X11 keeps wx's normal routing and no character is
typed twice.
The CHAR_HOOK gates on the editor's own is_open(), NOT on inline_busy().
inline_busy is a freeze flag for the sketch tool: cleared on commit, re-set only
when the next queued field opens, with a CallAfter between them. Gating on it
left a window where the field was on screen and the flag was false — typing
worked for a rectangle's Width and not its Height.
VERIFIED at the machine on behemoth: typing the first dimension directly, with
no click, works. NOT yet confirmed: the Width -> Height handover; the is_open()
gate is diagnosed from the handover code, not observed. The hook's
SNAPORCA_KEYTRACE=1 switch logs each key with the focused widget if it needs
chasing further.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Select a line in a sketch, right-click, and every sketch verb was greyed: Trim,
Extend, Fillet, Chamfer, Offset, Mirror, the arrays, Constrain. The menu was
right and the selection was gone — two independent faults, each of which hid
the other.
FIRST, the Select-mode RightDown branch called clear_selection() before handing
the click back. Handing it back is correct: the m_right_consumed flag means "the
tool USED this right-click", and a plain right-click is not a gesture
terminator, so the offer should open. Clearing first is not: the offer describes
WHAT IS SELECTED, so wiping the selection guaranteed it could only ever describe
nothing. Deselection keeps its own gesture — left-click on empty space, a few
lines above in the same handler.
SECOND, offer_selection_kind() returned SkNone for every sketch state. The offer
table has always carried verbs for a selected line, arc, point or pair, but
nothing ever RETURNED those kinds, so fourteen rows were gated on selection bits
no code path could set. Classify the selection instead: SkLine / SkArc / SkPoint
/ Sk2Ent, via a first_selected_type() accessor on the tool and two forwarders on
the canvas.
Either fix alone measures as a failure — the classification is handed an empty
selection, or the preserved selection has no kind to match — which is why both
land together.
This is the second half of the report behind 3eb6e5d608: a user comparing the
Design tab with Onshape said "adding constraints seems to be missing"
(OrcaSlicer PR #15238). Constrain was one of the fourteen dead rows, and the
gesture that would have shown it threw the selection away first.
Verified at the machine on behemoth by Tommaso: select a line of a rectangle,
right-click, and the sketch verbs are live.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
On a non-English desktop the offer menu came out mixed: "Create / Add material /
Rimuovi / Fillet / chamfer / draft / Repeat / Transform / Reference / Modify",
and under Modify, "Elimina" beside "Constrain sketch".
Nothing was mistranslated. The row names went through a bare wxGetTranslation(),
which searches EVERY loaded catalogue — including wxWidgets' own wxstd. That
catalogue is loaded in the desktop's language whether or not the application has
one, and it happens to contain exactly two of our eight row names:
wxstd it: 'Remove' -> 'Rimuovi', 'Delete' -> 'Elimina'
Create, Add material, Repeat, Transform, Reference and Modify are not wx
vocabulary, so they stayed English. Two words in one language, six in another,
in the same menu — and the same trap is set for every other locale wx ships:
Supprimer, Löschen, Eliminar.
Name the domain: wxGetTranslation(s, SLIC3R_APP_KEY). These strings are now
translated by our own catalogue or not at all, which is consistent either way.
Left deliberately alone: the accelerator still renders as "Canc" rather than
"Del" on an Italian system. That is wx naming the physical key, and on an
Italian keyboard the key really is marked Canc — telling that user to press
"Del" would name a key they do not have.
Verified on the rig with LANG=it_IT: the menu now reads Remove and Delete, and
the submenu shows "Delete Canc" beside "Constrain sketch".
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A user evaluating the Design tab against Onshape reported that "adding
constraints seems to be missing" — with nineteen constraint types and a solver
shipped behind it (OrcaSlicer PR #15238, exussum12).
They were not wrong about what they could see. The only ways in were an
icon-only toolbar button whose tooltip you have to hover to read, and an offer
row gated on sketch_mode with a sketch ENTITY selected, filed under "Reference".
Right after finishing a sketch — the moment you want to constrain it — neither
was in front of the user, so a shipped headline feature read as absent.
Add a model-mode row: "Constrain sketch", offered under Modify when a sketch
region is selected, routed through the new btn:constrain verb action.
on_begin_constrain() also gains a fallback to m_sel_sketch_feat. The offer
reaches it from a SkLoop selection, which carries no TREE selection, and the
function read only tree_selection() — so the new row would have answered
"Select a sketch in the tree first" about a sketch the user had visibly
selected. It now adopts the region's owning sketch and syncs the tree to match.
Verified on the rig: draw a rectangle, finish the sketch, click the region,
right-click -> Modify -> "Constrain sketch" enters Constrain mode with
"Pick 1-2 lines, then a constraint" and the Constraints (8) card listing the
sketch's inferred constraints. That path did not exist before.
Does NOT address the other half of the report: there is still no Pierce
constraint, so a sweep profile cannot be tied to its path. Tracked separately.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
mass_properties on an open shell returned volume 96000 with an inertia diagonal
of [-4.2e7, -4.2e7, -6.9e7] for a 60x60x40 four-walled box — negative principal
moments, which no real body can have. BRepGProp::VolumeProperties integrates the
divergence theorem over whatever faces exist; on an open shell that is not a
volume at all, and the old code hid the only obvious tell by taking std::abs()
of the mass. "valid: true" then asserted the number was trustworthy.
This matters because mass_properties is what an agent or a user reaches for to
confirm a cut removed the right material. Silent nonsense there means the check
passes on garbage.
MassProps gains is_solid. For a sheet we compute surface area only — that stays
exact — and report volume 0 with the inertia left zeroed. The MCP verb returns
is_solid plus a note saying volume and inertia are not defined for an open
shell; the GUI's Mass command says "sheet body — N cm² of surface, no volume"
rather than quoting material that is not there.
Verified on the rig: the sheet now returns volume 0.0, surface_area 9600.0
(exactly 4 x 60 x 40), is_solid false. The solid controls are unchanged and
exact — a 60 mm cube reports 216000.0 and 21600.0.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Clicking a face that already happened to be selected, while a Plane or Axis pick
was armed, read as a repeat pick: the click escalated to "whole body", the
capture was lost, and the card's label stayed "(none)" with nothing on screen to
explain it. On a cube it is easy to hit — the face under the cursor is often the
one already selected from the previous step.
The capture path in on_solid_picked already restores the flag for all three
tools, and reset_plane_refs()/reset_axis_refs() restore it when a pick is
abandoned — both were written as if the arm side disabled escalation. Only
CoordSys actually did (that was snaporca-u0wd). Plane and Axis never had it.
Verified on the rig: Midplane on a 60 mm cube now captures Face A (#5, top) and
Face B (#3, side) on the FIRST click each, and the resulting plane renders as
the 45-degree bisector between them, which is what a midplane of two
perpendicular faces should be. Before this, the first pick escalated to the body
and Face A stayed "(none)".
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Every method in CadDocument's plane dispatch falls back to offset_angle_plane()
when its references are missing. Picking Tangent and confirming with nothing
selected therefore produced an OFFSET plane, announced as "Plane added — pick it
as a sketch plane". The user asked for one construction and silently received a
different one, with nothing on screen to reveal the substitution.
Validate at the GUI boundary instead: Angle needs an edge, Midplane two faces
(and not the same face twice — that yields a plane coincident with the face,
which is well-defined and useless), Tangent a face, Two-edges two edges. Offset
and Coincident are unchanged: both are meaningful with no reference, since they
fall back to the base plane by design.
on_add_plane() now returns false when it refuses, and confirm_tool() skips
close_tool() in that case — a refusal that also threw away the picks the user
had already made would be worse than the bug.
The kernel keeps fallback_offset(): it must return something. It should just
never be reachable from a user gesture without a warning.
Verified on the Xvfb rig: Tangent with no pick refuses and creates no feature
(it created one before), the card stays open with the type preserved, Midplane
with no faces refuses with its own message, and Offset with no picks still
creates a plane as it always did.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
wxSpinCtrlDouble takes the mouse wheel whenever the pointer is over it. A card
taller than the panel therefore could not be scrolled past without silently
incrementing whatever field the cursor happened to be over — measured on the
rig: eight notches over the Coord Sys card turned an X hint from 1,00 into 6,00
while the panel did not move at all. The same gesture over an Extrude distance,
a Hole diameter or a mate Offset is a silent model change made by someone who
believed they were navigating, and nothing on screen reports it.
Every spin in this panel comes from one factory, so the guard goes there: an
unfocused spin hands the wheel to its parent, and the scrolled cards panel gets
it. A spin the user has deliberately focused still takes the wheel, which is the
one case where editing is what was meant.
Verified on the rig with an Extrude card: eight notches over an unfocused
Distance leave it at 10,00; clicking into it first and scrolling takes it to
13,00.
A mate between two connectors committed cleanly, recomputed without error, and
moved nothing. apply_mate was reached and computed a translation of exactly
(0, 0, 0).
The connectors were the cause, not the mate. Picking a face stores
coordsys_face and coordsys_body but left the Type combo on its default,
Point (world) — and datum_frame ignores the pick entirely for that type,
resolving the connector to coordsys_point, which is (0,0,0) unless the user
typed otherwise. Two connectors built that way share one frame, so the mate
between them is an identity transform: everything reports success and the
assembly never moves.
Capturing a face or an edge now switches the type to FaceAndDirection. Picking
a face IS the choice of a face-based frame; asking for it twice, with no hint
that the second half is required, is what made every mate a silent no-op.
Verified on the rig, two bodies with a face-based connector each:
before Extrude4 [-71.6, 402.6, 156.5] .. [-8.8, 412.6, 237.3]
after Extrude4 [-31.4, -40.4, -10.0] .. [ 31.4, 40.4, -0.0]
with the mate transform now (40.19, -196.88, 402.56) instead of (0, 0, 0).
That is the first mate in this tree that assembles anything.
Not the kernel: recompute applies mates exactly as preview does, proven by an
A/B harness over all five kinds — the two paths give identical bounding boxes.
1. A consumed loop WITH HOLES was never dropped from the sketch overlay.
sync_sketch_display compares each region's entity list against what a per-loop
extrude stored, but rebuilt the candidate from the region's OWN entities while
the extrude stores the region's entities PLUS every hole's (see
selected_loop_entities). For a plate with one bore that is 4 against 5, so the
match never fired and the extruded rectangle stayed drawn on top of the solid
it had become. Adds region_entity_indices_with_holes, which returns the same
order the extrude uses, and compares against that; a matching region now drops
its holes' entities too. Hole-less regions are unaffected.
This is also the artefact that made a correct plate-with-a-bore read as a
plate with a plug in it during rig testing.
2. The Constraints card drew its header and its first row on top of each other.
The rows and the delete buttons were parented to m_form rather than m_cards,
so they were laid out in the wrong window's coordinate space and started at
the card's top edge. Re-parented; nothing else about the card changed.
3. The mate hover ghost never appeared. The highlight handler asked for a repaint
with request_repaint(), which only queues a Refresh — and a wxMenu popup runs
its own modal loop, so the paint was not serviced until the menu closed, by
which time the ghost had been dropped. Adds DesignCanvas::repaint_now(), which
flushes the paint immediately, mirroring the m_status->Update() the status line
in the same function already needed for the same reason.
Delegated to opencode (DeepSeek V4 Pro) and reviewed by diff. Fix 1 needed an
accessor on DesignSketchTool because region_loops/RegionLoop are private — that
was outside the file list it was given, and it said so rather than working around
it.
Verified on the rig: a rectangle-plus-circle sketch extrudes to a plate with a
bore and NO overlay left on top of it, and the Constraints card shows its header
clear of eight readable rows.
Leaving Sketch clears the "N degrees of freedom" line, which is right — it
describes a sketch's constraint state and means nothing in Feature mode. But
entering CONSTRAIN left it blank too, and Constrain is where the number is the
whole point: the readout is fed only by a live solve, and no solve fires merely
because the mode changed, so the line stayed empty until the user happened to
change something.
The solve callback now caches its last result and the labelling is split into
apply_dof_status(), which set_ui_mode re-applies on entry to Constrain.
Verified on the rig: a rectangle reports "4 degrees of freedom" in Sketch, and
the same line is there after K enters Constrain.
The edge and vertex tolerances were fixed at 8 and 11 px. On a face that is
barely wider than that on screen — a thin plate, or any part once you zoom out —
every point on it lies within the edge budget, so the pick alternated edge and
whole body and the FACE level could never be reached at all. That is not just
awkward: a face pick is what gives a Coord Sys its owning body, and a mate needs
one, so thin parts could not be assembled.
Both tolerances are now capped at a third of the face's shorter on-screen side,
measured from the edge samples the picker already walks. They only ever shrink,
so a face with room keeps the full budget and nothing changes for ordinary
geometry; a narrow one keeps its middle for itself.
Verified on the rig on a 16.7 mm-wide plate at three zoom levels including one
far enough out to make the face a thin sliver: the middle reports "face 5
selected" every time, while points near the rim still take the edge.
A card that asks for a face ("Click a solid FACE in the viewport") could not be
satisfied. Clicking the same sub-element twice deliberately escalates to the
whole body — right for free picking, wrong here: the user clicks the very face
the card is pointing at, the escalation turns it into a whole-body pick, and the
armed capture rejects it and leaves "(none)". Both orders failed, so a
face-based Coord Sys was reachable only by accident of ordering. That mattered
beyond the card: a mate needs a connector with an owning body, and a face or
edge pick is the only thing that sets one.
Adds DesignSketchTool::set_escalate_on_repick, off while the Plane, Axis or
CoordSys card has a pick armed and back on as soon as it is captured or
abandoned. While armed, clicking a face means "this face", which is what the
prompt already says.
Verified on the rig: arm Pick Face, click the face, and it reads "#5" on the
first click. With nothing armed the escalation still alternates whole <-> face
as before.
The mate palette called all five kinds viable on connectors created as
Point(world), which belong to no body. Clicking one built the Mate feature and
the RECOMPUTE then failed with "mate: mate_cs_b has no associated body" — the
refusal arrived one step too late, after the feature was already in the tree,
and the user is told about it by an error line rather than by the palette that
offered the thing.
mate_options now checks the same two conditions the apply path throws on: B must
have a body (B is the connector whose body moves; A is the fixed reference and
needs none), and that body must still resolve. Either way all five kinds go
non-viable with a reason, so the offer and the kernel cannot disagree.
Verified on the rig: with two Point(world) connectors, every row is now dimmed
and reads "connector B is not attached to a body — a mate moves B's body". That
also exercises the dimmed-with-a-reason presentation for the first time, which
until now had nothing to show because every kind was always viable.
Tests: a new [mate] case covering no body, a body that no longer resolves, and
the revival once B is given one. One existing case needed its setup widened
rather than its assertion weakened: "an unrecorded fingerprint does not make a
type non-viable" built two body-less connectors, so it was asserting a side
effect of the old permissiveness instead of the property it is named for. Its
connectors now have a body, leaving the missing fingerprint as the only variable.
Full [CadDocument] suite: 2458 assertions in 185 cases.
Right-clicking a document with two coordinate systems builds the mate palette
exactly as designed: a separator, a disabled "Mate: A -> B" header, then five
rows in fixed order. Hovering a row showed no ghost and left the previous status
message on screen; clicking one created nothing at all. The palette enumerated
perfectly and fired nothing, which put the whole M8 mate epic out of reach from
the UI.
The mate handlers were never invoked. show_offer_menu binds two pairs of
handlers to the SAME wxMenu: the mate pair (by mate_base, at base + 500) and the
generic verb pair. wxWidgets pushes dynamic entries to the FRONT of the handler
list, so the pair bound LAST runs FIRST for every id — and the generic pair was
last. A mate id lands outside the verb table's range, so both generic lambdas
returned early WITHOUT e.Skip(), which wx reads as "handled", and the mate
handlers behind them never saw the event.
Binds the two generic handlers to the verb id range [base, base + 499] so each
half only sees its own ids regardless of bind order.
Verified on the rig with a purpose-built assembly (two bodies 60 mm apart, a
Coord Sys on each): before, clicking Fastened produced no feature and no status
change; after, it produces a Mate feature. The verb rows keep working — the same
run created both coordinate systems through Reference > Coord Sys.
Note the mate then fails its recompute with "mate: mate_cs_b has no associated
body", because a Point(world) coordinate system belongs to no body while the
palette still advertises all five kinds as viable. That is a separate defect and
is filed; this commit is about the palette being reachable at all.
Reopening a saved CAD design ended on a modal "The file does not contain any
geometry data." warning. A CAD project legitimately carries no mesh — the model
lives in the feature tree (Metadata/SnapOrca_cad.bin) until Commit to Plate — so
the warning is false for one, and it is the last thing the user sees after
opening a design they spent an hour on. It reads as "your work is gone" at the
exact moment the recipe HAS just loaded and the Design tab is about to rehydrate
it, and the main window sits disabled behind the dialog until it is dismissed.
Counts a non-empty model.cad_recipe as geometry.
Verified on the rig: save without Commit to Plate, reopen, and the Design tab
comes up with Sketch1 -> Extrude2 -> Body 1 editable, with no dialog at all.
A rectangle with a circle inside it, drawn in ONE sketch, could not be extruded
to a plate with a bore from the GUI. Five defects were in the way. Each was
found by driving the app on a headless rig and measuring the result — the code
reads correctly at every one of these points, which is why they survived.
1. Wire orientation (kernel). SketchEngine::wires_to_face added every hole as
wires[i].Reversed(), which is only right when the sketch happens to wind both
loops the same way. A circle drawn clockwise inside a counter-clockwise
rectangle came out matching the outer boundary, OCCT swept it as a SECOND
contour, and the prism was the plate with its bore filled and the disc's
volume counted twice. Measured: bbox 67.17 x 219.67 x 10 with volume
152088 mm3 against a solid box of 147542 — a body larger than its own
bounding box, which is the signature. Holes are now added as-is and
ShapeFix_Face::FixOrientation() classifies them; that is winding-independent
and is the idiom make_extrude_regions already used for imported glyphs, which
is why holed TEXT always extruded correctly while a holed SKETCH never did.
After the fix: 142996 mm3, implied bore radius 12.03 mm against the circle
drawn.
2. The live-sketch click threw the picked region away. region_at() served only
as a yes/no gate and on_face_selected() carried no argument, so Extrude fell
back to whichever loop the resolver found first — clicking the material of a
plate-with-a-hole extruded the disc. The region is now carried through, and
DesignPanel also hands it to the tool with set_loop_pick(), AFTER open_tool()
because that re-derives selection state, since extrude_uses_loop() reads
selected_loop_entities() and that lives on the tool.
3. region_at() had no hole awareness and no innermost preference: it returned
the first polygon containing the point. It now skips a region when the point
lies inside one of that region's holes, and picks the smallest containing
loop, so a click in the bore selects the disc and a click on the material
selects the plate.
4. Startup segfault. DesignCanvas::request_repaint probed the GL backend via
OpenGLManager::get_gl_info().get_renderer() before the canvas had initialised
GL — glGetString with no context current and, before init_opengl(), no loaded
function pointers. Anything that asked for a repaint while the panel was
still being built landed there, with no window and nothing in the log. It now
bails at the top on !is_initialized() and asks for a Refresh instead. Note
the crash was in the PROBE, not in render(), which already guards itself.
5. A holed sketch on a plane whose normal points -Z came out as the full box PLUS
a disc — 220274 mm3 where 163726 was due (192000 + 28274). wires_to_face took
a SketchPlane parameter it never used and let OCCT infer a surface from the
outer wire; when the inferred normal disagreed with the sketch's, the hole
classification produced no hole. Every face is now built on the sketch's own
gp_Pln.
Also in this change, from the same rig session:
- A right-click that only clears the sketch selection no longer reports itself
as consumed, so it stops suppressing the offer menu. With any geometry in a
live sketch there was no menu route left to add a second entity.
- Escape no longer discards a live sketch that holds drawn geometry; it says so
and keeps the work (live_sketch_has_work()).
- The holed-region fill is an even-odd scanline instead of a keyhole bridge, so
no corridor triangle leaks from the bore to the nearest corner.
- Cyan is reserved for the selection: an unselected region no longer wears a
shade one step off the selected one.
- The origin planes follow the mode, so pressing Sketch on a document that
already has a body offers them again instead of naming a plane you cannot see.
Tests: three [holes] cases over add_extrude_entities asserting the plate-with-bore
volume, solid and face counts on both a +Z and a -Z sketch plane, and the
by-name refusal of two disjoint regions. Full [CadDocument] suite green.
Delete a sketch on a document that still has a body and you could not start a new
one. Pressing Sketch said "click a face or a reference plane in the viewport" —
while update_reference_planes had already called clear_base_pick, because a body
existed. The instruction named something that was no longer there, and short of
finding a face to click there was no way back into sketching at all.
The planes are now offered when there is no solid yet OR while the UI is in
Sketch mode, and set_ui_mode refreshes them so they appear the moment you press
Sketch rather than at the next tree rebuild — which is not an event that pressing
Sketch causes.
Deliberately not always-on. m_dbp_active both RENDERS and picks, so leaving it
set would float three translucent planes over every finished model. Tying them to
the mode shows them exactly when they are the thing being chosen and takes them
away again on Finish.
Safe against stealing clicks: a base-plane pick is the last resort in on_mouse,
firing only on a click that hit no geometry, so solids and committed sketches
still win where they overlap.
Found on the rig by Tommaso: "if i remove a sketch, i cannot create sketches
anymore".
A rectangle with a circle inside it extruded to a plain box. Tommaso reported it
exactly right on the first attempt — "no intersection selectable, hence no plate
with hole" — and it was a causal chain, not a guess.
Two things were wrong and they compounded.
region_loops() returned N independent filled polygons with no notion of nesting,
so the only selectable things were the rectangle alone and the circle alone. The
region a user actually wants — the bounded area WITH its hole — did not exist to
be pointed at. Worse, the first polygon containing the click won, so clicking
inside the circle selected the rectangle.
And extrude_uses_loop() hands selected_loop_entities() to add_extrude_entities,
which copied only that one loop's entities into the feature. So the circle never
reached the kernel, build_sketch_face saw a single loop, and the multi-loop path
added in 5c4ced91e7 never ran. Proven from his saved project: Sketch1 held 5
entities, Extrude2 held 4, and the committed mesh was 8 vertices — a box of
260.40 x 220.91 x 10.00. Eight vertices cannot describe a bore.
A RegionLoop now carries the loops nested inside it. Containment is decided by
testing one vertex, which is sufficient because loops in a well-formed sketch do
not cross, and each loop is assigned to the SMALLEST loop containing it so a hole
belongs to the region that actually bounds it. Picking respects holes: a click in
the plate selects the plate, a click in the bore selects the disc. The selection
hands over the region's own entities plus its holes', which is what finally
reaches the kernel. The highlight lights the holes with their region, because it
has to show what will be extruded.
The status line said "Loop selected"; it now says "Region selected". What is
selected is a bounded area that may contain holes, not a single closed curve —
the old wording described the old, broken behaviour.
snaporca-txp8, and it is what makes snaporca-88v reachable from the GUI at all:
the kernel could build the holed face all along (verified on his own recipe:
2 closed loops, wires_to_face OK, area 74812.119 mm2), but nothing could ask it to.
Reviewed and compiled (RC=0). NOT exercised — the rig check is the point.
Modelling in the Design tab and pressing Ctrl+S saved a project with no feature
history at all, and the app reported success. Found on the rig: a project saved
after drawing a rectangle and a circle contained twelve archive entries, none of
them Metadata/SnapOrca_cad.bin, and a 3dmodel.model with zero vertices.
plater->model().cad_recipe was assigned in exactly one place — on_commit(),
immediately after load_mesh_object. The 3MF exporter was never at fault: it
faithfully wrote whatever the Model held, and on a save that had not gone through
Commit to Plate that string had never been set. The recipe reached the Model only
as a side effect of a different user action.
It now tracks the document instead. sync_recipe_to_model() is called after a
successful recompute, after tree edits (deletes, reorders and suppressions bypass
recompute_guarded), and from on_commit, which delegates rather than repeating the
rule. Only on success — a failed recompute leaves the document mid-edit, and
persisting that would save a model the user never had. An empty document still
clears it, so a non-CAD project carries no stale recipe.
Doing it here rather than in the save path is deliberate: Ctrl+S, Save As,
autosave and crash recovery all read model.cad_recipe, so keeping it current
after each change makes every one of them correct at once, instead of teaching
each save path to ask the Design tab. Commit to Plate means "send this to the
slicer" — making saving depend on it was the bug, not the cure.
Cost is one serialization per recompute, tens of KB against an OCCT rebuild that
has just run.
Why nothing caught it: the kernel round-trip tests serialize a CadDocument
directly, and the 3MF tests exercise the exporter with a recipe already present.
Neither can observe that the GUI never populates it, and every save in testing
happened to follow a Commit to Plate.
snaporca-vjk5. Reviewed and compiled (RC=0); persistence NOT yet confirmed on the
rig — that check is the reason the issue stays open.
This script builds only libslic3r_tests, which links libslic3r and no GUI code —
but cmake still processed the whole if(SLIC3R_GUI) block and every find_package
inside it, so the kernel suite silently depended on the GUI's dependency set.
That came due the moment upstream added wxInspector as a REQUIRED find_package:
the orcacad-deps image predates it, so configure died pointing at
src/CMakeLists.txt:92 with nothing about the kernel having changed. Turning the
block off is not a workaround for that one dependency — it is the suite finally
declaring what it actually needs, so the next GUI-side dependency added upstream
cannot break it either.
Surfaced by taking SoftFever's merge of main into the PR branch.
He merged upstream main into the PR branch himself on 2026-08-13. Taking it into
the local branch rather than force-pushing over it: the fork copy is what PR
#15238 shows, and discarding a maintainer's merge to make my own push
fast-forward would be both rude and a loss of 130 upstream commits.
Brings the branch far closer to main than the 2026-07-24 merge-base the PR body
describes, which is most of what snaporca-36u9 was filed for.
Every version bump so far has permanently orphaned every project saved before
it. deserialize_recipe refused anything that was not exactly the current
version, and with no migration path v2 and v3 projects are unopenable today —
the 3MF still carries the mesh, so the user gets a frozen solid and no feature
history, which is the whole point of the subsystem silently absent.
The cause was the shape of the data, not the gate. save/load is one flat
symmetric list of ~90 fields with no framing, so a reader has no way to know
where a feature ends unless it agrees on every field.
Each feature is now written as its own cereal stream behind a length prefix, and
the same few lines handle both directions of mismatch. Older file, newer build:
the sub-stream ends early, the read throws, and the fields already assigned are
kept while the rest default — cereal assigns sequentially, so a mid-list throw
leaves the earlier fields set, and that is what makes this work. Newer file,
older build: the sub-stream holds more bytes than the reader knows; it reads what
it knows and stops, and the outer stream is untouched because the length prefix
was consumed in full. A field a project predates is not a corrupt project, so
neither case is an error.
v4 keeps its own pre-framing flat path and opens exactly as before —
cad_recipe_v4.bin is untouched and now serves as the witness for that. v2 and v3
stay refused, by name: their field lists no longer exist in this code. This fixes
the future, not the past, and the comment says so rather than implying otherwise.
From here a new field only needs appending to save/load — no bump, no orphaned
projects. That removes the cost that had blocked snaporca-44m and snaporca-dgv.
The helix round-trip test was reading the blob back flat, reaching into the
format instead of through it; framing necessarily breaks that, so it now goes
through deserialize_recipe, which is a stronger assertion than it made before.
Every field check it carried is unchanged.
Tests: four new [CadDocument][recipe] cases, including the one the change exists
for — a deliberately truncated feature blob must LOAD, keeping what it could read.
Suite 177 -> 181 cases, 2366 -> 2411 assertions.
snaporca-2txy.
deserialize_recipe distinguishes three cases that matter very differently to the
person reading the message — saved by a NEWER build, saved by an OLDER one, or
genuinely unreadable — and names the version in each. load_recipe threw all of
that away and printed one generic sentence, so the user could not tell "update
SnapOrca" from "your file is damaged", and had no way to find out.
Same error-loss class as the 31 McpControl sites fixed in 1de72de9ed: the message
existed, it was simply not passed on. The generic sentence stays as the fallback
for the case where the kernel really has nothing to say.
This does not make old projects loadable — that is snaporca-2txy, which the audit
behind this change opened. It only stops the reason being withheld.
snaporca-2txy (partial). Reviewed and compiled (RC=0), not exercised.
entities_to_wire handled exactly two shapes of sketch: one lone Circle/Ellipse, or
any number of Line/Arc/EllipseArc/BSpline pushed into a single MakeWire. Everything
else fell off the end as a null wire, so a circle drawn inside a rectangle — the
most ordinary thing in this whole program — refused with "not supported yet". Two
separate closed polygons were quietly worse: both went into one MakeWire, which
does not mean "two loops" to OCCT.
entities_to_wires now returns one wire per loop. A Circle or Ellipse is a loop on
its own; chain entities are grouped by shared endpoints (union-find, 1e-6 in sketch
coordinates), and an open chain still comes back as a wire because a sweep path is
legitimately open. It is all-or-nothing: one loop that fails to build poisons the
whole result, because a partial profile would extrude a shape the user did not draw
— the failure 2e6a8f9e91 was written to stop.
entities_to_wire survives as a two-line wrapper returning the single wire when
there is exactly one loop and a null wire otherwise, so all nine of its call sites
keep their exact contract and Revolve/Sweep/Loft/Surface* are untouched. What a
holed profile means for each of those is a separate question.
wires_to_face takes the largest-area loop as the outer boundary and adds the rest
reversed, which is how OCCT is told a wire is a hole. Containment is CHECKED with
BRepClass_FaceClassifier, not assumed: a loop outside the largest one is a second
island, and one sketch producing several solids is a much bigger feature, so it is
refused by name ("two disjoint regions") rather than guessed at.
Only the Extrude case consumes the new face. Tapered extrudes of a holed profile
are refused — offsetting inner loops has to go the opposite way — and the guard
counts wires on the face already built rather than rebuilding every wire to ask how
many there are, which is also the more honest test: what matters is the profile
being extruded.
Tests: six new [CadDocument][sketchwire] cases, proved by VOLUME rather than by not
throwing — plate-with-hole, two holes, and two regression guards that a lone circle
and a lone polygon extrude exactly as before. Suite 177 cases / 2366 assertions.
No serialized field, recipe version untouched, golden fixtures unchanged.
snaporca-88v.
Third and last piece of the selection model. The other two turned out to be
built already — the rubber band is pick_bodies_in_rectangle and vertex picking
is SolidSel::Vertex with its camera-facing square, both live — so this closes
what the issue actually still described.
Vertex beats edge beats face is a rule the user cannot see until after they have
committed to a click. Showing the outcome under the pointer is what makes the
precedence learnable at all, and is the charter's L5 read honestly: one click,
one visible change means the change has to be predictable BEFORE the click, not
only explicable after it.
The resolution is now one function, resolve_solid_pick, const and writing only
into its out-parameter. The click applies it and then runs its escalation
unchanged; the hover applies nothing. Split this way the promise cannot drift
from the act — a second implementation of "what is under the cursor" would
eventually disagree with the first, and the disagreement would look like a
picking bug rather than a duplication one.
Rendering is likewise one function called twice. The pre-highlight draws first
so the committed selection paints over it, and is suppressed entirely when the
two are the same thing: two coats of the same colour reads as a rendering fault,
and a promise about a click that would change nothing is not worth making. It is
desaturated toward white rather than given its own hue — a distinct colour would
read as a distinct KIND of selection, when it is the same selection one moment
earlier.
Two things that would have been silent bugs. The edge ribbon and the vertex
square render with GL_BLEND off, so an alpha below 1 there is ignored; those two
are quietened by a muted rgb and only the blended face fill takes the alpha
multiplier. And the pre-highlight is cleared in clear_solid_selection, because it
names a face by an index into a shape a recompute has just rebuilt — left behind,
it would keep glowing on whatever now sits at that index, a real entity but not
the one meant.
Hover runs on plain motion only, with no button down and no band running: during
a drag the pointer is doing something else and a promise about clicking would be
a lie. It returns false so the event still reaches the camera — it asks for a
repaint, it does not consume the gesture.
snaporca-9xw. Reviewed and compiled (RC=0), not exercised.
Tommaso reported the array controls as missing. They were not — Shift+N opens a
Pattern card with every control correct — but the report was fair. With no body
the button accepts the click, opens nothing, and writes its refusal somewhere
other than where the click happened. From the user's seat that is
indistinguishable from a dead button, and the icon is one unlabelled glyph among
fourteen, which is how I mis-clicked it into Section view while reproducing this.
A control that cannot act should look like it cannot act, before it is pressed.
The three FEATURE buttons carrying a body-count guard — Pattern and Cut at one
body, Boolean at two — are now greyed below their threshold with a tooltip
naming what is missing.
Only those three. The same guard shape also appears on rows INSIDE the flyouts,
and those stay live: a drawer holds sketch-only entries too, so disabling the
drawer would hide tools that are perfectly usable. The keyboard shortcuts keep
running the guarded action rather than being gated — a key press has no
greyed-out state to see, so the sentence is the only feedback there is.
Re-evaluated in feed_bodies(), before its viewport early-return since this is
about the toolbar and not the canvas, and once after the toolbar is built: an
empty document is the state the bug was reported in and feed_bodies has not run
yet on a fresh tab.
snaporca-o9j. Reviewed and compiled (RC=0), not exercised.
Two independent leftovers, both in the same status area.
snaporca-752: the "N degrees of freedom" line described a sketch's constraint
state and stayed on screen after Confirm, Cancel and the Escape downgrade, in
Feature mode where it means nothing — visible in every Feature-mode screenshot of
the 2026-07-27 sweep. Cleared in set_ui_mode rather than at those three exits,
because that is the one place all of them pass through and a fourth exit added
later would otherwise reintroduce it. Constrain mode keeps the readout: that is
where the number is the whole point.
snaporca-8cc: moving the status out of the panel and into the viewport HUD
removed the clipping, but not the underlying problem. The chip is a top-level
popup that Fit()s to its text, so a long sentence grew past the right edge of the
canvas and hung over the window instead of being cut off inside it — the same
silent length limit wearing a different hat. The label now wraps to the room
actually available (canvas width minus the view-cube inset), which is what makes
the earlier promise that "a sentence can be a sentence" true at 1366 as well as
at 1920.
SetLabel + Wrap + Fit are now one function called from both the text change and
the placement. Wrap() rewrites the label it is handed, so it has to follow a
fresh SetLabel every time, and the placement path runs on resize — a chip wrapped
for the old width either overhangs a narrowed canvas or wastes a widened one.
The left inset is one constant now because the wrap width and the anchor have to
agree, or the chip wraps to a width it is not then given.
snaporca-752, snaporca-8cc. Reviewed and compiled (RC=0), not exercised.
The .pot, the Italian .po and list.txt are build product: 27,314 of the added
lines in this branch were regenerated catalogues rather than code, and a reviewer
running git diff --shortstat met that number before anything else. Restored to
the merge-base so their diff is zero; they regenerate from source with
scripts/run_gettext.sh whenever the maintainers want them refreshed.
The Romanian catalogue goes with them, for a different reason: it is a complete
new translation and deserves its own PR rather than riding along inside a CAD
feature, where nobody qualified to review it would think to look.
Nothing here changes what the Design tab does. The strings are still marked for
translation in the sources; only the generated catalogues are out.
m_hole_on_face and m_thread_on_face are cleared only by their tool's flyout and by
their plane combobox, so after any on-face hole or thread the flag stays true for
the rest of the session. load_feature_into_dialog restored the stored plane into
the dropdown but never touched the latch, so re-editing from the feature tree
ignored the plane it had just restored: hole_plane() returned the still-latched
face plane, which may belong to a different face, a different body, or a body
since rebuilt. Silent until snaporca-200 added the "On face" row, which then read
as a confidently wrong answer rather than as nothing.
The latch is now rebuilt from the stored feature, which is the only source that
describes THIS hole. Not from the dropdown row: index_from_plane snaps an
arbitrary face plane to the nearest XY/XZ/YZ, so driving the re-edit from the row
would MOVE a hole drilled on a slanted or offset face — that was the reason the
other candidate fix was rejected.
is_base_plane() decides which of the two a stored plane is. It compares the origin
as well as the axes (a plane parallel to XY but 12 mm up snaps to row 0 and would
come back at z=0), and adds modeling_origin before comparing, because hole_plane()
and thread_plane() add it to the dropdown plane before the feature stores it — a
document with a shifted origin would otherwise mistake every dropdown hole for a
face pick. Vector norms, not isApprox, which is relative to magnitude and useless
against the zero origin.
The face's (u,v) extent is not serialized, so m_hole_has_bounds is cleared: the
gizmo's footprint clamp goes unbounded, which is honest, where another face's
bounds are not. The label says which body the face belongs to instead of a face
number the feature does not carry; "(none — uses Hole plane)" is the one thing
that is definitely false there.
snaporca-uif9. Reviewed and compiled (RC=0), not exercised.
refresh_preview() listed Tool::Mate among the features that produce no solid and
cleared the ghost, with a comment saying a mate has no 3D ghost. The kernel never
agreed: preview() routes a Mate candidate through apply_mate on a throwaway copy
of the bodies, and build_candidate already filled the mate fields. That one early
return was the whole of epic gap G3.
A mate makes no NEW geometry but it MOVES a body, and the moved assembly is the
ghost worth showing. Both the Mate card and the offer's mate palette now show it:
hovering a palette row previews that kind, leaving the row drops it, and choosing
one commits. Nothing is written to the document until the click.
The committed bodies are hidden while the ghost is up — it is the whole assembly
in its post-mate pose, not an added lump, so leaving them visible would draw the
mated body twice and z-fight every other body against its own copy. Same reason
Dressup and Draft hide them.
Cleanup is after PopupMenu rather than on a close event: PopupMenu is modal, so by
then the menu is gone and any command it raised has run. A flag distinguishes a
ghost this menu put up from a preview that was already on screen.
snaporca-b4sp. Reviewed and compiled (RC=0), not exercised.
snaporca-lukg part B. The issue describes building a contextual viewport palette
with a stable icon set, non-viable options dimmed and explained rather than
hidden, and edge-aware placement. show_offer_menu() already does all three — its
dead-row branch appends a disabled row with " — " and a reason, and wxMenu
places itself against the screen edge. So this is not a new widget. It is one
section added to that menu, fed by mate_options().
The header row names the pair: "Mate: A → B". That is epic gap G4 — the mate card
is abstract dropdowns and never says which body moves. B is the connector on the
body that MOVES, so B is the arrow's destination; the parameter order invites the
opposite guess, which is why it is commented at the point of use.
Five rows in one loop over the kernel's result, never reordered and never
filtered. The palette addresses rows by position, so a shorter list would move
every row below it — which is the whole argument for dimming instead of hiding.
The pair comes from the Mate card's combos when that card is open, so the offer
and the card cannot disagree about what they are acting on; otherwise the first
two enabled connectors, which is defensible only because the header names them.
An offer acting on an unnamed pair would be worse than no offer.
REVIEW CATCH: the five type names arrived as an array indexed by kind, read as
_L(table[i]). That compiles and is silently untranslatable — _L is a gettext
macro and the extractor scans SOURCE for literals, so five strings would have
shipped that are never in the catalogue. These names appear nowhere else in the
tree, so that would have been their only occurrence. Now a switch of literal
_L() calls.
G3 INVESTIGATED, NOT BUILT, as specified. preview() DOES handle a Mate candidate:
it copies the committed bodies to a temporary and routes the candidate through
apply_mate on that copy, committing nothing, and build_candidate already fills
the mate fields. The only blocker to a hover preview is refresh_preview()'s
Tool::Mate early return, which clears the preview on the belief that a mate has
no ghost. Nothing in the kernel refuses it. Filed rather than built.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-lukg.
snaporca-lukg wants a palette offering all five mate types with the non-viable
ones DIMMED AND EXPLAINED rather than hidden — its reasoning being that a menu
changing shape between invocations destroys the motor memory experts rely on.
That needs an answer this document could not give. This is that answer, and
nothing else: mate_options(cs_a, cs_b) returns five MateOption{kind, viable,
reason}, always five, always in kind order, never filtered.
The geometry test rides on the fingerprint added for snaporca-kqih, which is why
it costs no new serialized field: coordsys_face_kind already records the surface
type. Revolute and Cylindrical need a cylindrical face at both ends because they
need an axis to turn about; Planar needs flat faces; Fastened and Slider
constrain frames rather than surfaces, so no geometry test applies to them.
UNKNOWN IS PERMISSIVE. A fingerprint of -1 means PointWorld or a connector that
has not resolved yet, and it does NOT make a type non-viable. Refusing on missing
information is the false-alarm behaviour that gets a whole feature ignored — the
same reasoning already recorded on kqih for the drift warning, applied again
because it is the same trade.
Reasons name WHICH connector is the problem when only one is. "needs a
cylindrical face at both ends" tells the user what the rule is; "connector A is
on a flat face" tells them where to look, and the second half is the one that
saves the time.
The stability contract has its own test, asserting five entries in kind order
even for a completely invalid pair. That matters more than any individual
verdict: the palette addresses rows by position, so a shorter list would move
every row below it.
Golden fixture unchanged — this is a pure query. Suite 167 -> 171 cases,
2284 -> 2348 assertions, green. snaporca-lukg part A; the palette is part B.
snaporca-rgbj measured the damage: four chamfers on a box remove
0.400/0.397/0.397/0.395 mm3 when each id is re-read, and
0.400/0.008/0.397/0.280 when the four ids are captured up front. The kernel is
right in both runs — the second one asks for the wrong edges. Neither errors,
because a stale id still resolves to a real edge, just not the one that was
measured.
That makes it an API problem rather than a script bug. Reading the scene once and
then issuing several operations is the natural way to drive a socket, it is what
every agent will write, and it produced silently wrong geometry with nothing
anywhere reporting it.
CadDocument::topo_generation is bumped where the bodies are replaced — the single
line in recompute() where the face and edge maps actually change, so a feature
type added later cannot forget to bump it, which a per-mutator counter would
invite. describe_scene and query_topology return it. A caller may pass it back as
"generation" on any call, and a mismatch is refused with a message that says what
to do about it.
Two deliberate choices:
OPTIONAL, not mandatory. Every existing script keeps working unchanged; passing
the generation is what buys the guarantee. Making it required would break every
caller to fix a mistake only some of them make.
CHECKED AT THE DISPATCHER, not in each handler. One site covers fillet, chamfer,
shell, draft, coordsys, thicken, cut, project, delete_face and everything added
after them. A per-handler check is a list that goes stale the first time someone
adds a method in a hurry.
Not serialized: an id means something only within the run that produced it, so
persisting the counter would promise a stability the ids themselves do not have.
No recipe version change.
describe_tools now carries an id_lifetime note, because the guard only helps a
caller who knows to ask for it.
Kernel suite 167 cases / 2277 assertions green; libslic3r_gui builds. The guard
itself is NOT exercised — it needs the socket, so it is on snaporca-bdco.
snaporca-o1l2.
kqih option (c). A FaceAndDirection connector stores a global face index, and an
upstream edit can renumber faces so the index silently names a different one. The
DANGLING case already threw; this is the in-range-but-wrong case, which nothing
detected.
Fingerprint the face on first resolve, compare afterwards, and report a mismatch
into mate_conflicts — the channel that already marks the tree row — never as an
error. A drift warning must not abort the recompute, because the alternative
makes a legitimate Draft on a mated face fatal.
WHAT THE FINGERPRINT IS, AND WHAT IT IS NOT. Surface type plus edge count. Not
centroid or area: legitimate parametric edits move and resize faces, which is the
entire point of the model, so either would fire on every dimension change. Not
the normal, which is the tempting one — Draft deliberately tilts a face and
Transform reorients a body, both legitimate. Type and edge count survive rigid
motion, tilting and resizing, and catch the case that actually happens: a planar
index sliding onto a fillet's cylindrical face after a dress-up inserts faces.
The accepted cost is that a slide between two planar 4-edge faces is invisible. A
partial detector that never cries wolf beats a total one that does, because a
false alarm on a valid connector teaches people to ignore the warning.
Connectors with no fingerprint record one on first recompute, so old recipes
self-heal and both writers (DesignPanel, McpControl) get it without changing.
THE VERSION BUMP IS THE IMPORTANT HALF. The task was specified with "do not
change the recipe version" — that was wrong, and the rule is written in the
header three lines above the constant: bump whenever save/load gains a field.
deserialize_recipe() gates on v == VERSION and then reads a FLAT symmetric field
list. A v3 blob under a v3 build that has grown two fields passes the gate and
reads two ints past the end of every connector, into the next feature's bytes.
That is silent corruption of a saved project, which is worse than any load error.
Now v4, and v3 gets the existing clean refusal.
cad_recipe_v3.bin is KEPT, unregenerated, with a test asserting it is refused and
that nothing half-read is left behind. It is the only artefact that can prove the
gate works, because it was written by an older build — regenerating it with
today's code would destroy the evidence, which the test says in as many words.
Suite 163 -> 167 cases, 2248 -> 2277 assertions, green. Fixture v4 34928 bytes.
snaporca-kqih.
Two tests that separate a hypothesis nobody had tested. The socket showed four
chamfers on a filleted rim removing 29.6 / 20.0 / 10.3 / 7.5 mm3, falling
steadily. That could be the chamfer maths degenerating on a filleted rim, or it
could be how the driver captured its edge ids. Those have completely different
fixes, so the first job was to find out which.
dressup_edge is a global index into TopExp::MapShapes(shape, TopAbs_EDGE),
resolved against the body AS IT STANDS at that feature's position, and every
dress-up rewrites that map. So the two usage patterns are:
ids re-read after each chamfer: 0.400, 0.397, 0.397, 0.395 mm3 (max/min 1.01)
four ids captured up-front: 0.400, 0.008, 0.397, 0.280 mm3 (max/min ~48)
The kernel chamfers uniformly when handed a fresh id. It degrades only when
handed ids snapshot against an earlier shape — and the second chamfer's stale id
landed on a nearly-consumed edge and cut two percent of what was asked. That is
the accumulating-drift signature the socket showed.
Conclusion: driver artefact. apply_chamfer and OCCT are not at fault.
The part that makes this worth a test rather than a note: IT DOES NOT THROW.
ok=1, error empty. A stale id still resolves to a valid edge — just the wrong
one — so nothing anywhere reports it. Silent wrong geometry, which is the class
this project does not tolerate, reachable by any caller that reads the scene once
and then issues several dress-ups.
Test 2 asserts the non-uniformity as CURRENT BEHAVIOUR and says so in the code:
it documents a defect, it does not bless one. When the driver contract is fixed
it should be rewritten, not deleted.
Suite 161 -> 163 cases, 2217 -> 2248 assertions, green. Tests only, no
production code. snaporca-rgbj.
These were the last tools from the charter audit with no handle at all, and the
issue filed against them offered three options, all of which changed the tool.
Reading on_add_surface_offset() dissolved the question instead.
The premise was that a distance handle needs a frame, a sheet body has no single
normal, and therefore the tool must start demanding a face. But the face was
never needed for the OPERATION — only for the ARROW. Both tools still offset or
thicken the entire sheet named in the combo. The picked face only says where to
stand the handle.
So the arrow appears whenever a face of that sheet is under selection, and its
absence costs nothing: the card alone works exactly as before. Purely additive —
no existing flow changes, and there is no new precondition for the user to learn.
That is strictly better than any of (a) anchor on the first face and be wrong on
a curved sheet, (b) sample a normal at the bbox centre and be arbitrary on a
folded one, or (c) require a face pick and change what the tool demands.
The arrow is refused when the picked face belongs to a DIFFERENT body than the
sheet in the combo. An arrow standing on one body while the tool acts on another
would name the wrong thing, which is worse than no arrow.
ThickenSurface was not on the audit's list — it is a distinct tool from Thicken,
with its own card and its own sheet-body combo, and it has exactly the same
shape. Fixing one and not the other would have left the same gap under a
different name.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-9fel.
detect_mate_conflicts() has been filling m_doc.mate_conflicts on every recompute
since the kernel half landed, and nothing read it. The diagnostics existed and
were invisible — a conflicting assembly looked exactly like a working one.
The tree row is where they go, because the tree is where the user is already
looking for which feature to change. Three states, in precedence order:
disabled -> dim. A SUPPRESSED mate is the user's answer to a conflict, so it
must read as suppressed rather than keep shouting about it.
conflict -> warn.
otherwise -> normal.
Selecting a marked row puts the reason on the status line — "Mate3 already
positions Body 2", the cycle, the self-mate — and names the eye as the way to
suppress it. A message that describes a problem with no action is a message that
gets ignored; the action here is already one click away on the row just selected.
Deliberately NOT a modal, and deliberately not treated as a document error. The
document still evaluates with a conflict present: the mate graph merely has more
than one answer for a body, and which one wins is the thing the user needs to
see. Blocking the loop to say so would interrupt without helping.
The dimming of non-involved bodies from the original UX proposal is still not
implemented, on purpose: under transform composition a failure mid-chain
propagates, so "not involved" is not a well-defined set, and dimming the wrong
bodies would hide the context needed to understand the conflict.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-bioq.
Two defects found by auditing the body-focus x-ray path, which shipped compiled
but never exercised. Neither is reachable from the happy path its test plan
walks, which is why compiling it proved nothing.
1. A STALE FOCUS KILLED THE VIEWPORT. The focus is a body INDEX held by the panel
across recomputes, so it outlives the body it names: delete a body and the
stored index can point past the end. body_pickable() then rejected EVERY body,
because none of them equals an index that no longer exists — a viewport that
silently accepts no clicks at all, with nothing on screen saying why. Out of
range now means no restriction. Fail open, never dead.
2. THE COMBO AND THE FOCUS COULD DISAGREE. refresh_cs_body_choice() rebuilds the
Body combo and, when the body list shrank, silently reset the selection to
"(all)" — while the viewport stayed focused on the old index. Every other body
kept its 25% alpha and picking stayed restricted to a body that might be gone.
That is the exact mirror of the open_tool ordering bug this feature already
fixed once: that one showed "Body N" over an opaque scene, this one shows
"(all)" over a dimmed one. They are one state and are now written together.
Guarded on CoordSys being the active tool, since it is the only card that owns
this focus. In the edit path the function runs BEFORE open_tool with the
previous tool still active, so the guard is false and the caller's explicit
set_xray_focus still wins.
Also confirmed while reading, since the header asserts it: set_solid_pick() does
NOT touch m_pick_only_body, so the focus really does survive the mesh feed. That
claim now has a check behind it rather than a comment.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-bgvk.
Two open findings from the first rig judgement of the connector glyph.
F4 — the quadrant collapses to a blob at grazing angles, which is exactly when
the roll is hardest to read. Adds a radial tick along +X extending past the disc
rim. As the disc flattens to a line the sector loses all its area, but a radial
spoke keeps its length and its direction along the one axis that still projects.
The alternative on the issue was to billboard the quadrant while the disc stayed
in-plane. Rejected, and not on taste: at true grazing the view direction lies IN
the connector's plane, so every in-plane direction projects onto the same screen
line and the roll is geometrically unrecoverable. Billboarding would not recover
it — it would face the camera and read as a definite orientation that is not the
frame's. Degrading to a direction that can still be trusted beats drawing a
confident lie. The tick is additive, so unlike billboarding it cannot make the
non-grazing case worse; it still wants judging on the rig at a true grazing view
before F4 is called closed.
F5 — roll-undefined was a loud red: the strongest colour in the viewport spent on
the least important connector, pulling the eye off the mate being made. It marks
"this one could not be derived", not an error. Muted amber says look-here without
shouting.
No tick is drawn when the roll is undefined — a tick there would assert a
direction that does not exist, which is the silent guess the hatched quadrant
exists to avoid.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-wgsc.
Rib's depth already reused the Extrude arrow. Its thickness could not: the arrow
points along the plane normal, and thickness is an offset either side of the rib
line, IN the plane. Different direction, different handle.
Two square handles at mid ± perp·half, plus the slab's actual footprint drawn as
a thin closed rectangle — the footprint matters more than the dots, because what
a rib thickness means is how wide that slab lands on the body, and until now
there was no way to see it before committing.
A drag on either handle sets the FULL thickness, twice the perpendicular distance
from the line, because the slab is centred on the line and the handle sits at
half. Both handles behave identically for the same reason, so they share one
colour rather than pretending to be two different actions.
A zero-length line has no direction to grow a slab perpendicular to, so the
shared rib_frame() helper returns false and render and drag both draw nothing
rather than dividing by zero. Non-Line entities clear the gizmo instead of
guessing: the kernel is line-only and a gizmo that guesses would be lying about
what Confirm will build.
Unlike the helix callback this one goes through refresh_preview(), because Rib
builds a real solid ghost that has to rebuild. The helix has none and skips it
deliberately.
Both gizmos coexist and resolve the sketch and entity the same way, so the depth
arrow and the thickness handles can never disagree about which line they are on.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-plew.
Driving the control socket: hexagon prism, six vertical fillets, four chamfers
on the already-filleted rim, an M8 hole. Afterwards describe_scene reported
bodies=3 and error='' — entirely healthy — while body 2's TopoDS_Shape was null.
Only mass_properties on that one body revealed anything was wrong.
So a feature destroyed a body, recompute() returned true, and the document went
on advertising it. Any downstream consumer — slicing, STEP export, a mass
properties report — met a null shape with no warning. That is the silent
corruption class, which is the one class this project does not tolerate.
recompute() now scans the freshly built bodies for a null shape, names the body
and the feature that destroyed it, and returns false. Returning false rather than
just setting error is the point: it hands the caller its normal rollback path, so
the operation that destroyed the body is undone instead of committed.
The message says "an unidentified feature" when source_feature is -1. "feature 0"
would be a lie, and a message that exists to tell you where to look has to be
trusted.
TEST IS A POSITIVE CONTRACT, AND THE REASON MATTERS. The reported order was
driven headlessly first, as the better test: it does NOT reproduce. The dress-up
step throws "fillet radius too large", which is an already-loud already-caught
path, so recompute fails honestly and never nulls a body. No public-API sequence
found so far reaches the guard's branch without a GUI, and faking a null into
`bodies` after the fact would not exercise it — the guard runs on `built`, before
the swap. So the test asserts what can be asserted: a box + fillet recomputes
true, error is empty, and no body is null. The guard's own branch is defensive
and currently unexercised; that is stated here rather than implied by a green
suite.
Kernel suite: 2217 assertions in 161 test cases, all passing. No existing test
relied on a null body surviving a recompute, so hardening this broke nothing.
snaporca-5425 (part a). Part b — why the chamfer chain degenerates on an
already-filleted rim — is untouched and stays open.
grep -i helix over the viewport code returned nothing at all. The tool was four
coupled numbers and a Confirm button — you typed radius, pitch, height and taper
blind and pressed OK to find out what you had made. So this is not only the
charter's L2 failure; the tool had no visible state whatsoever while it was open.
Adds a plane-anchored helix gizmo built on the datum-plane gizmo as its template,
being the closest existing thing: also plane-anchored, also driven by a card while
the sketch tool is inactive, also a render / hit-test / drag triad.
It draws the live curve and the axis, and puts a handle on each of the three
lengths: radius on the base circle, height at the top of the axis, pitch at the
end of the first turn — which is exactly where one pitch of rise lands, so the
handle means what it is standing on. Below one full turn the pitch handle moves
to the end of the curve rather than floating off a curve that does not exist yet.
Taper and handedness stay on the card. One is a shape modifier and the other a
flag; L2 governs numbers you can point at.
A drag reports the whole (radius, pitch, height) triple rather than one value,
because pitch and height are coupled through the turn count and writing one alone
would redraw a stale curve. The callback re-feeds the gizmo directly instead of
going through refresh_preview(), since Helix takes the produces-no-solid early
return and refresh_preview would rewrite the status line on every mouse move.
REVIEW CATCH, fixed here: the first cut read taper as a fraction of the radius
consumed over the turn count. It is an ANGLE IN DEGREES — helix_spine() builds a
Geom_ConicalSurface of half-angle taper and takes the top radius as R+H*tan(taper),
growing with the height risen. The wrong reading drew a preview that collapsed to
a point for any non-zero taper while the committed feature was perfectly fine. A
preview that lies is worse than no preview, which is what this commit replaced.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-i3jc.
Rib's depth is a distance along the sketch plane normal, so it is the Extrude
arrow for the fourth time — anchored at the midpoint of the line the rib is
built on, because a rib's line IS its profile.
This is half of Rib's L2 failure. The thickness is an in-plane offset either
side of that line and no existing gizmo draws that; it needs a handle that does
not exist yet, filed as snaporca-plew rather than left implied. One of two
numbers draggable is strictly better than neither, and saying which half is
missing is the point.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-i3jc.
Both tools produce exactly one number — a distance along a known normal — and
neither had a handle for it. That is the same shape as the Extrude depth arrow,
which was already written, already draggable and already had an editable label
on the geometry. So this adds no gizmo: it points the existing one at two more
tools.
SurfaceExtrude anchors on its sketch's plane, at the profile centroid.
Thicken anchors on the picked face, and reuses the face-as-profile recipe from
the Extrude path verbatim — including the two things that path learned the hard
way: look the face up on its OWNER body rather than the whole-document compound,
and carry that body's display Move transform onto both the origin and the
normal, or the arrow draws on the bed instead of on the face.
The drag callback routes by active tool. `second` stays Extrude's alone: it is
the two-sided pair, and the other two have a single distance each.
SurfaceOffset is the third tool in this group and is deliberately NOT here. Its
target is an arbitrary sheet body, which has no single normal to anchor an arrow
on — that is a design decision, not typing, and it stays on the audit.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-i3jc.
The Placement > Transform verb opened a card of spin controls — dx/dy/dz, an
axis combo, an angle — with nothing on the geometry. The 3-axis drag gizmo the
charter asks for already existed and was fully implemented (arrows, rotation
rings, click-to-type per axis), reachable only from a small icon button in the
tree card header. The prominent verb opened the form; the geometry-first
control was hidden behind an icon. That was backwards.
Transform now arms that same gizmo on the target body. The card stays as L2's
typed half: the drag writes dx/dy/dz, the axis and the angle, and the pivot is
seeded from the body's centroid so the parametric feature reproduces exactly
what was dragged.
Decomposition is exact for the interaction that matters — the gizmo's rings are
per-world-axis, so a ring drag is an axial rotation. A pose composed from two
rings is not axial and the card can only name one axis, so it reports the
dominant one rather than refusing to answer.
Three things this had to get right:
- The gizmo bakes its drag into the display transform so the body follows the
cursor, and the feature performs the same motion parametrically. Committing
without reverting first would move the body twice.
- tool_confirm() and tool_cancel() both tested moving_body() BEFORE the active
tool, so with the gizmo armed Confirm would have dropped the gizmo and never
created the feature. Both are now guarded on Tool::None.
- close_tool() is the single revert point. Esc, Cancel and switching tools all
pass through it, so a Transform that was never committed cannot leave the body
displaced.
Edit mode is untouched: re-seeding the gizmo from a stored feature is a separate
problem, so editing an existing Transform still gets the card alone.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised. snaporca-qtf4.
Three defects behind one report ("bodies cannot be moved or hidden/shown or
deleted, colour does not work"). They are unrelated to each other; only the
symptom was shared.
1. The Color tool wrote a per-body override that was correct end to end —
stored on CadBody, carried across recompute (CadDocument.cpp:3381), read
back by DesignCanvas::body_color() — and then overpainted every frame.
m_body_selected is a DOCUMENT-WIDE flag raised whenever a non-Sketch
feature row is selected, which is the resting state after any modelling
operation, and while it was true every body rendered gold. An explicit
colour now outranks the selection tint; unpainted bodies still tint, which
is all the tint was ever for.
2. The eye toggle re-selected the body row through m_tree, using item ids that
belong to m_parts. The row came back unselected, so the second press found
tree_body_selection() == -1 and fell through to the feature-level branch
instead of un-hiding. Hide worked exactly once. The sibling call in
refresh_parts() had it right.
3. The tree card's Delete button answered a selected body row with "select the
FEATURE that created this body" — an instruction the user cannot act on,
because the tree does not say which feature that is. on_delete_body()
already resolves CadBody::source_feature and confirms by name; it was
reachable only from the right-click offer. The button now routes to it.
Reviewed and compiled (libslic3r_gui, RC=0); not exercised — needs a session at
the machine to confirm all three in the viewport. snaporca-zjvg.
DesignCanvas::set_view() and fit_view() were both written and then never called
from anywhere in the tree. The Design viewport has had no way back to a standard
view since it existed: no key, no button, nothing but orbiting by hand until the
model happens to drift into frame.
That is worse than a missing convenience. A camera left pointing along the bed
plane renders a scene that looks exactly like a failed renderer — geometry
present, nothing visible — and an hour went into blaming the software GL stack
before the real cause turned out to be two uncalled functions.
Home rather than a letter: every letter A-Z is already a Shift+letter tool
shortcut. Home is also the reset-the-view key most users arrive with. The
dispatcher needed no change, it keys on the raw wx keycode. set_view() already
does select_view + zoom_to_volumes, so this is fit and orient in one call.
Doc row added to the View toggles table in docs/design_tab.md.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
tests/libslic3r/CMakeLists.txt added only test_caddocument.cpp under
SLIC3R_CAD. The other five shipped in the tree and were never compiled, so
49 TEST_CASE blocks looked like coverage and were not: sketch constraints,
sketch editing, sketch import, inference, and the libslvs constraint set.
They also still targeted Catch2 v2 — mainline is on v3, where the umbrella
header is catch2/catch_all.hpp and Approx lives in the Catch namespace rather
than at global scope. Both fixed; nothing else in the files changed.
Found by building the tree rather than reading it. Suite goes from 374 to 423
test cases, 54,424 to 54,620 assertions, all passing.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Two enabled mates driving the same body is not an error today — the later one
just wins, and the earlier mate looks ignored with nothing said. A cycle in the
mate graph is worse: composition still produces a result, but an arbitrary,
order-dependent one.
recompute() now fills a mate_conflicts vector of (feature index, reason) before
the geometry pass, so it survives a throw further down. It catches a second mate
on the same target body, a mate positioning a body against itself, and a cycle,
via an iterative three-colour DFS over the body graph. Broken mates are skipped
silently — apply_mate() already errors on those.
Deliberately non-fatal: recompute() still returns true and error stays empty.
Deliberately not "over-constraint" — that word promises DOF analysis from a
solver this kernel does not have.
Port of snaporca ec4ffeb979. Kernel half of snaporca-bioq.
Suite: 2213 assertions / 160 cases green on this fork too.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
In an assembly the face you want for a mate connector is nearly always behind
another body, and a click only ever returns the frontmost hit. Hiding the
occluder from the Parts list works but means leaving the tool mid-pick.
The CoordSys card now carries a Body chooser. Pick a body and every other one
drops to 0.25 alpha AND stops catching clicks, so the wanted face is both
visible and reachable in one gesture. "(all)" restores normal picking.
Deliberately NOT hit cycling: repeated-click cycling was removed from solid
picking as a charter L5/§10 violation (DesignSketchTool.cpp, "NO CYCLE"), and
re-introducing it here would make "click a face" a multi-click gesture again.
Mechanics: DesignSketchTool::set_pick_only_body() gates body_pickable(), which
every pick path already consults; DesignCanvas::set_xray_focus() drives both it
and the per-body alpha in reload(). The chooser stays a pick FILTER only --
coordsys_body still comes from the actual pick, so nothing in the kernel moves.
Body focus follows the CoordSys card: open_tool() reads it back from the combo
rather than clearing outright, because editing a CoordSys feature loads the card
(and its body) before open_tool runs.
snaporca-bgvk. NOT COMPILED: deps/build lacks OpenVDB so the GUI tree will not
configure here; reviewed by diff only.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
mate_cs_a/mate_cs_b are feature indices. remove_feature() remapped sketch_ref
through the deletion but not the mate connectors, and move_feature() swapped
sketch_ref but not the mate connectors. Deleting or reordering any feature
ahead of a connector slid both references onto whatever features landed on
those slots.
Nothing reported it. recompute() only rejects out-of-range and non-CoordSys
targets, and a shifted index normally lands on the assembly's other CoordSys —
an assembly carries at least two by construction. So the mate resolved against
the wrong frames and moved the wrong body, silently.
Extracted a remap lambda in remove_feature() and a swap_ref lambda in
move_feature(), applied to sketch_ref and both mate connectors.
Two tests, both confirmed red before the fix. [mate] tags green here:
395 assertions / 29 cases — the first end-to-end kernel compile of this fork.
Ported from snaporca; CadDocument.cpp is byte-identical across forks again.
Refs: snaporca-kqih
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Ghost GLVolumes baked their filament color once in calc_imex_ghosts and
only rebuilt when build_imex_ghost_cache_key changed. The key covers the
mesh-shaping inputs (mode topology, pem, object set, per-plate head map)
but not filament_colour, and no invalidation hook fires on filament
preset or color changes — so a ghost baked under a transient palette
(late-loading project colors, a subsequently edited filament color)
kept the stale color forever. A failed lookup at bake time baked
GLVolume::UNPRINTABLE_COLOR, which renders as a jet-black ghost while
the hover tooltip — which re-resolves live — shows the correct color.
Re-stamp each ghost's color in _render_imex_ghosts from
get_imex_head_filament_color, the same resolution the tooltip runs, so
the two can never disagree. Resolution is hoisted into a per-head map
so instances sharing a head resolve once per frame; the alpha baked by
calc_imex_ghosts is preserved. The render loop already stamps color per
ghost per frame (set_render_color + model.set_color), so the added cost
is a couple of map lookups. The cache key stays scoped to what it
actually protects: the expensive mesh bake and transforms.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
# Description
During the UI/UX improvements about a month ago, I got the transforms
wrong, and slicing at anything other than a 45 degree angle was
affected.
Validated on a baby belt pro at 30 & 45 degrees.
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
set_extruder skips the ooze-prevention standby cool-down when the outgoing and
incoming filament route to the same physical extruder. That is an IMEX behaviour --
an AFC/MMU lane swap keeps the same heater selected -- but the check was not gated,
so it ran on every printer.
73 shipping profiles enable ooze_prevention by default (37 Snapmaker, 21 WonderMaker,
9 Flashforge, plus Lulzbot, Prusa, re3D, iQ and the MyToolChanger), and the 150
multi-nozzle machines behind them author no physical_extruder_map. They were spared
only because every one declares a single variant per extruder, so the map came out as
the identity and nothing was ever suppressed. Correctness should not rest on that.
Gate on is_imex. Verified on one fixture with only is_imex differing, with two
filaments mapped to the same physical extruder: 0 cool-downs emitted with it on, 26
with it off.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
physical_extruder_map has one entry per logical extruder -- the index space of
nozzle_diameter -- and its consumers size their own arrays from that count. It was
being derived from printer_extruder_id, which is indexed by variant slot: one entry
per extruder+variant pair. An X1 Carbon has one nozzle and printer_extruder_id
{1,1}; an H2D 0.4 has two nozzles and {1,1,2,2,2}. The two spaces coincide only
when every extruder declares a single variant.
The visible effect was on the standby cool-down. set_extruder skips it when the
outgoing and incoming filaments share a physical extruder, and that check is not
gated on IMEX. With the map built from the wrong array, two filaments on a
dual-nozzle machine read as sharing one hotend and the cool-down was dropped --
caught by "Toolchange temperature commands are unchanged when the wipe tower wait
is off", which failed on all five CI platforms with the ;cooldown line missing.
Derive the identity over the nozzle count instead, the same fallback Plater.cpp
already applies where a profile authors no map. A profile counts as authoring one
only when its length matches the nozzle count, so the single-element PrintConfig
default is replaced rather than read as a one-extruder machine. Authored maps pass
through untouched, including the {1,0} numbering permutation the BBL dual-nozzle
profiles ship.
Tests pin the four branches and the length invariant the consumers depend on.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
## Summary
Fixes for the `belt-printer` branch ahead of upstreaming, in two groups
(10 commits). Targets `belt-printer` (not `main`) since group 2 fixes
the not-yet-merged Belt Printer Brims feature.
Every fix keeps non-belt (and brim-disabled) output unchanged; belt-only
behavior is corrected. All changed translation units and the two test
files were type-checked (`-fsyntax-only`); a full build + `ctest` still
needs to run in an environment with current deps.
## Group 1 — pre-existing belt-printer regressions
- **[HIGH] BuildVolume belt state not reset when leaving belt mode** —
toggling belt off (or switching belt→normal with matching bed geometry)
left the `BuildVolume` with `m_is_belt_printer=true` and inflated Y
bounds, so out-of-bounds objects were treated as printable on a normal
printer.
- **[HIGH] `GCodeProcessorResult::reset()` didn't clear belt fields**
(`belt_tilt_angle`, `belt_z_origin`, `preslice_remap_*`) — a reused
result corrupted a normal print's start-gcode preview Z.
- **[LOW-MED] `TreeSupport::drop_nodes`** — restored the single critical
section around node invalidation (the two `valid=false` writes had been
moved outside the mutex on the shared tree-support path); removed an
unused local.
- **[LOW] Support overhang hot paths** — avoid unconditional lower-layer
polygon copies when there is no build-plate tilt (`SupportMaterial`,
`TreeSupport3D`); untilted output matches upstream exactly.
- **[LOW] Render loop** — hoisted the frame-invariant slope
`up_direction`/`normal_z` (and their per-volume config lookup) out of
the per-volume loop.
- **[LOW] FDM-support "select by angle"** — restored the exact upstream
threshold when the build plate is untilted (the generalized form
differed for non-uniformly-scaled objects); tilted-gravity form kept
only under tilt.
- **[LOW] Printer tab tilt sync** — only clears the belt-derived
`build_plate_tilt` on a genuine in-place belt→off toggle (tracked,
seeded on preset load), no longer wiping a manually-set tilt.
- **[LOW / opt-in] Axis-remap G-code emission** — always emit full XYZ
under an active `gcode_remap_*`, apply the remap on all base
`travel_to_xyz` destinations, fall back to a linear lift for spiral/arc
under remap, sync `set_axis_remap` each export; fixed belt first-layer
travel speed. Identity/default output unchanged.
## Group 2 — Belt Printer Brims (#15155) fixes
- **[CRITICAL] Dropped brim at first belt contact** — a coincident brim
band on an object layer with no extrusion pass (zero-extrusion leading
slice, or belt support below the Z=0 floor with no coinciding object
extrusion) was never emitted. Now each coincident band's brim filament
is registered in `ToolOrdering`, each band is emitted exactly once in
its brim-filament pass, and an end-of-layer orphan sweep emits any band
whose object layer produced no visit.
- **[Multi-extruder] Wrong tool / double emission** — apron and
coincident bands now print once, in the correct brim-filament pass,
brim-first (were previously emitted with the active tool and could
double-emit per filament plan). Single-extruder / single-object output
is byte-identical apart from the previously-dropped bands now printing.
- **Inner-only predicate** — `has_belt_brim()` no longer reports a brim
(and no longer rejects the prime tower / spiral vase) for `inner_only` +
`brim_width=0` + leading/extra > 0, which produces no inner geometry;
mirrored in `wants_brim`.
- **ToolOrdering raft-gap comment** — clarified why raft-gap synthesis
is suppressed for all belt printers (belt has no rafts;
sub-object-bottom layers are apron / belt-support-below-floor /
lead-in). No behavior change.
- **Tests** — deterministic coverage: brim present at first belt contact
(support on/off), brim-before-perimeters once (no drop/double), single-
and multi-extruder tool selection with no doubling, multi-object
per-filament ordering, inner-only+leading-only not rejecting prime
tower/spiral, and inner-ring / leading-edge-only geometry units.
## Testing
- `-fsyntax-only` passes for all 16 changed source TUs + 2 test TUs
against this branch.
- Please run the full build and `ctest -R 'SkirtBrim|BeltBrim'` before
merging.
## Known follow-up (out of scope)
`extrude_arc_to_xy` does not remap its I-J center, so arc-fitted
*extrusions* under standalone axis-remap would be geometrically wrong —
a separate fix if that combination is supported.
Opened as **draft**.
Deterministic tests for: coincident brim at first belt contact not dropped
(C), single- and multi-extruder brim tool selection with no doubling (B),
multi-object apron ordering, inner-only+leading-only not rejecting prime
tower/spiral (D), and inner/holed + leading-edge-only geometry.
- Emit coincident belt_brim_by_layer bands even when the leading object layer
has no InstanceVisit (zero-extrusion lead-in / no coinciding support), so the
brim at first belt contact is no longer dropped.
- Register each coincident band's brim filament in ToolOrdering and emit each
band exactly once, in its brim-filament pass; emit ordinary-layer aprons in
the brim pass before object extrusion (correct tool, brim-first) instead of
with whatever tool was active.
- has_belt_brim(): inner-only brims need brim_width>0 (leading/extra produce no
inner geometry), fixing spurious prime-tower/spiral rejection; mirror in
wants_brim. Single-extruder/single-object output is unchanged except
previously-dropped bands now print.
- BeltGCodeWriter::travel_to_xyz final branch used config.travel_speed
instead of the computed first-layer-aware travel_speed.
- extrude_to_xyz decided emit_xyz vs emit_xy from pre-remap Z; emit full
XYZ whenever an axis remap is active so remapped machine-Z is never
dropped.
- base travel_to_xyz now applies apply_axis_remap() on all emitted
destinations (standalone remap on non-belt printers was unremapped).
- spiral/arc travels fall back to normal linear lift under active remap
(endpoint-only remap can't preserve arc plane/I-J).
- set_axis_remap() is now synced unconditionally each export to avoid a
reused writer retaining a stale non-identity mapping.
update_fff() zeroed any build_plate_tilt matching the dormant belt-derived
tilt (default X/45) within 0.01, wiping a legitimate manual tilt on a
non-belt tilted-bed printer. Track the belt->non-belt transition and the
exact values belt-sync wrote, clearing only those on an in-place toggle;
reset tracking on preset load so preset switches never wipe tilt.
select_facets_by_angle replaced upstream's limit.dot(down) threshold with
cos(threshold), changing facet selection for non-uniformly-scaled/mirror
objects on ALL printers. Restore the exact upstream computation when no
build-plate tilt is active; keep the tilted-gravity form only under tilt.
Belt slope-shading changes recomputed up_direction (with a printer-preset
config lookup) and normal_z per volume; both are frame-invariant. Compute
once before the to_render loop and reuse the already-hoisted
support_normal_z. Uniforms are still set per volume; visuals unchanged.
SupportMaterial::detect_overhangs copied lower_layer_polygons per region
even without build-plate tilt; hoist the tilted copy out of the region
loop and use the original polygons directly when untilted. TreeSupport3D
flattened lslices_extrudable to Polygons unconditionally; restore the
upstream ExPolygons offset on the untilted path.
The 2-node merge moved the two valid=false writes outside the mutex that
upstream held together with the contact_nodes push_back; restore a single
critical section per branch (belt branch also guards to_buildplate).
Remove an unused top_interface_layers local in drop_nodes.
reset() cleared the sibling machine_frame_transform_active but not
belt_tilt_angle/belt_z_origin/preslice_remap_*; a reused result carried
stale belt metadata into a subsequent normal print, flipping the store_z
branch and corrupting start-gcode preview Z for non-belt prints.
Non-belt branch of set_bed_shape reset only the 3DBed renderer, not the
BuildVolume; Bed3D::set_shape early-returns on unchanged bed, so a
belt->normal switch or in-place belt toggle-off left the BuildVolume with
m_is_belt_printer=true and inflated Y bounds -> out-of-bounds objects
treated as printable on a normal printer.
# Description
This adds brim support to belt printers.
Added a new belt printer specific mode, Leading Edge Only and two new
belt-specific parameters, Leading Edge Brim Length, which increases the
number of brim lines on the side of the part printed first, and Extra
Brim Width, which increases the width of brims along the X axis. Because
belt printer first layers are effectively a single line, getting them to
stick properly can be a pain. This PR aims to help alleviate that, or at
least give more options for control.
<img width="1849" height="1043" alt="Screenshot from 2026-08-06
12-20-21"
src="https://github.com/user-attachments/assets/f963ed8e-53e7-48f8-a495-123cb9ae27f7"
/>
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
"Leading edge only" describes where a part meets a moving belt, so it has no
meaning on a fixed bed and should not clutter the Brim type dropdown there.
Filtered the same way support_style and wipe_tower_wall_type already are a few
lines above in TabPrint::toggle_options(): the field holds its own copy of the
option definition, and Choice maps the combobox selection straight onto that
copy's enum_values, so rewriting the values, the labels and the combobox items
together keeps the mapping correct.
The entry is kept when it is the current value, so opening a project that uses
it on a non-belt printer cannot leave the control displaying an option it does
not offer - which would silently rewrite the setting on the next edit.
Print::validate() already warns that it prints as an ordinary outer brim there.
Matches the scope of the existing precedents: the per-object override panel is
not filtered.
Six issues found by reviewing the previous commit against belt-printer, two of
them release-blocking.
Data race (high). Print::process() runs generate_support_material() for all
objects in a tbb::parallel_for, and make_belt_brim() runs at its tail, but
belt_brim_obstacles() read every OTHER object's support_layers() - which a
concurrent task may be inside clear_support_layers() deleting. That is a
use-after-free, and even when it survives, the obstacle set depends on which
object finishes first. Only this object's own supports are consulted now; they
are complete at that point. Foreign objects still contribute their slices,
which are finished and immutable before the support phase.
Apron bands dropped (high), two separate causes. An apron band prints below
its own object's first layer, but another object can already be printing at
that print_z, in which case process_layer() takes the ordinary path and never
emitted the band - the emission is now shared by both paths. Separately, a
band whose print_z matched a support layer of the SAME object was overwritten
in the print-wide merge, which keeps one record per object per z and could not
detect the collision because LayerToPrint::layer() is null for a band. The
per-object pairing loop is now a three-way merge over object, support and apron
streams, so each object contributes at most one record per z.
Multi-instance was far too strict (medium). It refused belt brim for every
multi-instance object, killing plain brim width and inner brim too, and only
warned when a leading length was set. Only movement ALONG the belt changes an
instance's belt-floor Z, so copies side by side ACROSS the belt share one set of
bands perfectly well; belt_brim_instances_compatible() now tests just that, and
the warning fires whenever the brim is actually suppressed.
Apron layer bookkeeping (medium). Apron layers count toward m_layer_count and
advance m_layer_index, but emitted no Z/height tags, left m_last_layer_z,
m_max_layer_z and m_last_height stale - so the first object layer computed its
height against a pre-apron Z - and skipped before_layer_change_gcode and
layer_change_gcode entirely. All of that now matches the ordinary path.
Obstacle cost (low). belt_brim_obstacles() ran a full-plate union per band.
A bounding-box pre-filter drops non-overlapping objects before materialising any
polygon, and the union is skipped for trivial inputs.
Deliberately unchanged: every apron band still reports cooling layer_id 0.
CoolingBuffer uses it for the initial_layer_fan_speed override and the
close_fan_the_first_x_layers gate, and every band lies on the belt plane itself,
so it is all first-layer material by the only definition that means anything on
a belt. Numbering the bands would ramp the fan up while still printing on the
belt. Now documented at the assignment rather than left implicit.
A belt printer slices in a rotated frame, so the belt surface is a tilted
plane rather than the Z=0 bed plane. Each slicing layer touches the belt
only along a narrow strip at its leading edge - about 0.2mm at 45 degrees -
so a part's first layer is really a first line, with almost no contact patch
to hold it down while the belt drags it forward. Brim was hard-disabled on
belt printers, leaving no remedy at all.
Generate the brim on the belt plane instead. The object's belt footprint is
the union over layers of each slice clipped to that layer's contact band; the
brim is offset from it in a "flattened" frame where the shear axis is
stretched by 1/cos(tilt), so ordinary Clipper offsets measure true on-belt
distance. It is emitted as cross-belt lines, one per layer band, anchored to
a fixed fraction of the band so every line shares a nozzle-to-belt clearance
and therefore comes out the same width; flow is matched to the resulting band
pitch, keeping the sheet uniform and gap-free.
Three new controls, all belt-only:
* Leading brim length - extends the brim ahead of the part along the belt,
on every downhill-facing edge of its contact area. This apron necessarily
prints BELOW the object's first layer, since layer 0 is the part's leading
contact, so it needs brim-only bands of its own.
* Extra brim width - widens the brim sideways across the belt only.
* Brim type "Leading edge only" - brim at the part's first belt contact and
nothing after it. Appended last in BrimType so no existing value shifts;
degrades to an outer brim off belt printers, with a warning.
The apron bands are lightweight records rather than a Layer subclass, so no
fabricated Layer::id() can leak into initial-layer temperature selection, the
spiral vase probe, cooling or gradual interpolation. They are generated in
posSupportMaterial because their print_z values must exist before ToolOrdering
is built at psWipeTower, and they are emitted from a short dedicated branch in
process_layer that runs before any layer pointer is dereferenced.
The footprint is closed before offsetting outwards: a belt contact patch is
often a broken-up strip, and the merged offset rings of two islands closer
than 2 x brim_width would otherwise fill the space between them - space that
lies under the part.
Also fixes a pre-existing bug where PrintObject::get_first_layer_bbox()
overwrote a valid bbox with an unassigned one on any belt printer with a brim
configured, because has_brim() was true while make_brim() returned early.
Belt brim is refused alongside the prime tower and spiral vase, and requires
one instance per PrintObject - translating an instance along the belt axis
changes its physical belt-floor Z. Untilted belt printers are unchanged: they
still get no brim, since the plate brim is emitted out of skirt_brim_groups(),
which _make_skirt() never builds for a belt printer.
A mate connector was visible only to a program. resolve_datum_coordsys had exactly ONE consumer
in the whole tree -- McpControl.cpp, the agent socket -- so the frame every mate is built on
could not be seen at all, and the two questions a connector has to answer on sight had no
answer in the viewport: which way does Z point (the VERSE), and which of the pair is anchored
versus about to move (the POLARITY).
The glyph is Onshape's proven core plus the part nobody ships. Disc for the XY plane, one gold
quadrant for the roll -- the only in-glyph answer to "where is X", which matters because
Fastened and Slider lock the clocking -- and a Z arrow drawn on +Z ONLY, never double-headed.
Polarity is carried by the head: a filled cone travels, an open collar receives. Onshape,
Fusion, Inventor and FreeCAD all draw both ends of a mate identically, which is why "which part
moves?" is a standing complaint; nothing here invents new semantics, it just stops hiding them.
Polarity is read from the committed Mate features, not only from the open card. A connector some
mate drives must read as driven whenever it is on screen, or the glyph tells the truth only while
a dialog happens to be open. The card, when open, still wins -- that is the live intent.
Judged on the rig rather than in a mock, which changed three decisions:
- Three RGB axis arms lose to one Z arrow. Rendered side by side (SNAPORCA_GLYPH=A selects the
Onshape-style trio), the three heads are as large as the 22 px disc, they bury the quadrant, and
at an oblique angle they pile into a smudge -- and the trio is indistinguishable from the move
gizmo and the bed triad, which are already RGB arrow trios in this viewport.
- Depth off floats, depth on tears. With GL_DEPTH_TEST off, connectors on faces pointing AWAY from
the camera drew their discs over the solid, so the part looked covered in frames that were on its
back. Turning depth on fixed that and immediately z-fought: the disc is exactly coplanar with its
face and came out a broken dotted arc. Depth ON plus a 0.7*upp lift along Z buys both, and scaling
the lift by upp keeps it sub-pixel instead of opening a visible gap on zoom-in.
- Foreshortening degenerates an arrow into a dot when the axis points at the camera. It now draws a
ring instead of silently vanishing, which is what a naive projection does.
Everything is sized in screen pixels via upp = 1/zoom, like every other gizmo here: a connector is
a symbol, not a part, so it must not shrink with the model.
Research and the empirical findings are written up in DESIGN_MATE_CONNECTORS.md section 8b;
rig images in artifacts/shots/g-0*.png, vendor reference glyphs in artifacts/glyphs/.
Not fixed here, and recorded rather than papered over: the quadrant collapses to a blob at a
grazing angle, which is exactly when the roll is hardest to read (F4); roll-undefined in red makes
the least important connector the loudest thing on screen (F5); and a true grazing view, a curved
face, and overlap with the move gizmo are still untested. Also surfaced while testing and unrelated
to drawing: add_mate accepted a mate between two connectors on the SAME body and duly transformed
the body relative to itself -- a concrete instance of the missing validation already filed as G6.
Fork parity unchanged: DesignCanvas.cpp 16, DesignPanel.cpp 30, the other four files 0.
The belt-printer branch is failing two profile gates. Both stem from the
three belt-only vendors (Custom's generic belt printer, IdeaFormer,
Printcepts) not existing upstream, so upstream maintenance passed them
by.
Slice check: 4 of 1015 printers failed - Custom's MyBeltPrinter 0.2/0.4/
0.6/0.8 nozzle all fell back to "Default Setting". No process profile in
the Custom vendor listed any MyBeltPrinter in compatible_printers, and
Custom's fdm_belt_common pointed default_print_profile at "0.20mm
Standard @System", which does not exist in that vendor's index, so the
generic belt printer had no usable process at all. This gap dates to
when MyBeltPrinter was added (2026-04-07); it only started failing now
because the slice-check job is newer than that.
Adds two process profiles modelled on the sibling @MyKlipper ones:
- 0.20mm Standard @MyBeltPrinter - 0.4/0.6/0.8 nozzles
- 0.12mm Fine @MyBeltPrinter - 0.2/0.4 nozzles
The split is forced by hardware: the 0.2 nozzle preset caps
max_layer_height at 0.16, so a single 0.20mm profile cannot legally
cover
it. fdm_belt_common now defaults to the standard profile and the 0.2
nozzle preset overrides to the fine one.
setting_id: 14 files failed the rules introduced in #14432. That
migration renumbered 7425 files across 61 vendors but skipped these
three, leaving BabyBelt Pro, IdeaFormer IR3 V2 and MyBeltPrinter
squatting the "G*" id space reserved for Bambu (GMPC0BBP01, GMIF001,
GM_BELT_00x) and four instantiated filament/process presets carrying no
setting_id at all. Regenerated with
scripts/assign_vendor_setting_ids.py.
Also repoints the identical dangling "0.20mm Standard @System" in
Printcepts' and IdeaFormer's fdm_belt_common at their own real process
profiles. That is a no-op today because both concrete printers override
it, but it is the same landmine that took out MyBeltPrinter.
Vendor index versions bumped so check_installed_vendor_profiles() will
re-install the corrected profiles over an existing install.
Note: changing a shipped preset's setting_id can orphan user presets
that reference it as base_id. #14432 accepted that tradeoff for 61
vendors; this keeps these three consistent with the rest.
Verified: orca_extra_profile_check.py reports 0 errors across 66 vendors
(was 14 files with errors), and OrcaSlicer_profile_validator -s slices
all 1015 printer presets successfully (was 4 failures).
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The belt-printer branch is failing two profile gates. Both stem from the
three belt-only vendors (Custom's generic belt printer, IdeaFormer,
Printcepts) not existing upstream, so upstream maintenance passed them by.
Slice check: 4 of 1015 printers failed - Custom's MyBeltPrinter 0.2/0.4/
0.6/0.8 nozzle all fell back to "Default Setting". No process profile in
the Custom vendor listed any MyBeltPrinter in compatible_printers, and
Custom's fdm_belt_common pointed default_print_profile at
"0.20mm Standard @System", which does not exist in that vendor's index,
so the generic belt printer had no usable process at all. This gap dates
to when MyBeltPrinter was added (2026-04-07); it only started failing now
because the slice-check job is newer than that.
Adds two process profiles modelled on the sibling @MyKlipper ones:
- 0.20mm Standard @MyBeltPrinter - 0.4/0.6/0.8 nozzles
- 0.12mm Fine @MyBeltPrinter - 0.2/0.4 nozzles
The split is forced by hardware: the 0.2 nozzle preset caps
max_layer_height at 0.16, so a single 0.20mm profile cannot legally cover
it. fdm_belt_common now defaults to the standard profile and the 0.2
nozzle preset overrides to the fine one.
setting_id: 14 files failed the rules introduced in #14432. That
migration renumbered 7425 files across 61 vendors but skipped these three,
leaving BabyBelt Pro, IdeaFormer IR3 V2 and MyBeltPrinter squatting the
"G*" id space reserved for Bambu (GMPC0BBP01, GMIF001, GM_BELT_00x) and
four instantiated filament/process presets carrying no setting_id at all.
Regenerated with scripts/assign_vendor_setting_ids.py.
Also repoints the identical dangling "0.20mm Standard @System" in
Printcepts' and IdeaFormer's fdm_belt_common at their own real process
profiles. That is a no-op today because both concrete printers override
it, but it is the same landmine that took out MyBeltPrinter.
Vendor index versions bumped so check_installed_vendor_profiles() will
re-install the corrected profiles over an existing install.
Note: changing a shipped preset's setting_id can orphan user presets that
reference it as base_id. #14432 accepted that tradeoff for 61 vendors;
this keeps these three consistent with the rest.
Verified: orca_extra_profile_check.py reports 0 errors across 66 vendors
(was 14 files with errors), and OrcaSlicer_profile_validator -s slices all
1015 printer presets successfully (was 4 failures).
snaporca-200 asked which of the two models of "the card's face input" is right,
because clicking empty canvas now clears the selection (snaporca-od0) and made
them visibly disagree: Thicken / Shell / Draft read the LIVE selection and their
label reverts to "(pick a solid face)", while Hole / Thread LATCH the face they
were opened or picked on and keep it. The complaint was that Hole then drills a
face you can no longer see selected.
Taken to the rig, that turns out to be the wrong half of the story. With Hole
open and its face picked, a click on empty canvas leaves the Ø6.0 ghost and its
dimension gizmo drawn on that exact face — the card was never operating in
secret, it was showing its target the strongest way a CAD tool can. Meanwhile
Draft, whose behaviour was held up as the honest one, threw the pick away and
had to be told the face again.
So neither model replaces the other. They are different in kind: Thicken /
Shell / Draft are operations whose operand IS the selected face, and Hole /
Thread are placement tools with their own plane state that a pick merely seeds.
The latch is also the kinder of the two now that empty clicks are a deliberate
gesture — a stray one costs Thicken a pick and costs Hole nothing.
What was genuinely missing is that nothing in those two cards NAMED the latched
face, so after such a click the only words on screen were the viewport's
"Nothing selected" over a ghost about to drill. Both cards now carry an "On
face" row, the way the other three already do:
Hole Face 5 | (none — uses Hole plane)
Thread Face 1 | Edge 2 | (none — uses Thread plane)
Thread names an edge when the cylinder came from a circular rim rather than a
cylindrical face, which the code already distinguished internally and never
said out loud.
Verified on both rigs, every state driven through the GUI: face pick on open
and on live pick, survival across a click on empty canvas, and the fallback
after choosing XY/XZ/YZ from the plane dropdown. Thread's edge branch was
exercised on a revolved tube's rim, its face branch on the same tube's outer
wall.
Filed while here, surfaced by the new row rather than caused by it —
snaporca-uif9: re-editing a stored Hole/Thread from the feature tree restores
f.plane into the dropdown but never clears m_hole_on_face, so the re-edit
silently reuses the PREVIOUS card's latched face. Now visible by name instead
of invisible.
Fork parity unchanged: DesignPanel.cpp 30, DesignPanel.hpp 0.
Labels, tooltips, menu items and dialog text mixed "IDEX/IQEX" with "IMEX"
for the same feature. IMEX is now the user-facing name throughout: IDEX and
IQEX are hardware categories, IMEX is the feature spanning them. The
hardware terms remain only where they help a user tell whether the feature
applies to their printer.
Comments and log messages keep IDEX/IQEX, where the specific carriage
topology is the more precise term.
Also corrects the imex_parallel_mode tooltip, which pointed at a
"Printer -> IDEX/IQEX tab" that does not exist; the options live under
Printer -> Multimaterial -> IMEX Configuration.
None of these strings appear in any .po or the .pot, so no translation is
affected.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
IMEX placement validation only walked model instances. The prime tower is
not a ModelObject, so it could sit in a secondary zone or a carriage
collision strip and slice with no warning -- on mirror mode, a carriage
crash. Span (paired-gantry multicolor) is what made towers reachable in
parallel modes, so this is a gap in that feature, not inherited breakage.
The check now returns a cause instead of a bool so the message can name the
offender, and the tower and per-instance paths share one predicate,
imex_hull_violates_zones(), moved to libslic3r and covered by tests.
Overlap is area-based: a hull flush against a zone boundary is legal, only
a crossing violates. The tests pin that in both directions, since switching
to a touch-based test would silently block placements that work today.
The tower footprint comes from the same estimate the scene draws, so
validation matches what the user sees and drags. Three config reads there
are load-bearing in non-obvious ways and are commented at the point of use.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Both found by Kimi reviewing the previous two commits, both then reproduced here
before being touched.
snaporca-97z. The re-pick escalation required m_solid_sel, m_sel_body, m_sel_face AND
m_sel_edge to all match the previous pick. That looked stricter and was wrong: the
edge branch sets only m_sel_edge and m_solid_sel, leaving m_sel_face as whichever face
the ray happened to enter through — and a shared edge is entered through a different
face depending on which side you view it from. So picking an edge and picking that
same edge again from the other side compared equal edges, unequal faces, and refused
the escalation the status line had just promised. Now the comparison is made at the
level that was picked and nothing else. The edge id is already the stable global one
from edge_index_of, so it identifies the edge without help from the face.
Reproduced on the rig without needing to orbit, since two clicks 8px apart across an
edge enter through different faces:
pick -> sel=3 body=0 face=5 edge=3
ray -> body=0 face=0
re-pick -> escalated to whole body 0
pick -> sel=1 body=0 face=-1 edge=-1
Same edge, face 5 then face 0, escalation fires. The old condition could not.
The frame-move case. The chip is anchored at an absolute screen position, and until now
nothing told it the window had moved — only a resize, a status change or a tab switch
re-placed it. Dragging the window by its title bar left it stranded where it was,
verified on the rig by moving the frame and watching it stay put. wxEVT_MOVE on the
top-level frame, alongside the ICONIZE and ACTIVATE binds from the previous commit.
Two related cases are filed rather than bound, because the list of window-geometry
events to chase is exactly what snaporca-lcq argues should stop: a layout change that
translates the canvas without resizing it, and wxEVT_DPI_CHANGED.
Not fixed, deliberately, and recorded on snaporca-97z: clicking the same FACE but
landing within the vertex or edge tolerance resolves to a different kind and so does
not escalate — that is the "smallest thing under the cursor" rule working as
documented; and a vertex re-pick after moving the body compares stale world
coordinates.
Verified on both rigs. The cross-face edge case was exercised on orca_cad;
DesignSketchTool.cpp is byte-identical across the forks, so snaporca inherits it, and
its face-level escalation and empty-click clear were re-checked there directly.
Fork parity unchanged: DesignSketchTool.cpp 0, DesignCanvas.cpp 16.
Found by minimising the app on the rig with a face selected: the whole screen goes
black and the chip is still drawn on the bare desktop. A wxPopupWindow is
override-redirect — the window manager does not own it — so it neither iconises with
its frame nor stacks behind other applications. IsShownOnScreen does not catch this
either: an iconised frame still counts as shown, which is why the guard added for the
tab case sails straight past it.
So the frame has to say so itself: ICONIZE and ACTIVATE, both routed through the same
show_status_hud the page change already uses. Restoring is safe — a popup cannot take
focus, so our own Show() cannot re-trigger either event — and restoring while some
other page is up still leaves the chip down, because show_status_hud(true) goes
through place_status_hud's IsShownOnScreen guard.
Verified on both rigs: chip up, minimise -> screen black and empty, restore -> chip
back with its text and the face still selected. Restore while on Prepare -> chip stays
down.
This is the fourth defect from the same root, so snaporca-lcq now asks the question
these binds keep deferring: whether the line should be canvas content, like the view
cube and the round view buttons, rather than a window that has to be told about every
way a window can stop being visible.
Fork parity unchanged: DesignCanvas.cpp 16.
Two things a click on nothing should already have done.
snaporca-od0. A click that hit no geometry left the solid selection standing. A
rubber band swept over empty space has always cleared it (pick_bodies_in_rectangle),
and the two gestures cannot disagree about the same outcome. The visible cost was in
the escalation that landed last commit: "click the face, click away, click the face
again" arrived as the SECOND click on the same face and took the whole body, when the
click away was the user letting go of it. Now the miss clears and says so.
This gives up something real, deliberately: Thicken / Shell / Draft hold their input
face in the panel's selection, so a stray click on empty canvas with one of those
cards open hands that face back. Their handlers already write the "(pick a solid
face)" placeholder and rebuild the ghost when the selection empties, so the card SAYS
it lost the pick rather than confirming against a face the viewport has stopped
highlighting. An orbit drag never reaches this branch — it exits at the 8px budget —
so panning the view still does not deselect.
snaporca-dlj. The status line is a wxPopupWindow, which is a TOP-LEVEL window: hiding
the Design page does not hide it. Select a face, switch to Prepare, and the chip was
still there reading "selected (whole body) — right-click for what applies to it" on a
tab with no such selection and no such menu. Same cause, second symptom: a status
update arriving while the page is hidden anchored against a client size that is not
the size the page will have, and parked the chip on the tab bar. So: an
IsShownOnScreen guard in place_status_hud, show_status_hud(bool) to take it down and
bring it back with its text intact, driven from the page-changed handler.
Verified on both rigs, not by reasoning about it: face 5 selected -> click bed ->
"Nothing selected", tint gone -> click the same face -> face 5 again, NOT the body ->
click it again with no click away -> whole body, so snaporca-gem is intact. Prepare ->
chip gone; back to Design -> chip returns. KEYTRACE across the round trip shows
shift+S then R still reaching the canvas (ui_mode 0 -> 1, Rectangle armed), which is
the focus theft this popup replaced a wxFrame to avoid.
Fork parity unchanged: DesignPanel.cpp 30, DesignCanvas.cpp 16, headers and
DesignSketchTool.cpp 0. MainFrame.cpp is outside that set and was edited per fork.
BeltAffine activates the FirstLayerPlane evaluator unconditionally, so on a
non-belt printer on_first_layer(point) stopped agreeing with the legacy
slicing-layer-0 test. Every per-path first-layer call site in _extrude then
took the non-first-layer branch, and first-layer speeds were skipped: brim
came out at the volumetric fallback (24.6 mm/s) instead of initial_layer_speed
(10 mm/s). This is the shared speed path, so it affected all printers on this
branch, not just belt ones.
Auto resolves to BeltAffine only when belt_printer is set with a non-zero
slicing rotation, and to XY (evaluator inactive, legacy behaviour) otherwise --
exactly what the option's own description already promised.
Caught by "Brim uses first layer speed" (upstream #14616), which arrived with
the upstream merge; the bad default dates back to a9bae54f20 (#30). Verified
against a pristine upstream/main build, which passes the same test.
tests/fff_print: 100/100 test cases, 1085 assertions (was 99/100).
Both belt regression tests still pass, confirming Auto still resolves to
BeltAffine for belt printers.
Note: this changes a config default. Projects and profiles that stored
first_layer_plane explicitly are unaffected; those relying on the default will
now get correct first-layer speeds on non-belt printers, so their G-code
changes accordingly.
Processes a minimal belt start sequence through GCodeProcessor::process_buffer
and asserts the move preceding the first extrusion keeps its real Z, so it can
no longer back-transform to model Y~=0 and produce the phantom extrusion line.
Belt printers are non-Bambu, so the processor uses the compatible reserved
tags ("TYPE:"); the test sets s_IsBBLPrinter=false (saved/restored via an RAII
guard) to mirror the real printer. Proven to fail without the fix (the
prepare-stage move's Z is pinned to the first-layer height, 0 here) and pass
with it.
On a belt printer the sliced preview drew a stray extrusion-colored line
from Y~=0 to the model, rendered in the first extrusion role's color. It is
not a travel and does not occur on non-belt printers.
GCodeProcessor::store_move_vertex pins a move's stored Z to the first-layer
height during the start-G-code "prepare" stage. That is a harmless cosmetic
tidy-up on a normal printer, but on a belt printer the designed-view
back-transform couples machine Z into the rendered model Y (the belt tilt
mixes the height and belt-feed axes). Pinning Z back-transforms the last
prepare-stage move (the unretract before the first extrusion) to model
Y ~= 0, and libvgcode then draws a phantom extrusion segment from Y ~= 0 to
the first real toolpath.
Keep the real Z for belt printers (gated on belt_tilt_angle, parsed from the
G-code header before the body) so prepare-stage moves back-transform
correctly. Non-belt processing is byte-identical. The emitted G-code was
already correct; this is a preview-geometry fix.
Locks in the fix from the previous commit. A fresh BeltGCodeWriter has an
unestablished planar position (is_current_position_clear() == false) and its
m_pos.xy is the origin (0,0). With a pending NormalLift z-hop, travel_to_xyz
used to lift in place via _travel_to_z(), which in belt mode shears the origin
into a machine Y ~= the layer Z — a move far up the gantry.
The test configures an X-tilt 45 deg belt transform, defers a z-hop via
lazy_lift, travels to a near-belt first point (transformed gantry Y ~= 1mm),
and asserts no emitted move has Y anywhere near the layer Z. Verified to fail
without the fix (max emitted Y = 100.0 vs the destination's ~1.0) and pass with
it.
On a belt printer the first travel of the print emitted a bogus move to
the bed corner with the nozzle far up the gantry, e.g.
G1 X95 Y168.19 Z237.857 F12000
right after the first "; printing object" line. Y168 (≈ the layer Z)
is out of the gantry's range.
Root cause: the layer-change z-hop is deferred via lazy_lift and consumed
by the first BeltGCodeWriter::travel_to_xyz, whose NormalLift branch does a
separate lift-in-place via _travel_to_z(target.z()). On a normal printer
_travel_to_z emits a Z-only move, but in belt mode Z is coupled to Y/X, so
_travel_to_z re-emits the current m_pos through the belt shear. At print
start (and after custom gcode) m_pos.xy is still the uninitialised origin
(0,0), which the back-transform + axis-remap shear into machine
(X=bed_max, Y=layer_z) — the illegal move.
Guard the NormalLift branch on is_current_position_clear(), matching the
SlopeLift branch directly above it which already does so. When the position
isn't established there is nothing to lift over, and the xy_z_move that
follows travels straight to the destination with full XYZ, establishing the
correct position. Bookkeeping is unaffected: in this path m_lifted stays 0,
so no spurious restore move is produced.
Verified by re-slicing the repro project: the start-of-print move is now
G1 X44.946 Y.621 Z237.857 (straight to the first object point), no move
touches the bed-max X edge, and the max Y over the whole file is 62.8mm
(printable_height 100).
Upstream retyped travel_speed and travel_speed_z to ConfigOptionFloatsNullable
and initial_layer_travel_speed to ConfigOptionFloatsOrPercentsNullable, so the
scalar .value / get_abs_value() accessors no longer compile. BeltGCodeWriter.cpp
is belt-only and merged without conflict, so this only surfaced at build time.
Index them the way the base GCodeWriter does -- .get_at(m_cached_extruder_idx)
and get_abs_value_at(..., m_cached_extruder_idx) -- keeping belt's per-point
first_layer_for_point test rather than the base class's m_is_first_layer.
m_cached_extruder_idx moves from private to the existing protected block that
already exposes writer state to subclasses, so the belt writer resolves the
per-extruder index identically to the base writer instead of guessing one.
Brings the belt-printer work up to date with 591 upstream commits.
Conflict resolutions (12 files, 42 hunks):
- GCode.cpp: adopted upstream's per-filament/per-nozzle config refactor
(get_filament_config_index, NOZZLE_CONFIG), the extracted
generate_timelapse_gcode + farthest-point timelapse, and the
ConfigOptionFloatsNullable calibration options. Re-applied the belt
hooks on top: init_belt_writer / axis remap / FirstLayerPlane setup,
on_set_origin, the belt-corrected calib_z for the volumetric speed
tower, and path_on_first_layer (belt's per-path first-layer test) in
place of upstream's layer-index on_first_layer() in the acceleration,
jerk and overhang-detection paths. Swept upstream's new m_writer.
uses to m_writer-> since belt holds the writer by unique_ptr.
- interpolate_value_across_layers: kept upstream's banded stepping and
belt's object-Z-span ratio; dropped upstream's duplicate ratio decl.
- Plater.cpp: took upstream's guarded add_model(...) early-returns and
the VFA vfa_layer_height plumbing; kept the belt temp-tower path,
_calib_apply_belt_mode and belt_calib_flip_ringing_tower. Dropped the
VFA "cut upper" block, superseded upstream by model scaling.
- Brim.cpp: upstream's ObjectInstanceID-keyed brimAreaMap, keeping the
belt early-return.
- 3DScene.cpp: kept both the belt build-plate tilt up_direction and
upstream's per-extruder printable-height shading.
- GCodeViewer.cpp: kept upstream's dim-previous-layers setup and belt's
exemption from the same-result early return.
- TreeSupport.cpp: upstream's >= 0 roof-layer fix inside belt's
belt-floor branch.
- calib.cpp / GCode.hpp / GCodeWriter.{cpp,hpp} / Print.hpp: upstream's
additions adapted to belt's pointer-held writer and helpers.
- Custom.json: kept profile version 02.04.00.03 (belt) over upstream's
02.04.00.01; both bumped from 02.04.00.00.
Building this tree needs the wxInspector dependency, which upstream
added in the interim (python3 and wxWidgets 3.3.2 were already present
in the shared deps prefix).
A click could point at a face, an edge or a vertex, but never at the body those
belong to: offer_selection_kind() can only return BodySolid when all three are
clear, which a viewport click never produces. The rubber band was the only door,
and the status line said "face 5 selected" while the user believed they had taken
the body. A second click on the SAME sub-element now escalates to it (snaporca-gem).
Not the pick cycle that was removed in bc2b741ce9 -- that one was silent and three
deep, so no click had a predictable meaning. Here the status line names the next
click before you make it, and a further click just takes the face under the cursor
again, which needs no teaching. Double-click is untouched: wx sends Down/Up/DClick/Up
and only the first Up carries a pending press, so a fast double-click still zooms to
fit and picks once.
The status line itself moved to the base of the viewport. In the side panel it was
clipped at ~73 characters with no warning and no wrap -- set_status()'s Wrap() never
took effect (snaporca-8cc) -- which silently length-limited every hint in the tab; the
first version of this change lost a clause to it. m_status is kept, hidden, as the
owner of the text and its colour, and the line is drawn in a bottom-left twin of the
readout HUD where there is a whole window's width.
Three defects found driving it on the rig, none of which the build could see:
* the HUD as a wxFrame took the WM's keyboard focus every time it was raised, and
the canvas then received NO key events -- every sketch shortcut silently dead.
Caught with SNAPORCA_KEYTRACE: shift+S logged a line, the following R logged
nothing. It is a wxPopupWindow now, which cannot be focused. SetFocus() on the
canvas does not fix it: focus was on another toplevel.
* zero vertical padding fits the popup tighter than the font's line box and clips
the glyphs; 6 (what the readout uses) reads as a two-line box. 3 is right.
* "has a caller chosen a colour?" compared the label's foreground against its
PARENT's, which differ by default, so every line counted as chosen and the
neutral text came out the panel's dark grey -- invisible on a dark chip. Compare
against the colour the label was created with, captured before any caller writes.
Verified on both rigs against fresh binaries: sketch -> extrude -> click face ->
click again -> whole body tinted, offer opens with the body rows live and Create /
Add material correctly greyed. Keyboard drives the whole sequence.
Filed and NOT fixed here: snaporca-od0 -- a bare-plate click does not deselect the
solid, so "click away, click back" escalates. Pre-existing; clearing there would also
drop the face the Thicken/Shell/Draft cards hold, which needs its own pass.
Refs: snaporca-gem, snaporca-8cc, snaporca-od0
The third door onto the offer, after the viewport right-click and the Menu key. A body ROW is
an unambiguous body, so the offer reports BodySolid and the body verbs act on the row you can
see highlighted — the confirmation a face pick cannot give, since pointing at a face lights the
face and never the body the verb will change. The status line has been promising exactly this
("Body N selected — right-click for what applies to it") since before any handler existed on
that list; the product was advertising a gesture that did nothing.
WHAT THE RIG CAUGHT THAT THE BUILD DID NOT. The first version hung the state normalisation off
wxEVT_TREE_SEL_CHANGED. But SelectItem() on a row that is ALREADY selected fires no selection
event, so a stale vertex from an earlier viewport pick survived — and offer_selection_kind()
tests vertex FIRST, so right-clicking the body row served the VERTEX offer while the row sat
highlighted: Fillet/chamfer/draft greyed, Mirror standing where Repeat belongs, "vertex
selected" still in the status line and the cyan marker still on screen. The happy path (fresh
row, nothing else picked) looked perfect, which is why only the deliberate stale-state sequence
exposed it. Reading the code would not have shown it — SelectItem looks like it selects.
So the normalisation is no longer a selection handler. apply_body_row() is called
UNCONDITIONALLY by both doors, because taking a body from the list means the same state change
however it was asked for. It also clears m_sel_solid_vertex, which the original handler never
did — latent while nothing opened the offer from that list, and immediately fatal once
something did.
Verified on both rigs with the failing sequence itself: pick a vertex, then right-click the
already-selected row. Fillet/chamfer/draft enabled, Repeat back in place, status reads "Body 1
selected", vertex marker gone.
Does NOT touch the feature tree. That needs new selection kinds (offer_selection_kind has no
notion of "a feature is selected") plus verbs the atlas does not contain — Suppress, Rename,
Reorder, Roll back — and is filed separately.
Tommaso: "i deleted a body using rubber band selection, but this is not intuitive as all
the ux revolves around clicking". Correct on both counts, and a correction to what I said
last round: the rubber band IS implemented and shipping (pick_bodies_in_rectangle, m_rubber,
the drag branch in on_mouse). What is unbound is whole-body picking via CLICK; I read the
comment about the click path and wrongly generalised it to the gesture as a whole.
The handlers were never the problem either. Move, Mirror, Cut, Mass and Colour all resolve
their target through selected_body_default() / m_sel_solid_body, and that is already set when
you click a FACE — level >= 1 records the body. They would have worked from a click all
along. The only thing keeping them out was the atlas gate: accepts listed body_solid and no
face kind, so offer_selection_kind() returning FacePlanar filtered the rows away. This is
therefore an atlas-only change, no handler edits.
Cut, Split, Mirror, Transform, Mass and Colour now accept face/edge/vertex as well, matching
what Delete Body already did. Edges and vertices are included deliberately, not just faces: a
click resolves to a vertex, an edge or a face depending on where inside the pixel it lands,
so accepting only faces would make Move vanish whenever you clicked near a corner — a flicker
that reads as a bug and gets reported as "sometimes it works".
NOT widened: Extrude on a face means push/pull THAT face, and Thicken consumes the face you
point at. Both have genuine face-specific meaning, so widening them would change what they
do rather than where they can be reached from.
The rubber band keeps its job — it is still the only way to take a body without also naming
one of its faces. It just stops being the only door.
Verified on both rigs from a plain face click: Transform > Move opens with Body = Extrude2
(resolved from the face pick), Modify > Edit / Delete Face / Colour / Delete Body, and
Reference > Mass.
Reported by Tommaso: select a body, and there is no Delete in the offer. Two independent
faults stacked behind that.
FIRST, clicking a body never selects the body. Whole-body picking is deliberately unbound
(DesignSketchTool.cpp) pending the rubber band, so a viewport click only ever yields
Face/Edge/Vertex. The offer therefore saw face_planar, and "delete" accepted body_solid but
no face kind, so the row was filtered out entirely — while the status line read "Body 1
face 0 selected", which actively teaches the wrong model.
SECOND, even selecting the body from the Bodies list, Delete refused in red: "Select the
FEATURE that created this body". CadBody had no link back to its maker, so the offer was
advertising a verb it could not perform — worse than the action:null rows fixed earlier this
session, because this one is ENABLED and its refusal reads like user error.
CadBody::source_feature fixes the second. It is stamped in ONE place, the recompute loop,
and the rule is just "still unset?". That is sufficient because of an invariant worth
stating: no feature ever replaces a whole CadBody. Every in-place op writes only .shape
(boolean, cut, mirror-fuse, transform, dress-up — all 8 sites checked), so a body keeps the
stamp it was born with; a consumed body is erased outright, taking its stamp with it; and
the only bodies still at -1 are the ones the current feature just pushed. A feature type
added later needs no change here as long as it keeps to that invariant.
"Delete Body" fixes the first, sitting beside "Delete Face" in Modify and reachable by
pointing at any face/edge/vertex. The two names cannot be confused, and "delete" gave up the
body kinds so both can never appear for one selection. Deleting a body removes the feature
that made it, which is a real edit to the recipe, so it asks first and NAMES the feature — a
body vanishing from the viewport is not evidence of which feature went, and this is the one
action here that cannot be eyeballed.
Multi-body delete is NOT offered. bodies_2 was in the first draft of the verb; the handler
deletes exactly one body, so a two-body selection would have silently deleted whichever was
m_sel_solid_body. Caught before it reached a binary, at the cost of one rebuild.
Verified on BOTH rigs, full round trip: click a face -> Modify > Delete Body -> "Delete
Extrude2?" -> body gone, Sketch1 correctly left behind, panel falls back to the idle hint ->
Undo -> Extrude2 and Body 1 restored.
The previous commit cites snaporca-y7q, which does not exist — I wrote the ID from memory
instead of reading it back from the bug I had just filed. The real one is snaporca-kgx,
"Offer: Thicken (and peers) open with the picked face discarded". The comment is corrected
here; the commit message above it cannot be, so this note is the pointer.
snaporca-y7q. Thicken's opener cleared m_sel_solid_face outright. That was right when the
only door was a toolbar button — a button carries no selection, so pressing Thicken had to
clear and ask you to point at something. The offer inverted it: the verb is now invoked ON
a face, and the same line threw away the only thing the user had said. The card opened
reading "(pick a solid face)" over an immediate "thicken: face not found" — you pointed at
the face and were told none could be found.
Keep the pick when the body combo landed on the body it came from (the index is per-body,
and selected_body_default() returns exactly that body when it is valid).
Two neighbours had the mirror-image flaw, both invisible for the same reason — the value
was right and the ghost updated, so only the label lied:
- Thicken had NO live label update at all. Nothing outside the opener ever wrote
m_thicken_face_label, so while the card was open you could pick face after face and it
still read "(pick a solid face)".
- Shell and Draft wrote theirs ONLY from the pick handler, which runs while a card is
already open — so opened from a selection they showed the previous pick, or the
placeholder over a face they were about to use.
So the label is now written once in open_tool(), which every door goes through. The
edit-feature path already restores m_sel_solid_face from the stored feature BEFORE calling
open_tool, so it agrees rather than fights.
Verified on the snaporca rig: face 4 of an extruded plate, offer > Add material > Thicken
now opens "Face: Face 4" with "Preview — 24 triangles" and confirms to a real Body 2. Draft
opened from a face shows "Face 3" and previews the taper. This fork is code-identical here
bar the two permitted DropDown divergences; it still owes a build of its own (snaporca-5pl).
Project keeps its clear: there "(all edges)" is a legitimate default mode rather than a
failure, so changing it would alter behaviour with no reported problem behind it.
snaporca-7ih's remaining half. Both flyout factories registered their verbs INSIDE the
widget-building loop, so the ~40 retired tool buttons had to be constructed and then
Hide()n: skipping construction would have deleted 42 offer verbs (26 fly:<family>#<row>
+ 16 Shift+keys) while their rows still rendered and did nothing when picked.
Register first, build second. The addresses are pure data; the widget is one door onto
them, not their owner. A family absent from kBarKeep now returns before any wxWindow is
made. The keep-list stays a one-line data decision, not a structural one.
And close the class of bug for good: the constructor now verifies, once, that every verb
the atlas marks wired resolves to a real registration, logging each break and asserting in
debug. Rows that render and do nothing have shipped three times (edit_feature and sk_move
with action:null, then this) and are invisible from either side alone.
Verified on the snaporca rig by walking the offer, not by reading the code — all four
at-risk address kinds run with no widget behind them: fly:design_rect#2 drew an OBLIQUE
rectangle (the third variant, not the family's first), key:S+E opened Extrude with its
10 mm gizmo, fly:material#4 opened Thicken. Hover hints, icons and nesting intact. This
fork is code-identical here bar the two permitted DropDown divergences; it still owes a
build of its own (snaporca-5pl).
Two hints were wrong and are fixed: Cut said "Split the body with a plane", colliding with
the Split verb one row away and pointing at a card for a value the canvas already offers as
a draggable arrow; Split never said its plane comes from a picked face.
Also, because it blocked the verification and will block the next one: gui-session.sh
killed by full path while its own app_pid() matched by basename, so a differently-pathed
instance survived, held the single-instance lock, and got reported as a healthy session —
a Jul-30 binary nearly passed as this build. It now kills by basename and prints which
binary is actually on screen. Traps 6 and 7 documented.
Mirror of snaporca 2b3e890165 (DesignPanel.cpp applied as a patch; parity 30 / 16, shared
files byte-identical).
All 86 verbs now carry a hint: 55 extracted from the C++ tool definitions so the offer and
the armed-tool hint cannot drift, 31 written by hand. One wxEVT_MENU_HIGHLIGHT binding
shows the hovered verb's hint in the status line. The generator asserts that no wired verb
lacks one.
Also: all 200 status writes go through set_status(), which wraps instead of clipping at the
panel edge; and the empty-document hint is called from on_tab_shown() as well, since
after_tree_edit() never runs on a freshly opened tab.
Verified on the rig.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca 033347d062 (parity 30 / 16).
Retiring the toolbar made ten hints untrue: each named an action whose door had moved to
the offer, or a button no longer on the bar. They now name the gesture. An empty document
blanked the status line entirely and now says how to start.
Known and not fixed here: m_status does not wrap, so long hints clip; and the 86 offer
verbs still have no per-verb hint of their own.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca b3d4cf85af (parity 30 / 16).
Hiding it with the drawing tools was wrong: Construction is a persistent MODE, not a tool —
the Bed checkbox, not the Line button. Q and the offer's Construction row kept toggling a
checkbox nobody could see, so you could not tell whether the next line would be construction
geometry.
Scoping unchanged and already correct: m_tb_sketch is shown only in UiMode::Sketch, so it
appears exactly while a sketch is open or being edited.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca 809aa9df87 (DesignPanel.cpp applied as a patch; parity 30 / 16, shared
files byte-identical).
fadd() and sadd() now gate what reaches the bar: file operations, Bed, Undo/Redo, Delete
selected, Commit to Plate, Confirm/Cancel, plus Place on Face and Section view — the last
two because they are chrome_only in the atlas and have no offer row to fall back on.
The tool buttons are still built and then hidden, deliberately: their fly: addresses and
Shift+key bindings are registered inside the widget-building loops, so not building them
would silently drop 42 verbs from the offer while they still rendered. snaporca-7ih covers
hoisting the registrations so the construction can go too.
Four separators whose groups are now empty were dropped; they rendered as stray rules.
Verified on the rig in both modes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca bcab67f8ce (DesignPanel.cpp applied as a patch; parity 30 / 16, shared
files byte-identical).
tool_atlas.json now names an icon for 80 of 86 verbs, derived from the toolbar's own
definitions rather than invented, and every one of the 54 distinct names was checked to
exist in resources/images first.
The first attempt drew nothing despite a green build: wxGTK builds the GtkMenuItem inside
Append() and reads GetBitmap() there, so setting the bitmap on the returned item is a
silent no-op. append_offer_item() constructs, sets, then appends — the same order Orca's
own append_menu_item() uses.
Verified on the rig.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca 1fb7786d9a (DesignPanel.cpp and DesignCanvas.cpp applied as patches;
parity 30 / 16, shared files byte-identical).
With a sketch open, Text/SVG outlines become ordinary Line entities via push_closed_lines()
instead of a separate Sketch feature carrying rigid imported_regions — so the letters can
be constrained, trimmed and extruded like anything drawn by hand. The buttons and offer
actions arm Select first when in Sketch mode, since begin_sketch() does not run until a
tool is armed.
add_imported_regions() calls reset_autoedit(): without it the glyph contours entered the
draw-then-edit queue and opened a Length field on the first segment, which freezes the
canvas. Caught on the rig, not by reading.
No sketch open: unchanged — a new Sketch feature, still dropped on a picked face.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca 6724ea27c5 (DesignPanel.cpp applied as a patch; parity 30, shared files
byte-identical).
chrome_only's rule is "acts on the DOCUMENT, not on a selection". Text and SVG both call
add_imported_sketch(), which drops the art on a picked solid face via
SketchPlane::from_face() — a selection-consuming profile creator, like Sketch. Now
sk_text / sk_svg in the sketch half's Create row, where their toolbar buttons already sit.
Verified on the rig: Create ends Point, Text, SVG.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca e5e223a794 (DesignPanel.cpp applied as a patch; parity 30, hpp
byte-identical).
Every body combo opened on index 0, so picking a body and pressing Mirror acted on a
different solid while the card showed that other body as the target. Nine sites now read
the viewport selection; Boolean takes the picked body as target and a different one as
tool, since defaulting both to the same body is a no-op.
Verified functionally on the rig: picked the 20x20 body, mirrored, and the new body
measures 20x20 — not the 80x50 one it would have used before.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca e2ef3018cd. Data only, and only the `why` prose in two slots — the
generated DesignOffer.hpp is byte-identical, so there is nothing to rebuild.
Ratified 2026-08-01: the offer deliberately splits the toolbar's dressup family. Shell
hollows a solid so it sits in Remove; Delete Face edits an existing solid so it sits in
Modify. Recorded in both slots' `why` so either half explains the split.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca 352cf1c259 (data only — tool_atlas.json + the regenerated
DesignOffer.hpp; both byte-identical across the forks).
"Fillet / chamfer / draft" rather than naming shell too: Shell is in Remove and Delete
Face in Modify. Only the toolbar's dressup dropdown groups all five.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca 3677937964 (DesignPanel.cpp applied as a patch; parity 30, shared
files byte-identical).
The card header read "Fillet 1" over a chamfer because the offer's Chamfer address opened
the tool before setting the type, and open_tool() titles the card from that combo. Choose
first, then open.
"Dress-up" removed from the offer row (-> "Fillet / chamfer"), the card field (-> "Type",
it was a label reading Dress-up whose value said Chamfer) and the toolbar tooltip.
Verified on the rig: header "Chamfer 1", field "Type: Chamfer", row "Fillet / chamfer".
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca 6d5510734b (DesignPanel.cpp applied as a patch; parity re-checked at
30 lines, shared files byte-identical).
Polygon's Sides/Circumscribed card is deleted — the choice is made in Create > Polygon,
which names the counts and the two fits, because the side count cannot be recovered after
drawing. Dress-up, Combine and Pattern were single verbs hiding several behind a combo
and now name each one in the offer. Fixes fillet and chamfer both carrying key:S+F, which
made the offer's Chamfer open a Fillet.
Built green and verified on the rig: the Dress-up card opened from Chamfer reads Chamfer.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca 0c83f59f13 (DesignPanel.cpp applied as a patch, not copied, so this
fork's 30 permitted divergent lines survive; parity re-checked at 30/16 with the shared
files byte-identical).
The sketch dropdown never registered "fly:<family>#<row>" addresses the way
feat_dropdown does for model verbs, so the offer could name a family but only ever arm
its first tool — Rectangle always gave a corner rectangle. Adds the registration, 14
atlas verbs (including the entire array family, which was absent, and rotate/scale), an
action for the sk_move row that previously did nothing when picked, and a second submenu
level so variants nest under their family instead of flattening 19 create tools.
Built green and verified on the rig: Oblique rectangle arms oblique, not corner.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Mirror of snaporca ac85277bac..0e7cb3ec78 (six changes, applied as a patch to
DesignPanel.cpp rather than copied, so this fork's 30 permitted divergent lines survive
— parity re-checked afterwards: the five shared files are byte-identical, DesignCanvas.cpp
and DesignPanel.cpp differ by exactly 16 and 30 lines).
- The straight slot's inline field says Radius, which is what it sets. It stores the
half-width and passed the typed number through unchanged, so 30 produced a 60 mm slot.
- The offer opens from the keyboard (Menu, Shift+F10), anchored on the viewport rather
than wherever the pointer happens to be. The card hint names the new route.
- The sketch card's Plane combo is gone; the plane comes from the viewport. Also stops
build_candidate collapsing a face plane to a base plane while editing.
- Mass properties and the dead Edit row are wired into the offer; DesignOffer.hpp is
regenerated from tool_atlas.json, verified by re-running the generator and diffing.
- docs/rig_build_traps.md + scripts/rig-build.sh, which derives its fork identity from
project() so it cannot be pointed at the other fork's image or volume.
- docs/design_tab.md refreshed (44 commits stale) + a PR description, with this fork's
own merge-base and diff shape rather than snaporca's.
Built green in the deps container with the new script and verified on the rig: Menu and
Shift+F10 both open the offer at the viewport centre with the pointer parked off-canvas,
and the sketch card shows no Plane row.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Confirming a sketch while an inline value field was open left the field behind. The
editor is a top-level frame, so it survived the session that owned it, and inline_busy
stayed set with it — on_mouse_impl then returned true at its first branch for every
later click and the viewport was simply dead. No refusal, no message: exactly the
"click the geometry, nothing happens" the pick bugs above it were mistaken for.
finish() and cancel() now call close_session_chrome(): dismiss the open field
(keep-as-drawn, the same contract the polyline terminators already use), drop the
queue of fields behind it, and clear the corner readout — which had the same defect
for the same reason, sitting on 336.8° over a committed sketch because nothing redraws
the HUD once the tool stops.
Verified on the rig on the exact reported sequence: line on XZ, Return to accept the
length, Confirm with the Angle field still open. The field goes, the sketch commits,
and the next click reaches the pick (pick trace shows down/up consumed) and selects
Sketch1.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Type a new length into a sketch quote and the number changed while the geometry
sat still, with the solver dropping to "Conflicting constraints" — a broken
constraint state the user never asked for, arrived at by doing the one thing the
status line invited.
An experiment separated the two candidate causes. Same gesture, one variable:
a line committed WITH a driving length refused the edit and conflicted; a line
committed with none (Esc keeps it as drawn) accepted it and visibly shrank. So
the value was never the problem and the solver was not wrong — it was being
handed two contradicting constraints and correctly declining to choose.
The quote-edit path appended unconditionally:
a.con = int(m_constraints.size());
m_constraints.push_back(constraint_for(a));
Accepting the length at draw time creates Distance(P0,P1) = 64.9. Clicking the
quote later creates a SECOND Distance on the same two points asking for 30. Over
-constrained by construction. A line with no dimension yet only ever gets one
constraint, which is exactly why this looked intermittent rather than total.
upsert_constraint() finds an existing constraint with the same type and operands,
overwrites its value and returns its index; it appends only when there is none.
Operand order is ignored — a Distance from A to B is the same constraint as B to
A, and so is an Angle. Returning the index matters as much as the update: it
keeps the annotation's `con` pointing at the constraint that is actually live, so
the NEXT edit is an update too rather than reverting to appending after one good
round. upsert_dimension() applies the same rule to the visible quote, which had
been stacking labels reading different values on the same pixel, and keeps the
existing label position so a placed quote does not teleport.
set_dimension_value already did the right thing through a.con. The machinery
existed; these two call sites never consulted it.
Verified on :11 on the exact failing case — draw a line, Return to lock the
length, Confirm, double-click to re-open, click the line, type 30 into the quote:
the line shrinks, the quote reads 30.0 mm, one label not two, and the solver
stays at "3 degrees of freedom" with no conflict.
NOT included: record_dimension_constraint() has the same unconditional push_back
in all six of its branches. It belongs to the legacy Constrain mode with
different selection semantics and I could not exercise it, so it is flagged
rather than changed blind.
Refs snaporca-e1p.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Selecting a committed sketch line lit the right tree row and then told the user
to go and press Edit in the panel. That is the side-panel dependency this tab
exists to remove, and it made "selectable" true while "editable from the
geometry" stayed false.
on_edit_feature already does the whole job — re-open the entities in the sketch
UI with handles and live quotes — and was only ever reachable from a tree row.
A double-click on a committed stroke now calls it for that feature. Double-click
on empty space still fits the view, so nothing is taken away.
The stroke hit test is now one hit_display_sketch() shared by the click and the
double-click. Two copies of "what is under the pointer" drift, and a double-click
acting on a different entity than the click before it is a miserable thing to
chase. It also reports the entity index, which the tracer prints, so a pick that
lands on the wrong stroke can be seen rather than inferred.
Verified on :11 end to end: draw an open line, commit, double-click it. The
tracer prints "double-click -> edit sketch feature 0 (entity 0)", the panel
reads "Editing sketch — drag a handle or click a quote to edit", and a click
inside the session selects the line with endpoint handles, live quotes and
"1 selected — Delete removes them".
NOT delivered by this commit, found while verifying it: typing a new value into
a length quote is accepted and displayed but the geometry does not move and the
solver drops to "Conflicting constraints". Filed separately — it lives in the
constraint layer, not in selection, and nothing here touches it.
Refs snaporca-e1p.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Reported from the rig: a committed sketch holding a single open line rendered on
the plate and could not be selected, so it could not be edited or deleted from
the viewport at all.
The viewport pick for committed sketches iterated region_loops(). That function
exists to find EXTRUDABLE regions and, as its own walk comment says, an open
chain "stalls" and is discarded. So for a sketch of open entities it returns an
empty list, the pick loop has nothing to iterate, and every click falls through
to bare plate. Not a tolerance problem and not a focus problem: there was no
candidate geometry to test against.
Whether a stroke bounds a closed region has nothing to do with whether the user
can point at it. The stroke test now covers every non-construction entity, and a
separate entity->region map preserves what a hit REPORTS, so a click inside a
closed loop still names that loop exactly as before. Region membership decides
the report, not whether the hit can happen.
The rest of the path was already written and simply unreachable: the panel's
handler has a region < 0 branch that selects the feature, highlights it in the
tree and says "Sketch selected — Extrude it, or Edit / Delete from the tree".
This makes existing behaviour reachable rather than adding new behaviour.
Verified on :11 with the pick tracer (SNAPORCA_PICK_TRACE=1), which is what
distinguished the two failure modes: before, the click reached the handler and
fell through to handle_solid_click ("no solid data"); after, it is consumed by
the display-sketch test and never reaches it, and the panel reads "Sketch
selected" with Sketch1 lit in the tree.
Construction geometry stays unpickable, matching region_loops' own filter. It is
the same class of bug — a construction line cannot be selected to delete it —
but including it risks construction stealing picks from real geometry, so it is
left as a separate decision.
Refs snaporca-e1p.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
contradicting itself about the plane
Reported from the rig: pick a plane, press Sketch, and you are told to pick a
plane. The app prescribed a sequence and then refused to acknowledge that you
had followed it. Right-click did nothing, so there was no way to reach a
drawing tool except the toolbar this tab exists to retire.
act_sketch was two lines: set the mode, then print "Click a face or a reference
plane in the viewport, then a sketch tool" — unconditionally, without ever
asking whether a plane was already chosen. The plane was never lost; m_ref_plane
held it and begin_sketch captures it when the first tool is armed. The sentence
was simply false.
It now asks. sketch_plane_target() is a companion to sketch_plane_from_selection
that distinguishes "the user chose XZ" from "nothing chosen, falling back to XY"
— a distinction m_ref_plane cannot express on its own, being always a valid
index, so m_plane_picked carries it. With a target the readout names it and the
offer opens on the Create row; without one the old prompt stands, because then
it is true. The card above the status line was a local wxStaticText that nothing
could update, so it went on asking for a plane two inches from a line saying the
plane was chosen. It is a member now and the two are written together.
Right-click inside a sketch was excluded wholesale so that it could end a
polyline chain, abandon an anchor, exit a tool. That made every sketch row in
the atlas unreachable. The honest test is not which mode we are in but whether
the tool actually USED this right-click, and only the tool knows: on_mouse now
wraps on_mouse_impl and records that once, for every terminator, instead of
threading a flag through the twenty-odd sites that consume a RightDown. The
canvas read-and-clears it on the matching release.
Underneath all of it was one confusion — MODE versus SESSION — at four sites.
begin_sketch does not run until the first tool is armed, so is_sketching() is
false for exactly the interval between "press Sketch" and "pick a tool", which
is precisely when the drawing tools must be on offer. The keyboard learned this
once already (snaporca-0ud, whose comment states the rule) and I reintroduced it
in offer_selection_kind and again in show_offer_menu, where the offer built from
the FEATURE map and rendered nine rows that all refused the sketch selection.
Both now call sketch_map_applies(), so they cannot drift apart again. The
keyboard keeps its own split: its "sketching" gates undo and delete-last-entity,
which genuinely need a live session.
Verified on :11 against a fresh build. Pick XZ, press Sketch: card reads
"Drawing on XZ", status reads "Sketching on XZ — pick a tool", offer opens with
Create and Reference live and the six rows needing geometry greyed. Right-click
while idle opens the offer. Right-click as a terminator does NOT — the chain
ends, the line lands on XZ, its length field arms at 49.36 mm. That last one is
the regression the blanket exclusion was buying and the reason this shape of fix
was chosen over a mode test.
Refs snaporca-6vs.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
The pick cycle died two commits ago and left no viewport route to a whole
body at all: one click resolves vertex, edge or face, double-click is
already zoom-to-fit, and the only way to take a body was the Bodies list —
a geometry-first violation for as long as it stood. The rubber band is that
route.
Left-drag is the gesture, as asked. That button was orbit, so this canvas
now maps the mouse the way every CAD the user already knows does: left
selects, middle orbits, right pans. The change is a single flag on
GLCanvas3D set only by DesignCanvas, so Prepare and Preview keep the mouse
their users learned. Sketch mode inherits the same mapping, which is the
consistent reading — Design is one modality, not two.
Past an 8 px budget a press becomes a sweep, anchored at the ORIGINAL press
point rather than at the frame where the threshold was crossed, so the
first few pixels are not lost. Below the budget it is still a click and the
existing vertex/edge/face pick runs untouched. Sampling is the display
mesh's triangle vertices plus centroids — the same points the ray pick
tests, already in world coordinates — and the body with the most samples
inside wins, because the selection callback downstream carries one body.
Crossing semantics: touching selects. Enclosed-only for left-to-right and
crossing for right-to-left is the fuller CAD convention and is deferred,
not forgotten; with one selectable body it would have bought nothing.
Two defects fixed on the way, both found by exercising this:
Right-drag pans, and every pan ended by popping the offer over wherever the
camera stopped — the context menu arriving as the reward for moving the
view. The offer is now the release of a STATIONARY right-click, at the same
8 px budget the pick uses.
The selection handler wrote m_status twice. Only the later write ever
reached the screen, so the earlier block had been dead since it was
written, and its labels drifted out of step with the live ones unnoticed —
including a vertex fix I made this morning in the branch that never
renders. Deleted, with a note saying why, rather than left as two writers
for the next person to pick the wrong one.
Verified on :11 against a fresh build: click takes face 5; left-drag across
the body reports "selected (whole body)" with the whole solid tinted and
the camera unmoved; left-drag over empty space clears; stationary
right-click opens the offer; right-drag pans with no menu; middle-drag
orbits. Precedence re-checked after the deletion — face at 25 px from the
corner, vertex from 10 px in.
Refs snaporca-9xw.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
Completes the precedence Tommaso asked for: vertex, then edge, then face, all
from ONE click, all decided in screen pixels. Verified on :11 by sweeping into
the right corner of a plate — x=1250 and 1290 report "face 5 selected", x=1308,
1318 and 1323 report "vertex selected" — and the cyan marker lands exactly on
the corner in the render.
The vertex tolerance (11 px) is deliberately LARGER than the edge one (8 px). A
corner lies ON its edges, so equal radii would make vertices unreachable: every
click near one would resolve to the edge underneath. Bigger-wins-first is what
makes the smallest entity actually pickable.
Vertices come from the sampled edge polylines' endpoints rather than a separate
topology walk — every corner of a face is the end of one of its edges, so the
data was already in hand.
The highlight is a camera-facing square scaled by 1/zoom, the same trick the
edge ribbon uses, so it reads as a constant dot at any zoom. This is why vertex
picking did not ship with the previous commit: the Edge render path billboards
a ribbon and degenerates on a two-point input, and a selection you cannot see
is not a selection (L5). Better to add the primitive than to fake the feature.
DesignPanel now distinguishes level 4: a picked corner sets neither face nor
edge, because a corner is not its face, and the offer classifies it as
OfferSel::Vertex — where Plane, Axis and Coord Sys already accept it.
snaporca-9xw (rubber band still open — see the issue).
Tommaso, correctly: fix selection before building on it. I had taken the
whole→face→edge click cycle as terrain and hung the tool offer off it, when §10
of the charter already listed that cycle as an L5 violation. An offer can only
ever be as truthful as the selection beneath it, so this is the foundation and
it should have come first.
NOW: one click selects the SMALLEST thing under the pointer — the edge if the
cursor is within tolerance of one, otherwise the face. No repeat clicks, no
state, no memory of what was picked before. Verified by sweeping a column of
single clicks down a plate on :11: y=800..915 all report "face 5 selected", and
y=925/935 — within a few pixels of the front edge — report "edge 3 selected".
One gesture, one deterministic result, which is what L5 asks for.
TOLERANCE IS IN SCREEN PIXELS. The old edge step compared a ray-to-segment
distance in millimetres, so the same gesture meant different things at
different zoom levels. The pointer is a screen object; its tolerance has to be
one too. 8 px, measured against the edge polyline projected through the camera.
WHAT IS NOT HERE, AND WHY IT IS NOT FAKED. Whole-body selection has no viewport
gesture in this commit. Double-click is ALREADY zoom-to-fit, bound earlier in
the same on_mouse, and this pick runs on LeftUp where LeftDClick() can never be
true — so a double-click branch here would have been dead code that reads like
a working feature. I wrote one, found it unreachable, and deleted it rather
than leave it. The body gesture is the rubber band, which is its own piece of
work; until it lands bodies are selected from the Bodies list, and the hole is
named in a comment at the site instead of being left for someone to trip over.
Six status strings that promised the cycle ("click again for a face", "click
again for an edge", "click again to reset") are gone — they described a
behaviour that no longer exists, and a hint that lies is worse than none.
Both forks build. Parity: DesignSketchTool.cpp byte-identical, DesignPanel.cpp
30 divergent lines — the invariant exactly.
snaporca-6vs.
Row order is RATIFIED (charter 4.1, 2026-07-31) and this is the first working
implementation of it: right-click in the Design viewport and a vertical list
opens at the pointer with the eight families in their fixed order, the verbs
that apply live, and the ones that do not disabled IN PLACE carrying their
reason.
THE MAP EXISTS ONCE. DesignOffer.hpp is GENERATED from docs/ux/tool_atlas.json
by docs/ux/mockups/gen_offer_table.py — the same file the 113 mockups are drawn
from. A drawing and the product therefore cannot drift apart, which is the only
way row constancy survives contact with a codebase. Never hand-edit the header.
NOTHING IS RE-IMPLEMENTED. Each row routes to the code that already runs that
verb: "key:S+E" through m_keys_feature, "key:L" through m_keys_sketch,
"fly:material#4" through the feature flyout's own action, "btn:colour" through
the standalone button. The offer is a second door onto the same room, so the
toolbar, the shortcuts and the menu cannot drift into three behaviours. The 8
verbs with kernel support but no GUI path show disabled, which is honest and
matches section 10 of the charter.
Right-click only fires the offer when the canvas is IDLE. Right-click already
ends a polyline chain and finishes the move gizmo; taking those over would
break two working interactions to add a third.
Two things the running build corrected, both found by looking at screenshots:
- THE REASON MUST BE TRUE FOR WHAT IS IN FRONT OF THE USER. Taking the first
refusal in a family printed "Transform needs a body — add or import one
first" on a document that HAS a body, because the real obstacle was that
nothing was selected. Now the reason comes from a verb that accepts the
current selection and fails only on document state; if no verb in the family
accepts this selection at all, it says "select something first" or says
nothing. A menu whose whole value is telling the truth cannot ship a lie.
- Classification follows the level the pick cycle has REACHED, not the face the
ray happened to hit, so the header cannot name a face while the whole body is
lit. Sketching on the face you merely clicked is untouched — that path is
sketch_plane_from_selection (snaporca-3a2).
Verified on :11 end to end: nothing selected shows Sketch live with Shift+S and
seven greyed rows each explaining itself; a selected solid shows Move directly
with Shift+Y (one applicable verb, so no submenu and no extra click) and five
families as submenus. Both forks compile and link.
Fork parity re-checked after the port: DesignPanel.cpp 30 divergent lines,
DesignCanvas.cpp 16, every other CAD file byte-identical — the invariant exactly.
snaporca-96r.
Folds the form-factor decision into the charter. 4.1 is rewritten around
Tommaso's proposal — left-click selects, right-click opens a vertical list of
icon / name / shortcut — and the radial is demoted to a "Rejected" subsection
rather than deleted, because it is a good idea that loses on evidence and
somebody will propose it again. The evidence is recorded with it: mean fill of
3.45 of 8, only two live slots on a fresh document, sketch Create needing nine
addresses on an eight-slot ring, names that do not fit around a circle in
translation, and a 380px disc over the model on a 1366x768 screen.
The invariant survives the change of geometry, which is the useful proof: same
eight families, same fixed order, nothing re-sorted or compacted. It is now
stated as ROW constancy, and the one substantive gain is that unavailable verbs
are disabled IN PLACE carrying their own reason, in strings the product already
ships. An empty ring slot was mute; a greyed row teaches. On a first-run
document the offer stops being a mostly-blank control and becomes a map of what
the product does and what you must do first — which is the section 2 audience
in one picture.
Opening it is now a table rather than an assumption, because "right-click" is
not a universal gesture: two-button mouse right-clicks, trackpads two-finger
tap, a one-button Mac LONG-PRESSES or Ctrl-clicks, touch and pen long-press,
and the keyboard uses the Menu key or Shift+F10. The long-press is explicitly
an ADDITIONAL route — 6.2 forbids press-and-hold as a sole path and that
stands, so the rule now names its own exception and closes it — and it must
show that it is charging, or a user who lets go early concludes the product is
broken (L5).
Consequently: 6.2's keyboard bullet describes opening and walking the offer by
key rather than by compass direction; the gate gains question 13 (every new
pointer gesture declares its keyboard equivalent and what a one-button Mac, a
trackpad and a touch screen do) and question 12 now asks whether tool_atlas.json
was updated and the atlas regenerated; section 10 points at the rendered atlas
and says the outstanding thing is ratifying row order, not drawing the map.
snaporca-96r.
Tommaso was not sure about the ring and proposed a vertical list: left-click
selects, right-click exposes icon / name / shortcut. Drawn, it is better, and
the reasons are visible in the renders rather than arguable.
THE DISABLED ROW CAN SPEAK. This is the one that decides it. A ring slot that
does not apply is an empty circle: it says nothing, and on a fresh document six
of the eight are empty. A list row that does not apply is greyed IN PLACE with
its own name and its own reason — "Create a sketch, or pick a solid face,
first", "Create a solid body to pattern first" — which are strings the product
already ships and which tool_atlas.json already carries. The first-run picture
stops being a mostly-empty ring and becomes a map of what the product does and
what you must do first. For the audience section 2 puts first, that is the
whole ballgame.
THE OVERFLOW DISAPPEARS. Sketch Create needs nine addresses; a ring of eight
pushed Polygon and Point behind a "More" slot. Nine rows is just nine rows. The
one measured defect in the ring design is not a defect in this one.
SHORTCUTS READ AS A COLUMN. Right-aligned in a list they stack into something
the eye learns passively, which is exactly the graduation path 4.1 claims —
and it is the mechanism by which the power user Tommaso describes stops opening
the menu at all. Around a ring the same keys are eight loose chips.
Also, unglamorously: long translated names fit, arrow keys and screen readers
work natively where a radial needs special handling, and a 324px box costs the
1366x768 machine far less than a 380px disc over the model.
What the ring keeps: equidistant targets and a future flick gesture. Since the
brief is that power users live on the keyboard, that buys less than it looks.
The invariant is untouched — same eight families, same fixed order, nothing
re-sorted, nothing compacted. Only the geometry changed, which is the point:
the map survived a change of form factor, so it was a real map.
Both forms are now rendered side by side for the same states, and the atlas
opens with the pairs.
snaporca-96r.
The charter fixed the slot-constancy invariant but carried one hand-written
eight-cell table as an illustration, and nothing of the offer exists in the
product. Before any GUI code, the group needs the map itself: what verbs there
are, what each one needs before it can be offered, and what the ring actually
looks like in every situation a user can put it in.
tool_atlas.json is the source of truth and it was extracted from the code, not
from memory: verbs, shortcuts and the exact refusal strings from DesignPanel's
six feat_dropdown call sites and the sk_key table, kernel coverage checked
against CadFeatureType, headless coverage against McpControl's dispatch, and
the selection kinds taken from the callbacks DesignCanvas actually exposes. It
carries each verb's preconditions, so "why is that slot empty" has an answer
already written in the product's own words. The eventual C++ table generates
from this file too — the map exists once.
gen_offer_mockups.py renders it: 20 selection kinds x 2 document states = 40
primary rings, plus 73 sub-rings, plus comparison sheets. 113 states, none of
them hand-drawn, because a human drawing 113 rings is exactly how an address
quietly changes. Everything is framed at 1366x768, the charter's own reach
target, so L11 is tested in the mockups before it is tested in code.
Three things the drawing found that the prose had not:
- MEAN FILL IS 3.45 OF 8. The empty-slot rule is cheap in argument and
expensive on screen; on a fresh document exactly two slots are live. That
picture is the anti-clutter thesis made literal and it is the strongest image
in the set.
- SUB-RINGS MUST ANCHOR ON THEIR PARENT. Fanning them from north put Extrude at
N, which is Create's address in the primary map, so the second level
contradicted the first. Anchored, an address is two consistent strokes: Add
material is NE and its first verb is NE again.
- ONE FAMILY OVERFLOWS, AND ONLY ONE. Sketch-mode Create needs nine addresses
on an eight-slot ring. That is the ninth-position pressure the charter
predicted, arriving on schedule and measured rather than argued: either Point
moves family, or the tail goes to a third level, or the ring is not eight.
Every model-mode family fits. The generator refuses to wrap a tenth verb onto
a first — silent collision is the one outcome worse than an ugly ring — and
reports the overflow instead.
Also fixed while looking at renders: the selection pill sat on top of the north
slot's shortcut chip and hid it, and the scrim at 0.55 swallowed the very face
the ring had been opened on, which is 4.1 failing inside its own mockup.
Fork-neutral: the generator and everything it emits name no product, so both
forks carry byte-identical copies.
snaporca-2is.
Tommaso's requirement, and it changes what the offer IS: a tool must sit in the
same physical position whatever you selected. Click a face, an edge or a text
and fillet is in fillet's place every time. Position becomes an address the hand
learns, and the eye stops being needed.
That kills the ordering rule this section had two commits ago. "Most-used first
for that kind of selection" is adaptive ordering, and adaptive ordering destroys
the one property that makes a spatial menu fast — worse, it destroys it exactly
for the user who has just started to learn the layout. Office 2000 shipped that
idea and withdrew it. So: NO adaptive ordering, ever, in any form.
The invariant, written to survive every future feature: every tool has exactly
one address; that address is identical in every selection type where the tool
appears; slots for inapplicable tools are left EMPTY rather than compacted; and
adding a tool never re-addresses an existing one. Empty slots are the price of
constancy and they are cheap — a compacted offer is denser and unlearnable, a
sparse one is memorised in a week. An empty slot also answers a question ("this
cannot be done to this thing") that a silently-inert tool does not.
Radial rather than a strip, reversing what I proposed last time and for a reason
that only appears once constancy is the requirement: a direction from the click
point is an absolute address that survives the offer opening anywhere on screen
and survives being clamped at a screen edge, while "third item down" does not.
Centre is a hole so the picked geometry stays visible, and it names what is
selected, so a mis-pick is caught before a verb is chosen.
Every slot carries its keyboard shortcut beside the icon and the word. This is
the graduation path and it is why power users never see a conflict: you reach
for the place, the place says "F", and one day your hand types F before the ring
finishes drawing. The offer is the mechanism by which a beginner stops needing
the offer — one interface at two speeds, no advanced mode in between.
Also here: a proposed eight-position compass map across face/edge/body/text
(create, add, remove, dress-up, repeat, transform, reference, modify) offered as
the group's first ratification, with families opening a secondary ring under the
same rule; arrow/numpad direction addressing so the spatial map works from the
keyboard; a gate question 12 that treats re-addressing an existing tool as a
breaking change to every user's muscle memory.
snaporca-2is.
Two changes to section 4, both from Tommaso.
FIRST: the selection does not merely feed the tool, it DETERMINES WHICH TOOLS
EXIST. Point at a planar face and the product offers the small set of things a
planar face can become; point at an edge and it offers fillet, chamfer and the
sketch tools that can reference it. Nothing else, because nothing else is
possible. This is the largest single thing available to us for a first-time
user, and the reason is worth writing down: a beginner's difficulty is not
operating a tool, it is not knowing which tools apply to what they are looking
at. Sixty icons answer a question they cannot yet ask; a face that offers its
own five verbs teaches the product by being used. It also deletes a whole class
of failure — a tool that silently does nothing because the selection was wrong
becomes unreachable.
The offer is an accelerator, not a toll gate: toolbar and single-letter
shortcuts keep working unchanged and consume the same selection, so an expert
never looks at the offer and a beginner never needs the toolbar. Both routes
land in the same place, which is how one interface serves all three audiences.
SECOND: "click empty space to commit" is withdrawn. It was an invisible gesture
with a destructive meaning — nothing on screen said it, and a stray click
committed a feature still being adjusted. Exactly what L5 forbids. A pending
feature now carries a confirm/cancel puck attached to its own geometry, beside
its handles, with Enter/Escape mirroring it; empty space reverts to the safe
meaning, clear the selection. The puck is an object in the scene, not a dialog:
the camera orbits, the values stay editable, nothing is blocked (L4 intact).
Two cases the rule has to get right or it damages the inner loop: continuous
tools (line, rectangle, circle) still commit each entity on its own gesture — a
tick per line would be miserable — and Enter/Escape end the tool rather than
confirm an entity. And ambiguity resolves toward keeping work: starting another
operation with a valid feature pending commits it rather than discarding it,
because undo reaches everything and the recoverable direction is the right
default.
Section 10 gains the two honest consequences: today a selection offers nothing
(the largest single item of new work this charter asks for) and committing is
still the invisible empty-space click.
snaporca-2is.
The charter had accessibility only in the assistive sense — keyboard, contrast,
colour, targets — and said nothing about who can get through the door in the
first place. That was the larger omission. The premise of an OSS CAD tool is
that a kid on a school laptop, with no licence, no account, no fast machine and
nobody to teach them, can open it and make a real thing; a tool that only the
equipped can run reaches people who were already going to design something.
So the fourteen-year-old is now the FIRST of three audiences, ahead of the maker
and the mechanical designer, with an explicit rule that when audiences conflict
the earlier one wins unless someone writes down why not.
L11 states the floor: runs completely on a low-end laptop with integrated
graphics at 1366x768, offline, no account, and no capability withheld behind a
tier, a plugin or a cloud service. Section 6 splits into 6.1 reach and 6.2 the
assistive floor: the reference machine, the small screen as the layout target
rather than the stretch case, files that belong to the user, learnable with no
documentation, plain language at the entry tier, and exploration that is never
punished — undo reaches everything, nothing asks the user to be sure.
Consequences elsewhere: the screen budget is set by the smallest screen we
serve, not the reviewer's monitor; the PR gate gains a reach question; B6 joins
the benchmark (the inner loop on the reference machine, offline, fresh install)
and every task is measured there rather than on a workstation. The side-panel
debt now fails L11 as well as L1 — on that screen the cards leave the model a
strip.
One role addition: the absent audience needs a seat. The kid cannot file an
issue, so someone owns B5/B6 and the group watches real first-timers quarterly.
Approachability is the one thing here that cannot be argued from principle.
snaporca-2is.
The doctrine is shared but the document is not: each fork's copy now speaks
about its own product only, so it reads as that project's own charter rather
than as a note about a sibling repository. This is a deliberate divergence —
the two copies must NOT be reconciled by a parity sweep. The CAD sources stay
byte-identical; only this doc branches.
snaporca-2is.
The call with SoftFever settled that Orca-CAD is one of the branches to be
implemented and that a design+dev group forms around it. A group without a
written doctrine reviews by taste, and a CAD reviewed by taste becomes
FreeCAD one locally-reasonable side panel at a time.
So the doctrine is written first, as something a reviewer can FAIL a pull
request against: ten laws each with its own test, the interaction grammar
they compose into, the accessibility floor as a merge requirement, and a
ten-question gate answered in every UI pull request.
The position is Shapr3D's interaction economy, not its feature list —
direct, gestural, almost no chrome, depth revealed by what you touch. Depth
for mechanical designers arrives as progressive disclosure of tools that
never move, in three tiers, non-modal, with assemblies and exploded views
obeying the same point-then-act grammar as a beginner's extrude.
The one thing neither Shapr3D nor FreeCAD has is that we live inside a
slicer: plate, nozzle, material and build volume are known at design time,
so print-domain failures are warnings on the geometry, not a report.
Section 10 is an honest inventory: what already complies, and the six
things that violate the laws today, none of them defended. The appendix
keeps the anti-patterns we have already paid for, because each one is
cheap to reintroduce.
snaporca-2is.
The relaunch sequence was an ad-hoc pile of docker exec one-liners, and it
had a real bug: it dismissed the first-run dialogs by computing the
titlebar close box from `xdotool getwindowgeometry --shell` and clicking
it. When the dialog had already closed, that eval left the geometry
variables stale or empty, the click landed at a garbage coordinate, and it
kept hitting the Sketch button in the toolbar underneath — so the app came
up in sketch mode with a stray Sketch feature that then had to be
cancelled by hand. Three times in one session.
The fix is not a different mechanism. `xdotool windowclose` looks cleaner
and KILLS THE APP: it destroys the GdkWindow out from under the dialog and
the process dies with "GdkWindow unexpectedly destroyed", three
GLib-GObject criticals and a segfault. Measured, not guessed — that is
what the first version of this script did. Escape does not close the Setup
Wizard either, which is why it needs handling at all. So the titlebar
click stays, and what changes is that it refuses to click geometry it has
not validated: the window id is re-resolved immediately before, all four
geometry variables are unset first and must come back numeric, and the
computed point must be inside the screen. Any of those failing logs why
and clicks nothing.
Two further honesty fixes in the status output, both caught by reading it
rather than by it failing: app_pid skipped nothing, so with a container
full of <defunct> instances it printed a dead pid as though the session
were healthy — it now walks /proc/<pid>/stat and ignores zombies. And a
container without x11vnc reported "vnc: DOWN" as if something had broken,
when nothing was ever installed; it now says so, and does not try to start
what is not there.
Also replaces the fixed post-launch sleep with a wait for the main window,
because cold starts under software GL vary by a lot, and adds --status for
diagnosis: "no windows but the desktop is up" means the app died, "cannot
connect at all" means the desktop did. That distinction cost real time to
work out by hand.
Verified on both containers: one run each, wizard closed on validated
geometry, no stray sketch mode, live pid reported, and the app still up.
snaporca-e1p adjacent (tooling, not the tab itself).
The plane combo is gone. A sketch takes its plane from what is picked in
the viewport: a face on a solid, or one of the reference-plane ghosts,
clicked in 3D. The card is now a single line of instruction instead of a
control.
The combo had become worse than redundant. Once a picked face could be
the plane it displayed a row that CONTRADICTED the actual target — it
still said XY while the sketch went onto the face — so the one place a
user could look to confirm where they were drawing was the one place
guaranteed to be wrong.
What replaces it is state, not UI: m_ref_plane records which reference
plane was last clicked in 3D (0/1/2 = XY/XZ/YZ, >=3 indexes the datums)
and ref_plane_name() turns it into text for the on-geometry hint. Clicking
a ghost plane while a session is live re-planes it immediately, which the
combo's own handler used to do; that behaviour is kept, just driven from
the geometry instead of the widget. A plane click also drops a stale face
pick, so last pick wins in both directions.
populate_plane_choices() stays — seven other pickers use it (Plane base,
Axis A/B, Helix, Project, Mirror, Cut). Those are the next candidates,
tracked on snaporca-e1p; this commit only removes the one that had become
actively misleading.
Also: tessellation now matches Orca's OWN STEP importer, linear deflection
0.003 instead of 0.01 (Format/STEP.hpp default; angular was already 0.5
rad and unchanged). The Design viewport was never using a different
rendering technique — it hosts a real GLCanvas3D, builds a real
Model/ModelVolume and goes through the same reload/GLVolume path and the
same shaders as Prepare and Preview. What differed was the mesh handed to
it: 3.3x coarser than anything else in the application, which is why a
curved face read as faceted beside an imported part. Suite unaffected at
154 cases / 2125 assertions, so nothing depended on the old density.
Verified on :10: the card shows no dropdown, one click on a face then a
sketch tool still reports "on the picked face", and the circle is drawn in
that face's plane (artifacts/shots/h3a2-02-sketch.png, h3a2-03-drawn.png).
snaporca-e1p, snaporca-3a2.
The previous commit made a picked face the sketch plane and I verified it
by clicking the face TWICE. That was the wrong test. handle_solid_click
cycles whole -> face -> edge, and "First click on a (new) body/face
selects the WHOLE solid; refine on repeat clicks" — so at level 1
m_sel_solid_face is -1, and one click on a face, which is what selecting
a face means to anyone, still fell through to the plane combo. The fix
was real and unreachable, which from the outside is indistinguishable
from no fix at all.
The face id was never missing. handle_solid_click resolves it by ray on
the FIRST click and passes it to on_solid_selection_changed regardless of
the cycle level; the panel simply discarded it whenever level < 2. Keep
it in m_pick_face/m_pick_face_body and let a sketch use it, preferring an
explicit face-level selection when there is one. Nothing about the cycle
changes, so body operations that rely on whole-body selection are
untouched.
The new state is dropped wherever the existing picks are, so a stale face
cannot come back: choosing a body from the Bodies list (an explicit
choice with nothing pointed at), picking a committed sketch loop (last
pick wins), undo/redo (recompute invalidates topology ids), and when a
sketch consumes the face.
Verified on :10 with ONE click, which is the flow that was broken: build
a box, single-click its top face, S then C, and the hint reads "Circle —
click center, then radius · on the picked face" with the circle drawn in
that face's plane (artifacts/shots/g3a2-01-one-click.png,
g3a2-02-sketch.png, g3a2-03-drawn.png).
GUI-only, so the kernel suite is unaffected — plane_of_face and its 154
cases / 2125 assertions are unchanged from the previous commit.
snaporca-3a2.
Selecting a face and sketching on it is the most common gesture in solid
modelling, and it was impossible. The plane came from a combo holding
XY/XZ/YZ plus datums, and plane_from_choice had no face branch at all, so
the only route onto a face was to build a Coincident datum plane on it
first, confirm that, reopen the sketch and find the datum in the
dropdown. Three extra steps and a junk feature in the tree.
The fix is not another combo row. A new sketch now takes its plane from
what is SELECTED IN THE VIEWPORT: a picked planar face wins outright, and
only when nothing is picked does it fall back to the reference plane —
which is itself normally set by clicking one of the ghost planes in 3D,
not by opening the combo. The tool hint names the target ("Circle — click
center, then radius · on the picked face") so the choice is visible on the
geometry side rather than needing a control to read back.
CadDocument::plane_of_face is the shared derivation, so the sketch path
and the Coincident datum method cannot drift apart. It refuses
non-planar faces: face_normal_world evaluates at the mid parameter, which
on a cylinder or a fillet is a tangent plane at one arbitrary point —
fine for offsetting a datum, wrong as a sketch plane, and silently
sketching on a tangent is worse than declining.
Picking the face also CONSUMES it. Leaving the pick live meant the next
Extrude saw a selected face and push/pulled it instead of extruding the
sketch just drawn — the same trap the imported-art path already guards
against.
Verified on :10 end to end with no combo interaction: build a box, click
its top face twice to cycle whole -> face, press S then C, and the circle
is drawn in the plane of that face with its Radius tab on the geometry
(artifacts/shots/f3a2-03-face.png, f3a2-04-sketch-on-face.png,
f3a2-05-circle-drawn.png). Kernel side: 154 cases / 2125 assertions green
on both forks, including that a cylinder resolves exactly its two flat
caps and refuses the barrel.
Still side-panel-shaped and to be dealt with separately: the Plane combo
remains on the card and now merely displays a stale row when a face is
the real target. It should show the actual target or go away.
snaporca-3a2.
entities_to_wire handles exactly two shapes: one lone closed entity
(Circle/Ellipse), or a chain of open ones (Line/Arc/EllipseArc/BSpline).
Anything else -- a circle coexisting with a line, two circles -- returns a
null wire. build_sketch_wire answered that by falling through to its
legacy tail, which ends in a rectangle built from width/height. For an
entity sketch those fields are whatever they were initialised to, so the
extrude produced a box the user never drew, silently and with ok:true.
That is how the ellipse+stray-arc case in the P2 Tier-B.1 verification
turned into a default-rectangle solid.
Throw there instead. The legacy profile/shape paths below are still
reached by sketches that legitimately carry no entities at all, so the
enum and profile constructors are untouched -- only the case where
entities exist and cannot be turned into a wire now fails, which is
exactly the case that was fabricating geometry.
This does NOT implement the multi-loop support the issue asks for. Doing
that properly means deciding containment -- a circle inside a rectangle is
a hole, a circle beside it is a second region -- and make_extrude_regions
cannot be reused because it takes flattened Vec2d contours for imported
Text/SVG art and would discard the analytic circle. Guessing containment
would trade a visible failure for a wrong solid, which is the opposite of
the point. Left scoped on snaporca-88v.
Also converts the three float comparisons in the two test cases added this
session from Approx to WithinAbs/WithinRel, per tests/CLAUDE.md, which
rules Approx out for being asymmetric and double-only. The rest of the
file's pre-existing Approx uses are left alone.
153 cases / 2090 assertions green on both forks; no existing test depended
on the default-rectangle fallback.
snaporca-88v (partial: the silent-fallback half).
A boolean cut whose tool misses the target is a perfectly legal
operation: OCCT reports IsDone(), the shape comes back unchanged, and the
feature lands in the recipe reporting ok:true. Driving the MCP socket,
that produced two consecutive {ok: true, bodies: 1, error: ''} responses
for a hole that was never drilled -- same viewport, same 46939.11 mm3 --
and the tree grew two Hole features that will never cut anything. A
caller, an agent especially, has no signal at all that the thing it asked
for did not happen.
Measure the volume across the op in route_feature's in-place branch and
refuse the no-op. Only for removals: Hole, Thread, and Extrude / Revolve
/ Sweep / Loft in Cut mode. Everything else may legitimately leave the
volume alone -- a Transform certainly does. The tolerance is relative,
because an absolute epsilon is wrong across the mm-to-metre range of real
parts, and a cut that shaves a numerically invisible sliver is a miss
too. The existing rollback in the MCP actions already preserves the
reason, so a missed hole now answers ok:false with the error and undoes
the feature.
The confusion underneath was not itself a bug: hole's x/y are in the
sketch plane's frame, whose origin is describe_scene's modeling_origin,
and describe_tools documented them only as "number, unit mm". Passing the
world centre put the hole 135 mm clear of the solid. All six x/y params
on hole / hole_styled / hole_standard now say which frame they are in,
since the wrong guess was silent.
Regression test drives the reported failure directly: a hole at x=135 on
a 20x20x20 box is rejected and leaves the body untouched, the same hole
at the origin still removes exactly pi*4^2*20, and a cut-mode extrude
whose profile sits at x=200 is rejected too. 152 cases / 2083 assertions
green on both forks.
snaporca-daf.
All 17 single-letter sketch shortcuts were dead. The dispatch gate was
circular: the sketch key map was consulted only when
m_viewport->is_sketching() was already true, but is_sketching() is a
whole-session flag whose only riser is begin_sketch(), called from
select_tool() -- which is precisely what every sketch key closure calls.
So the first letter after entering sketch mode fell through to the
feature map, where the keys are Shift+letter, matched nothing, and did
nothing. The mouse worked only because the toolbar flyout row reaches
select_tool() directly, bypassing the gate.
Which key MAP applies is a question about the mode. Split the flag:
'sketching' (live session) still drives undo/redo, Delete and the
section-view branch, where a session genuinely has to exist; a new
'sketch_mode' (m_ui_mode == UiMode::Sketch alone) drives the map choice.
Verified headless end to end, keyboard only: Shift+S, then R draws a
143.4 x 133.7 rectangle on XY reporting 4 degrees of freedom, then F and
L arm Fillet and Line (artifacts/shots/0udb-01..03). Note the toolbar
strip does NOT change when a tool is armed by keyboard -- the family
buttons are flyouts and only show their own pressed state -- so the
pixel diff on that strip, which is how this was originally measured,
reads 0 for a tool that is live. The status line is the surface that
actually reflects the armed tool.
Also fixed, from snaporca-d9i's list: the plane-pick status line said
"press Sketch to draw on it", naming a button that exists only in
Feature mode. It is now mode-aware.
Adds a SNAPORCA_KEYTRACE=1 trace in the CHAR_HOOK printing key, ui mode,
is_sketching, in_text and the focused window's class. It is what
separated "the fix does not work" from "the surface being measured never
moves", and it costs a full GUI build to re-add, so it stays.
snaporca-0ud, partial snaporca-d9i.
Two P0s in the same gesture. Filleting a corner of a parametric rectangle
produced either nothing at all or a sharp corner with a stray arc floating
above it.
Trigger (snaporca-cq2): the only routes that ever reached confirm_op were
finishing the whole sketch and an unsignposted click on empty space.
set_tool() dropped a ready op, so typing a radius or dragging the arrow and
then touching any other tool threw the value away. Commit a ready op on tool
change (before m_mode is reassigned — op_ready() and confirm_op() both switch
on it), commit on Enter in the radius editor, and drop the pending op before
Esc's tool downgrade so Esc still cancels rather than applies.
Substitution (snaporca-pl5): libslvs writes its last Newton iterate into the
params whether or not it converged, and SketchSolver read them back
unconditionally, so every REJECTED solve deformed the sketch. The fillet
ladder tries a deliberately over-constrained rung first (a tangent on each
leg, against the legs' own H/V); it is correctly rejected, but its wreckage
then failed rungs 2 and 3, which solve cleanly on their own. The arc ended up
with no constraints at all, the rigid loop won, and the corner snapped shut.
Measured: from pristine geometry rung 1 gives result=INCONSISTENT with 3 bad
constraints, rung 2 gives dof=6 with the arc's radius intact.
Read the geometry back only on success. try_add_constraints then needs no
"restore" re-solve — the entities still hold the prior solved state.
Kernel suite 151 cases / 2072 assertions green on both forks; the GUI check
ran on the Snapmaker fork (9d72c4377a).
Ported from the Snapmaker fork. snaporca-pl5 snaporca-cq2
The card decides between a single picked edge and a whole face-group from a
viewport pick it never mentioned. With no edge picked the user saw only the
group combo and concluded per-edge rounding did not exist; with an edge picked
the combo still read "All" — the opposite of what Confirm would do.
Adds a Target row that names the actual target and greys the group combo out
while an edge is picked, wired at the same three points Shell already uses:
the selection-changed handler, the re-edit load, and open_tool.
Ported from the Snapmaker fork (33771a0e95). snaporca-40d
The CAD sources are meant to be byte-identical across the two forks, with exactly
two permitted divergences: DesignPanel.cpp's flyout plumbing (30 lines — mainline's
DropDown is Item-based where the other fork takes three parallel vectors) and
DesignCanvas.cpp's Bed3D::set_shape signature (16 lines). DesignPanel.cpp had drifted
to 105.
Nothing failed to announce this. The kernel suite does not compile the GUI, and all
four defects are interaction-level, so both forks stayed green while only one of them
had the fixes. The parity diff is what found it.
* combo_append_index and its five call sites. Orca's ComboBox keeps client data in
its own vector, so Append()'s clientData argument never reaches wxItemContainer
and m_clientDataItemsType stays wxClientData_None. GetClientData() opens with a
wxCHECK_MSG, which is an early return — so every read came back NULL and every
caller resolved it to index 0, silently, because 0 is a legal answer. Affects the
rib sketch picker, the sheet-body picker, both mate coordinate-system pickers and
the sweep path picker.
* Wrap(240) over the card labels. wxStaticText never wraps itself, so a one-sentence
hint sets its card's minimum width to the width of the whole sentence and every
control in that card is clipped at the sidebar's right edge.
* Shell and Draft face-label initialisation in open_tool. Both read the face from the
live pick at Confirm time, but only the pick handler wrote their labels, so picking
a face and then opening the card left the card describing one operation while
Confirm performed another.
* SurfaceOffset and ThickenSurface added to build_candidate's negative list. Both read
a sheet body from their own combo and both had it overwritten by the picked solid's
index. The list is a negative one, so a tool that picks its own body breaks by
omission — noted in a comment now.
Also drops a stray <cstdio> left behind by a debug probe.
Verified by building the GUI here for the first time since these landed: 461/461,
liblibslic3r_gui.a links. DesignPanel.cpp is back to exactly 30 divergent lines and
every other CAD file is byte-identical again.
snaporca-7xx snaporca-aqu snaporca-gu9 snaporca-c03 snaporca-5pl
The inline editor special-cased Escape and Skip()ped every other key, so Tab fell
through to wx's default navigation. Its popup frame holds exactly one control, so
focus came straight back to that control with its text re-selected.
Type 60, Tab, 40, Enter — expecting to fill two dimensions — and the 60 is gone: Tab
neither committed it nor advanced, so the 40 just replaced the re-selected text. A
re-selected field is pixel-identical to a freshly opened one, so nothing on screen says
a number was dropped. Tab-to-next-dimension is what Onshape, SolidWorks and Fusion do,
which is exactly why it is the key a user reaches for.
Tab now calls do_commit(), the same path Enter takes; the caller's on_commit is already
what walks to the next dimension. Verified by driving the GUI: 37 Tab 24 Enter now
produces a 37.0 x 24.0 rectangle, where before it produced 24 and a mouse-derived value.
snaporca-xah
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
A filleted solid arrived on the plate as a broken model: the slicer reported
"8 non-manifold edges" on an 80x50x12 box with r=3 on all edges, and advised
repairing it in another CAD application -- the exact round trip the Design tab
exists to remove. The same box without the fillet committed cleanly.
Measured rather than guessed. Each of the 8 bad edges is degenerate, both
endpoints the same vertex:
open tri=145 edge=1 face=5 v59(3.000000 3.000000 0.000000) v59(3.000000 3.000000 0.000000)
open tri=538 edge=1 face=6 v87(3.000000 3.000000 12.000000) v87(...)
... one per corner, 8 corners
OCCT triangulates a degenerate surface parameterization with a triangle at the
pole; a corner sphere patch has exactly one. Its two pole nodes are distinct in
the per-face triangulation and collapse to a single vertex when the faces are
welded, leaving a zero-area triangle whose v->v edge can never pair with a
neighbour. its_face_neighbors counts it as open, and the field the object panel
prints as "non-manifold edges" is in fact stats.open_edges.
So the geometry was never wrong -- the B-rep volume matches the Steiner formula
for a box dilated by a ball to 0.016%. Only the bookkeeping was.
Dropping those triangles after the weld removes 8 of 3492 and takes open_edges
to 0. Zero area, so nothing about the shape changes. tri_face is compacted in
the same pass, since it must stay index-aligned with the triangle list that the
face picking and per-body colouring both index into.
Guarded by a new [CadDocument] case that asserts open_edges == 0, no degenerate
triangle survives, and both per-triangle maps still match the triangle count.
The existing suite only ever checked B-rep volumes and areas, which is why a
mesh defect this visible went unnoticed: 150 cases / 2049 assertions green on
both forks.
snaporca-agw
OCCT_LIBS is an explicit single-pass static link order — dependents first, TKernel
deliberately last. The CAD block appended its two extra toolkits to the END of that list,
which puts them after everything they depend on:
list(APPEND OCCT_LIBS TKFillet TKOffset)
TKOffset references BRepAlgo_Loop, and nm against the built deps prefix shows TKBool is the
only toolkit that defines it (TKTopAlgo, TKBO, TKPrim, TKFillet and TKOffset all define it
zero times). TKBool sits first in the list, so a single-pass linker has passed it long before
it reaches the appended TKOffset and will not go back:
libTKOffset.a(BRepOffset_MakeLoops.cxx.o): undefined reference to
BRepAlgo_Loop::BRepAlgo_Loop()
Only one configuration ever objected — the Snapmaker fork Flatpak (aarch64). Ordinary Linux,
macOS and Windows links resolve it regardless, and the mainline fork Flatpaks pass, so six
green platform legs said nothing about whether this list was correct.
Prepended via set() rather than list(PREPEND), which needs CMake 3.15 while this project
supports 3.13.
Worth knowing for later: TKFillet and TKOffset are mutually dependent, 20 symbols needed in
each direction, so a stricter single-pass link could still trip on that pair. It does not on
any current platform, so no --start-group or duplicate entry is added here; if something ever
complains about ChFi or BRepFill symbols, that cycle is the reason.
snaporca-2kj
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
The header forward-declares a long list of wx types and omitted wxBoxSizer, wxTextCtrl and
wxListCtrl. All three are used as pointer members only (m_expr_text, m_var_list,
m_parts_hdr, m_hdr_tree_row), so a forward declaration is all they need — but there was
none. Every ordinary build compiled anyway because the wx/panel.h + wx/scrolwin.h chain
happens to pull the real headers in transitively.
The Snapmaker fork Flatpak build (aarch64) has a wx that does not, and it failed outright:
DesignPanel.hpp:511: error: 'wxTextCtrl' does not name a type; did you mean 'wxTreeCtrl'?
DesignPanel.hpp:521: error: 'wxListCtrl' does not name a type; did you mean 'wxFileCtrl'?
DesignPanel.hpp:663: error: 'wxBoxSizer' does not name a type; did you mean 'wxSizer'?
plus a cascade of "m_var_list / m_expr_text / m_parts_hdr was not declared in this scope".
Not an environment quirk: the header was simply not self-contained, which is exactly what
breaks a reviewer building in an unfamiliar configuration. The mainline fork Flatpaks passed
on both arches, so only that one manifest exposed it.
Audited the rest of the header afterwards: every other wx pointer type is either
forward-declared or genuinely included — only wxScrolledWindow and wxWindow are undeclared,
and both come from the real wx/scrolwin.h include.
snaporca-4dn
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
store_bbs_3mf reaches Model::get_backup_path(), which builds
temporary_dir() + "/orcaslicer_model/" + timestamp. temporary_dir() returns a file-static
that ONLY OrcaSlicer.cpp's startup sets, so in a test binary it is the empty string and the
backup path becomes "/orcaslicer_model/..." — absolute, at the filesystem root. An
unprivileged process cannot create that, so the save returned false and the scenario died
on REQUIRE(store_bbs_3mf(sp)).
This was the SINGLE failure in this fork's Unit Tests — 1 of 566, on Linux x86_64, Linux
aarch64 and macOS arm64 — from CI run 30191490709:
Failed to create backup path "/orcaslicer_model/Sun_Jul_26/08_49_41#5398#1":
boost::filesystem::create_directories: Permission denied [system:13]
It hid because that job had never run to completion on this branch before: every earlier
run was cancelled by the concurrency group first. It also passed on Windows x64, where the
drive-root path is writable, and it passes in the local build container, which runs as root.
Verified against the same defect in the Snapmaker fork by running the built binary as
uid 1000: permission denied before, 4 assertions passing after.
snaporca-vg8
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
snaporca-tkz, the last quarantined test. Root cause read out of the vendored
source rather than guessed: slvs/constrainteq.cpp, Type::ARC_LINE_TANGENT does
ExprVector ap = SK.GetEntity(arc->point[other ? 2 : 1])->PointGetExprs();
so it dereferences the ARC'S ENDPOINTS. A full circle entity carries only
point[0], its centre. point[1] and point[2] are zero handles, FindById throws
"Cannot find handle", and the process ABORTS rather than failing the solve —
taking every later test in the binary with it. That is also the wrong equation
for a circle regardless: it only makes the line perpendicular to the radius at
an endpoint that does not exist.
CT::Tangent no longer hands a full circle to that constraint. For a circle it
emits PT_LINE_DISTANCE(centre, line) = radius, which is precisely what tangency
to a circle means. Arcs keep the ARC_LINE_TANGENT path they are built for.
One limitation, stated rather than buried: the slvs C API takes a constant
distance and offers no way to reference the circle's radius parameter, so the
radius is captured when the constraint is emitted. That is exact whenever the
radius is fixed or is simply not driven by another constraint in the same
solve, and re-solving restores tangency if something else moves it. Tying them
would need an auxiliary point constrained onto both the circle and the line.
With this and eeca6794e7, both quarantined tests are gone and the exclusion in
kernel-test.sh goes with them. A green run now means the whole CAD suite
passed, not "everything except the two we gave up on":
149 cases / 2043 assertions, no filters.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Follow-through from eeca6794e7. The header claimed two pre-existing failures
are excluded and named both; the internal-thread case now runs like any other,
so only the solver SIGABRT is left. A comment that lists a test which is no
longer excluded sends the next reader looking for something that is not there.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
snaporca-kzy was filed as "internal thread cuts too little material". It does
not. Measured on the test's own fixture, a 40x40x20 box:
plain Ø12 bore removes 2261 mm3
internal thread removes 2157 = 1571 (minor bore) + 586 (groove)
apply_thread bores at the MINOR radius (radius - depth = 5) and then carves the
groove out to radius + depth = 7. A tapped hole therefore keeps the crests
between turns and holds MORE material than a plain clearance hole at the
nominal radius — which is what every real tapped hole does. The test asserted
the opposite, so it was asking for something physically wrong and had been
quarantined for it since it was written.
One hypothesis discarded on the way: that the shortfall was a tessellation
artefact, since chords on a helical surface undercut a concave bore. Exact
BRepGProp::VolumeProperties agreed with the tessellated volume to within
2.5 mm3, so that was not it and is not offered as a hedge.
The reference is now the tap-drill bore the thread actually starts from (Ø10),
against which the groove's 586 mm3 is the meaningful quantity — that is what
"the thread cuts" means. Test re-tagged [CadDocument][thread], so CI covers the
thread path again instead of skipping it.
Also documented the (void)internal in make_thread_profile. It reads like a bug
and is not: the V is the same shape either way and the caller decides, fusing
it onto a shaft or cutting it out of a wall. Someone "fixing" it to point
inward for the internal case would make the groove sweep already-empty bore
space and cut nothing — the exact failure the old comment described.
Suite 148 cases / 2035 assertions, with this test now among them.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Found by sweeping the index-space defect class deliberately rather than by
hitting it: that class produced 4 of the 8 defects found by hand yesterday, so
it was worth auditing every combo in the panel that maps a row selection onto
a document index.
Most of it came back clean — the loft sidecar vectors are consistent at all
four read sites, the sheet-body pickers go through the helper everywhere, mate
connectors carry client data. Six did not. A stored target_body indexes the
body list AS IT WAS just before that feature ran during replay, but Transform,
Mirror, Thicken, Rib, Project and DeleteFace all populated their combo from
the live m_doc.bodies. Boolean and Cut already replayed to the right slot.
The failure is concrete: model a body, Thicken it, then Cut something later in
the tree. A Cut replaces one body with two, so every index at or after it
shifts. Reopen the Thicken and the combo lists the post-cut bodies while
selecting the pre-cut index — showing, and on confirm re-targeting, a
different body than the feature actually used. A Boolean that consumes its
tool body shifts them the other way for the same result.
fill_body_choice() does the truncated replay populate_body_choices() already
did, for the single-combo tools. Six call sites, and 60 lines of duplicated
population loops go with them.
Visible change when testing: re-editing an early feature now lists FEWER
bodies, because it lists only those that existed then. That is correct — you
cannot target a body that did not exist yet — and it is what Boolean and Cut
have always done.
The new kernel test pins the invariant the GUI now leans on: a Cut turns one
body into two, and replaying to just before it yields the earlier, shorter
list. If body ordering after a split ever changes, that assumption fails
loudly here instead of silently in a dialog.
NOT click-tested — GUI wiring, compile-verified only. Filed as snaporca-oz7
and added to snaporca-cfi.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
on_edit_feature had two types falling into default: with "This feature type
can't be edited yet". Boolean was already handled, so the follow-up note was
stale on that point; the real gap was Cut and Import.
Cut now re-edits like any other feature: plane, offset and target body are
restored and the generic replace_feature path commits the change. The body
list is rebuilt with populate_body_choices(m_edit_index) for the same reason
Boolean does it — a Cut splits one body into two, so the live body list no
longer matches the one this feature's target index was recorded against.
Replaying to just before the feature makes the stored index land on the right
entry.
Import deliberately gets no dialog. An imported solid has no parameters to
re-edit: its geometry is rigid data read from a file, not something rebuilt
from numbers, and moving it is what the Transform feature already does.
Building an "edit" for it would duplicate Transform behind a second name. So
it now says that instead — the previous message implied a dialog was coming
that should not.
Imported 2D Text/SVG art is a different thing and stays re-editable; it
arrives as a Sketch feature carrying imported_regions and is handled above.
The default: arm is kept as a guard so a feature type added later announces
itself rather than silently swallowing the click.
NOT click-tested — GUI wiring, compile-verified only. Added to snaporca-cfi.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The Shellcheck workflow has failed on every push and every scheduled run since
2026-07-24, on exactly one finding: SC2029 in scripts/kernel-test.sh, the file
the CAD branch added. So the CAD work is what turned that job red, and a PR
arriving with a red job is a bad way to open a conversation with a maintainer.
Client-side expansion of $REMOTE is the intended behaviour — it is derived from
$VOL locally and the remote has no such variable, exactly as the rsync
destination two lines down relies on. So this is a disable with a reason, not a
silencing: the note says why the warning does not apply.
Verified by running the workflow's own command over all 24 matched scripts:
exit 0, no findings.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
docs/design_tab_upstream_portability.md explains the subsystem to a
maintainer; nothing explained it to a user. This is that: what the tab is,
how to get a first solid out of it, every tool grouped the way the toolbar
groups them, and the keyboard shortcuts read out of the source rather than
remembered.
The limitations section is deliberate. Rib needing a sketch with an explicit
open line, Surface Loft and Surface Fill having no hands-on verification, mates
composing transforms instead of solving simultaneously, move-face and
replace-face being absent, and the two quarantined kernel tests are all things
a user would otherwise discover by hitting them.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Maintainers will ask what the Design tab costs before they will look at the
diff, so measure it rather than assert it.
The headline correction: the OCCT delta is THREE toolkits, not two. The
comment in deps/OCCT/OCCT.cmake claimed "TKFillet + TKOffset (3.77 MiB,
Windows only)". Walking OCCT's own adm/MODULES and each toolkit's EXTERNLIB
shows ModelingAlgorithms holds twelve toolkits, that eight of them are built
either way because DataExchange (the STEP path upstream already ships) depends
on them, and that the true delta is TKFeat, TKFillet, TKOffset and TKXMesh —
of which TKXMesh is never produced. So three archives are built: 7.40, 5.38
and 4.42 MiB.
TKFeat is the interesting one. Nothing in the Design tab references it and it
is absent from the TKFillet/TKOffset dependency closure, so it is built for
nothing — OCCT's module flag is all-or-nothing per module. On static-link
platforms that is build time and zero shipped bytes.
Two numbers are deliberately absent, marked as absent, and not approximated:
the Windows DLL delta needs a Windows build (snaporca-gix), and a clean-build
time delta needs the deps prefix built twice on one machine. The old 3.77 MiB
figure is withdrawn rather than reused — it covered two of the three toolkits.
Also recorded: the vendored solver is 9,339 lines under GPLv3 with its LICENSE
preserved, which combines into this AGPLv3 fork without difficulty (AGPLv3
§13), and it is live code driving every sketch constraint — not a carried
corpse.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
DesignPanel.cpp was never listed in localization/i18n/list.txt, and xgettext
extracts only what that file names. All 934 of the Design tab's _L() calls
were therefore invisible to every translator in every language — not merely
untranslated, unextractable. Adding the one line is the actual fix; the rest
follows from it.
Regenerating the .pot brings the catalogue from 6007 to 6536 msgids. 549 of
the new ones are now translated into Italian: 5230 to 5779 translated
messages. Verified no existing work was destroyed — every msgid that survived
into the new .pot kept its translation, and the 36 that lost one are genuinely
gone from the sources. msgfmt --check-format passes, which matters here
because a large share of these carry %d / %s / %zu.
CAD terms follow Italian CAD convention rather than literal glosses: Fillet ->
Raccordo, Chamfer -> Smusso, Draft -> Sformo, Rib -> Nervatura, Mate ->
Accoppiamento, Shell -> Svuotamento, Pattern -> Serie, Sheet body -> Corpo
superficie. Strings identical in both languages are deliberately left
untranslated so gettext falls back to the msgid.
The long mixed-filament / Local-Z dithering tooltips are left untranslated on
purpose: slicer internals, outside a Design i18n task, and untranslated before
this commit too.
One string changed rather than translated. The Coord Sys hint read "Without an
edge the frame's rotation about its normal follows world X, not the body" —
that described the defect fixed in 1726e93760, so it was a lie as of that
commit. It now says X comes from the face's first edge.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
datum_frame took a FaceAndDirection frame's Z from the face normal, which
follows the body, but its X from coordsys_x_hint, a world constant, whenever
no explicit edge reference was set. Spinning a body about its own face normal
therefore left the frame bit-identical: the connector could not encode that
rotation at all, so Fastened and Slider mates claimed to fix an orientation
the frame could not see.
X now comes from the face's own first usable edge, which rotates with the
body. The hint survives only as a last resort, for faces that offer no
in-plane direction — a full circular edge has coincident endpoints, and a
seam projects to nothing in-plane.
The new test spins a box 90 degrees about its top-face normal and asserts the
frame's X turned with it. Reverting just the X_tent derivation and rerunning
makes it fail with "1.0 is within 0.000001 of 0.0" — cos(angle) between the
before and after X is exactly 1, i.e. the frame did not move — and that is the
only failure in 2019 assertions, so the test discriminates this defect and
nothing else.
Note for anyone replaying an older document: a face-only connector's frame
can now differ from what that recipe produced before, so a mate built on one
may place its body differently. Nothing in the suite or the golden v3 fixture
changed, but the semantics did.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This fork's Unit Tests job failed on every single commit, because CI runs the
whole ctest suite including the two cases tagged [known-broken] that
scripts/kernel-test.sh has always excluded locally. A job that is red
unconditionally is worse than no job: it trains everyone to ignore it, so the
next genuine regression arrives invisible.
No workflow change was needed. scripts/run_unit_tests.sh already passes
-LE NotWorking, and tests/CMakeLists.txt registers Catch2 tags as ctest labels
via catch_discover_tests(ADD_TAGS_AS_LABELS) — so the exclusion upstream
already ships works as soon as the cases carry the tag. Verified against the
built test tree: 337 tests unfiltered, 335 with -LE NotWorking, i.e. exactly
these two dropped and nothing else.
The second cause recorded in the issue, the test-reporter step failing with
"Resource not accessible by integration: 403" on a fork, is already fixed
upstream: the Publish Test Results step now carries continue-on-error: true.
The comment these cases carried claimed CI kept the bugs visible by reporting
them forever. That is now false and was never a good mechanism anyway, so
visibility moves to the tracker: snaporca-tkz for the solver SIGABRT, and
snaporca-kzy, filed now, for the thread groove volume. Neither is fixed;
neither is forgotten.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Rib guarded with `sk.type != CadFeatureType::Sketch`, while every other
sketch consumer — Extrude, SurfaceExtrude, SurfaceRevolve, the loft paths —
tests `!= Sketch && != Project`. A Project feature carries a plane and Line
entities, which is all a rib reads, so the guard blocked "project a body
edge, then rib along it" for no stated reason.
The picker in the Design tab offered Sketch features only, so it is widened
to match: a kernel that accepts Project refs and a GUI that never lists them
would have left the path unreachable anyway.
Worth recording for whoever hits this next: Rib also needs a sketch carrying
EXPLICIT entities. A parametric Rectangle sketch (add_sketch with
width/height) has an empty entities vector — build_sketch_wire synthesises
its profile on demand — so rib_entity 0 is out of range there and it fails
with "rib: bad entity". That is why Rib could not be driven headlessly at
all before this change; a Project feature is now the one programmatic way to
produce a ribbable line.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Two defects found while driving the tools that Phase B wired but nobody had
exercised yet.
apply_project had no all-edges branch: with no face picked and no explicit
edge list it threw "no edges or face selected". That is precisely the state
the Project card opens in, and its label reads "(all edges)" — so the card's
default could never be confirmed. It now projects every edge of the source
body. Edges perpendicular to the target plane collapse to a point when
projected, so segments whose endpoints coincide are dropped instead of being
emitted as zero-length lines that would poison the sketch downstream.
The second defect is why the first one was invisible. 29 of the 31 rollback
sites in McpControl ran `if (!ok) doc.undo();`, and undo() recomputes the
restored feature list — which succeeds and clears doc.error. Every failing
command therefore reported `error: ""`. Yesterday's fix covered 2 sites and I
treated the file as done; it was not. All 31 now capture the reason before
the rollback and restore it after. Failures that read as `""` now read as
"rib: bad entity" / "surface-revolve: revolve failed".
Verified on the running GUI through the control socket: the Project call that
previously returned ok:false now returns ok:true, and failures carry a reason.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Six of the surface entries, both Thicken variants, Rib, Axis, Coord Sys,
Mate and Delete Face all reused a sibling's glyph, so a drawer opened as a
column of identical faces and the card that opened rarely matched the entry
clicked. Fixed both halves: entry icons are now unique within their drawer,
and each card header uses the icon of the entry that opens it.
Two new glyphs, design_thicken and design_rib, are the only ones added —
everywhere else an existing icon already carried the right meaning
(design_revolve, design_offset, design_line, design_point,
design_c_coincident, design_delete).
The Surface drawer BUTTON deliberately keeps design_surface; only its
entries may reuse the solid glyphs, because a menu row carries its own
text label while two adjacent toolbar buttons do not.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
refresh_preview() exempted only Sketch and Plane from the ghost-preview path.
Axis, CoordSys, Helix and Project produce no solid either, so they fell through
to it, found nothing, and reported "invalid: preview produced no geometry" —
which also DISABLED Confirm, so all four tools were unusable rather than merely
noisy. Guaranteed on an empty document; with a body present the ghost path
finds something and masks it, which is why it survived until the tools were
tried on a fresh project.
All six non-solid tools are now exempt, each with its own ready message. The
list is not a guess: recompute() skips Sketch, Helix, Plane, Axis and CoordSys
outright and routes Project through apply_project(), which emits sketch entities
and no solid — so the panel and the kernel now agree on exactly what is not a
solid. Mate already had its own branch, since it needs Confirm gated on having
two distinct CoordSys features.
Introduced when Axis/CoordSys (batch 1) and Helix/Project (batch 3) were wired
without extending this exemption. Confirmed fixed on hardware: Axis ->
Plane Intersection with XY and XZ now resolves on an empty document, which also
exercises 60b04feea1.
Compiles clean; kernel untouched, suite unaffected at 143 cases / 1980
assertions.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The merge of upstream main into the PR branch resolved the conflict in
test_gcodewriter.cpp by keeping both sides but dropping the closing brace
of the set_pressure_advance BBL SCENARIO. That nested the incoming
TEST_CASE inside it and broke the build on every target that compiles the
test suite:
test_gcodewriter.cpp:1087: error: namespaces can only be defined in
global or namespace scope
Close the SCENARIO before the TEST_CASE. Both tests are preserved and no
test logic changes.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
axis_plane_a/b are filled by the GUI from populate_plane_choices(), whose rows
are XY / XZ / YZ followed by the datum planes, and the row is stored verbatim.
The kernel's base_plane() indexed datum_planes[ref] directly, so the two spaces
were off by three: picking XY resolved to datum plane 0, and picking the first
datum ran past the end and failed with "plane ref not found". The
PlaneIntersection axis type could not work from the GUI at all.
base_plane() now uses the encoding CadFeature::plane_base already uses — 0/1/2
are the base planes through the modeling origin, >=3 indexes datum_planes[ref-3]
— so there is one convention for plane references instead of two. That also
makes two base planes usable, which the previous code rejected as out of scope
even though XY x XZ is an ordinary way to define the X axis.
Removed the dead find_plane lambda directly above it. It was never called and
half-anticipated this exact offset ("if (ref >= 3) // base plane offset"),
which is presumably where the confusion started.
Tests: the existing parallel-planes case encoded the OLD convention, passing
axis_plane_a = 0 to mean "datum 0" — values the GUI cannot produce — so it is
re-based onto rows 3 and 4. Two new cases cover what the GUI actually emits:
base x base (XY x XZ -> X) and base x datum, the latter pinning the +3 offset.
Both verified to FAIL against the previous indexing, at test_caddocument.cpp
:2325 and :2346.
Found by auditing the remaining tools for the index-space defect class that had
already produced three bugs in the GUI; this is the first instance of it
crossing the GUI/kernel boundary.
Suite 143 cases / 1980 assertions (was 141/1972). GUI compiles clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The Surface drawer used design_extrude, the same face as the Add-material
drawer, so the two buttons were indistinguishable in the feature bar.
design_surface.svg is a draped patch — deliberately unlike design_plane (a flat
parallelogram) and design_extrude (a box with an up-arrow) — with a faint
interior rule so it reads as a skin rather than a solid face. Same visual
language as the other 71: 24x24, no fill, #b6b6b6, stroke-width 0.85, round
caps and joins.
The five surface ENTRIES that also used design_extrude now use it too. That is
not cosmetic tidying: a flyout button's face follows the last-picked entry
(SetBitmap_(icon_names[i])), so changing only the drawer's default icon would
have been undone the moment the user picked anything. Surface Loft keeps
design_loft, which already suits it.
The six rows still share one glyph between them, so they are told apart by
label alone inside the flyout. Per-entry icons belong with snaporca-vrg
(Draft/Shell reusing design_dressup), not here.
Compiles clean; kernel untouched. Icon confirmed legible on hardware.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
checkpoint() snapshotted `features` and undo() restored `features`, but
`variables` is a separate member of CadDocument. Every caller of the documented
checkpoint -> mutate -> recompute -> undo-on-failure pattern therefore failed to
roll a variable back: the bad value stayed in the document and every later
recompute failed, which is exactly the corruption the pattern exists to prevent.
Feature `expr` bindings were unaffected only because expr lives inside
CadFeature and rode along in the features snapshot — which is why the feature
side appeared to work.
This was a kernel gap, not a GUI one: McpControl::action_set_variable has the
same sequence and was equally broken.
The undo/redo stacks now hold a {features, variables} Snapshot. Nothing here is
serialized, so no recipe version change and no golden-fixture regeneration.
Two tests, both verified to FAIL against a faithful reproduction of the bug
(undo() leaving `variables` untouched) at test_caddocument.cpp:4420 and :4441:
one covers restoring a variable's previous value, the other covers removing a
variable that did not exist before the checkpoint. Worth recording that the
first mutation attempt was NOT faithful — it dropped the restore but kept
std::move(variables) into the redo stack, which empties the map as a side effect
and made the second test pass for the wrong reason. A mutation has to reproduce
the original defect, not merely break the code.
Second defect, same area: undo() calls recompute(), which succeeds and clears
doc.error, so the reason an edit was rejected was destroyed before anything
could display it. Six sites — four in DesignPanel, two in McpControl — now carry
the message across the rollback. on_remove_variable additionally asserted
"referenced by a feature expression" as fact; it now offers that as the likely
cause and appends the real error, since that diagnosis is wrong for any other
failure.
Suite 141 cases / 1972 assertions (was 139/1960). GUI compiles clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The sheet-only pickers filtered correctly and then threw the filtering away. Their
rows are the SHEET bodies, but GetSelection() was passed straight through as an
index into m_doc.bodies. With a solid at 0 and a sheet at 1 — the normal order,
since you extrude a solid before making a surface — the single row is row 0 but
body 1, so Surface Offset and Thicken Surface targeted the SOLID. The kernel then
refused with "target is not a sheet", which reads as a kernel bug rather than a
picker bug, and the row's own label ("Body 2") disagreed with what was targeted.
Four sites per tool were wrong, including the re-edit path, which compared a body
index against the sheet-only row count and so restored the wrong row.
populate_sheet_body_choices() now carries the real body index in client data, and
two helpers make the row/body distinction hard to get wrong again:
sheet_choice_body() reads it back, select_sheet_choice() finds the row holding a
given body. No caller touches GetSelection()/SetSelection() on these pickers.
This is the third instance of the same index-space confusion in this file, after
the 0-based body labels in the interference report and the Rib sketch picker. The
kernel suite cannot catch any of them: the kernel receives whatever index the GUI
computed, and its own tests pass correct ones.
Delete Face was structurally right — its picker uses the all-bodies populate, so
its indices genuinely match, and accumulation appends with a running list. Two
gaps closed: clicking "Add picked face" with nothing picked was a silent no-op,
indistinguishable from a broken button, and the same face could be added twice,
putting a duplicate id into delete_faces that the defeaturing has no reason to
cope with. Re-adding is now a no-op with a message, not an error.
Both confirmed working on hardware. Kernel untouched: 139 cases / 1960 assertions.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Port of snaporca ca25352d74. Hand-applied rather than cherry-picked: unlike the
Phase B commits, which only touched DesignPanel.{cpp,hpp} and transfer verbatim,
this one reaches into GLCanvas3D and DesignCanvas, where the forks genuinely
differ — mainline passes m_show_world_axes to _render_bed where Snapmaker passes
a local show_axes, and the surrounding code sits ~90 lines further down. git am
refused, correctly; the six edits were applied against mainline's own context and
the DesignPanel half came across as a patch.
Bed toggle: a "Bed" checkbox in the document/view row, on by default, in that row
rather than in a card because a view option must stay reachable with no tool open.
It drives GLCanvas3D::m_show_bed (default true, so Prepare and Preview are
untouched) and gates _render_platelist as well as _render_bed — hiding the bed
while leaving its grid and outline floating would read as a rendering fault.
Bound to wxEVT_TOGGLEBUTTON, not wxEVT_CHECKBOX: Orca's CheckBox derives from
wxBitmapToggleButton, so a wxEVT_CHECKBOX handler never fires.
Also gives the Placement drawer its own "placement" toolbar slot. put("place")
already holds the Place-on-Face button, and put() formats slot item 0 as the
control and later items as its chevron, so sharing the slot bottom-aligned the
drawer's button like a chevron.
196/196 targets, 0 compile errors, orca-slicer links (165 MB). The build script
still exits non-zero at the AppImage bundling step on libpython3.12.so.1.0 —
that is snaporca-96t, packaging only, and does not affect the binary.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The sidebar opened with eight tool cards stacked in it — Transform, Mirror,
Thicken, Rib, Project, Delete Face, Helix and Mate. A card added to the cards
sizer is visible until something hides it, and close_tool()'s hide-all only
runs on a tool SWITCH, so anything missing from the construction-time hide
block is on screen from the moment the tab opens. Those eight were wired into
close_tool() but never added here. All 37 cards are now hidden at startup.
Worth stating because it invalidates a check I ran while diagnosing this: every
card IS hidden somewhere in the file, so grepping for "hidden anywhere" says
nothing. The block that matters is the one in the constructor.
Second, the drawers mixed unrelated operations, and two group tooltips no
longer described their contents — Dress-up listed eight tools spanning three
different kinds of operation, and Add material still claimed to hold only
extrude/revolve/sweep/loft after Thicken and Rib were added to it.
One concept per drawer now:
Add material extrude, revolve, sweep, loft, thicken, rib
-> grows new solid material, whether from a profile, a face or
a line
Surface unchanged; already coherent
Datum / Curve plane, axis, coord sys, helix, PROJECT
-> reference geometry and derived curves. Project consumes a
body but PRODUCES sketch entities, so it is curve creation,
not a finishing operation
Placement TRANSFORM, MIRROR, MATE (new)
-> moves a body without changing its shape; a mate places one
body relative to another
Dress-up fillet/chamfer, draft, shell, delete face
-> finishing on the faces and edges of an existing solid
Hole / thread unchanged
The new drawer costs no toolbar width: the layout order already contained an
empty put("place") slot between "material" and "plane" with nothing registered
to it. Shift+Y and Shift+Z follow Transform and Mirror; every tool still
appears exactly once.
Compiles clean; kernel untouched, so the suite is unaffected at 139 cases /
1960 assertions.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
M6 landed the kernel side as two plain public maps — CadDocument::variables and
CadFeature::expr — reachable only through MCP's set_variable / set_feature_expr.
Nothing in the GUI could create a variable, so the parametric layer was
unreachable from the Design tab.
Variables get a wxListCtrl (name, expression) with add/edit/remove below the
feature tree, since they are document-scope and must not live in a card that
only exists while a tool is open. Expression bindings get one generic row in the
feature-edit path — a field-name combo plus an expression box — rather than an
extra control on each of 32 cards.
Every mutation copies McpControl's sequence exactly, and the rollback is the
part that matters: checkpoint, mutate, recompute, undo() on failure. Without it
one typo leaves the recipe permanently unrecomputable, since load() replays the
whole list. Removing a variable a feature still references fails that recompute,
so it reports the reference rather than a bare evaluation error.
The field-name combo is deliberately editable: only 11 feature types get a
curated field list, and free text is what makes the other 21 reachable. That is
safe because assign_field() throws "unknown parameter: <name>" for anything it
does not know, inside recompute()'s try block — so a wrong name gives a clear
message and a rollback, never a silently dead binding.
Two fixes on top of the generated wiring:
- make_combo() passes wxCB_READONLY, under which Orca's ComboBox HIDES its text
ctrl (ComboBox.cpp:51). There is no SetEditable() to undo that, so the field
combo is constructed directly with style 0; that shows the ctrl with
wxTE_PROCESS_ENTER and makes GetValue() return typed text.
- the field-list helper had been made a file-static function taking
DesignPanel::Tool, which required moving Tool out of private and into the
public API. It is now a private static member instead: 32 values of internal
card state should not be published to satisfy a signature.
Compiles clean (0 errors); kernel suite unchanged at 139 cases / 1960
assertions. Phase B is complete on this fork — all 16 previously GUI-less tools
plus the variables panel. Not yet exercised on a display.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Mate completes the set: every CadDocument feature type now has a card. It is the
one that needed CoordSys wired first, since a mate connector IS a CoordSys
feature and cs_a/cs_b are feature indices into the recipe.
The card states what each kind constrains rather than just naming it, because
the five kinds are not distinguishable from their labels — the useful fact is
which DOF each PRESERVES: Fastened fixes all six, Planar leaves in-plane
sliding, Revolute leaves spin, Slider leaves axial travel, Cylindrical leaves
both. The offset/angle spins retitle per kind, since offset is a plane distance
for Planar and a position along the axis for the three joint kinds.
Confirm is blocked, with the reason in the status line, when fewer than two
CoordSys features exist or when A and B are the same one: a mate with cs_a ==
cs_b is meaningless and a dangling index recomputes to nothing useful.
check_interference() gets a button in the feature-tree header behind a rule, not
a tool card — it adds no feature, so it must not checkpoint(), recompute(), or
touch the undo stack, and a card would imply it does. Results go to the status
line as count + worst volume, with the per-pair list in a message box, named as
the parts tree names them.
Fixes on top of the generated wiring:
- the button's sizer adds sat after the closing brace of the block declaring
trow, so trow was out of scope ("'trow' was not declared in this scope");
- the CoordSys client data was typed const void*, which Append rejects;
- the report labelled bodies 0-based while all 28 other body labels in this
panel (and the parts tree) are 1-based, so it would have called the tree's
"Body 2" an interference on "Body 1" — and it was the only unlocalised label.
Compiles clean (0 errors); kernel suite unchanged at 139 cases / 1960
assertions. Still to come: the M6 variables panel. The GUI has not yet been
exercised on a display.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Every CAD feature added across M1-M8 got a kernel API and an MCP method, and
almost none got a card. The MCP path was the only way to reach them, so from
the Design tab these features did not exist: add_axis and add_coordsys have
been callable since M1 with no UI at all.
Wired here, in three batches:
datum Axis (5 construction types), CoordSys
surface Surface Extrude / Revolve / Loft / Fill / Offset / Thicken Surface
body Transform, Mirror, Thicken, Rib, Project, Delete Face, Helix
Grouped into existing dropdowns rather than widening the 16-slot toolbar:
surfaces get one new "Surface" dropdown, body ops join Dress-up, Thicken/Rib
join Add-material, Helix joins Datum (renamed "Datum / Curve"). Shift+Y and
Shift+Z went to Transform and Mirror; the remaining five are shortcut-less
rather than getting invented chords.
Three places where the UI has to encode a kernel distinction, not just expose
a field:
- Surface Offset and Thicken Surface consume a SHEET body and fail with "target
is not a sheet" on a solid, so their pickers filter on
CadDocument::is_sheet_shape() and say so when no sheet exists. Thicken (solid
face -> plate) is a different tool and is kept visibly separate.
- delete_faces is a vector, so Delete Face accumulates picks via "Add picked
face" and shows the running list. Supporting one face would have been a
silent downgrade of the kernel field.
- CoordSys labels its edge pick with the consequence of omitting it: without an
edge, datum_frame() takes x from coordsys_x_hint (world constant) and the
frame cannot express rotation about its own normal — which is snaporca-en4,
and is why a Fastened mate built on a face-only connector cannot fix spin.
The Rib sketch picker needed the 3-arg Append(text, wxNullBitmap, clientdata):
ComboBox's own Append(text, bitmap) hides wxItemContainer's (text, void*), so
the 2-arg call resolves to the bitmap overload and fails with "conversion from
void* to const wxBitmap is ambiguous". The Sweep picker already documents this;
Rib now matches it.
Compiles clean (0 errors) against snaporca-deps; kernel suite unchanged at
139 cases / 1960 assertions. Mate, the interference report and the M6 variables
panel are still to come; the GUI itself has not been exercised on a display yet.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
First successful link of orca-slicer in this fork's history (158 MB, 738/738
objects, plus OrcaSlicer_profile_validator). 1633005bba got the test binary
green; the GUI had still never linked.
The blocker was in the deps image, not the code. That commit's message says
OCCT V7_6_0, Boost 1.84.0 and OpenCV 4.6.0 "are pinned identically in both
forks and were reused as-is". That is true of the VERSIONS and false of the
FLAGS: this fork's deps/OpenCV/OpenCV.cmake passes -DWITH_JPEG=OFF,
-DWITH_TIFF=OFF and -DBUILD_TIFF=OFF, and snaporca's does not. So the reused
build shipped lib/opencv4/3rdparty/liblibjpeg-turbo.a, which collides with the
deps' own lib/libjpeg.a — wx pulls that in via wxUSE_LIBJPEG=sys — on
jpeg_stdio_dest. Under -flto that duplicate is fatal, not a warning. snaporca
never sees it because its GUI does not link libopencv_world.a at all.
Fixed by rebuilding OpenCV 4.6.0 in orcacad-deps with this fork's own flags
(read out of the recipe rather than retyped), from the source already cached in
the image, into a clean build dir so no stale cache entry survived, and deleting
the orphaned bundled jpeg archive. The rebuilt libopencv_world.a has no jpeg or
tiff symbol references and its CMake config no longer names either library.
--allow-multiple-definition would have hidden this while leaving OpenCV carrying
codecs mainline deliberately turns off.
Lesson for the next dep: comparing deps recipes by version is not enough, diff
the CMAKE_ARGS.
Two more mounts, same root cause as the CMakeLists.txt mount this script
already documents — the baked tree is snaporca's:
- build_linux.sh, which builds `--target Snapmaker_Orca`; here the target is
OrcaSlicer and its output name is orca-slicer, so configure passed and ninja
then died on "unknown target". The binary check was looking for the wrong
name too.
- scripts/, because the packaging step needs scripts/appimage_lib_policy.sh;
without it a fully successful link still exited non-zero with "missing
AppImage helper" and the binary check never ran.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Both scripts defaulted IMAGE to snaporca-deps. That image is Snapmaker-based
and lacks Eigen 5.0.1, CGAL 5.6.3, wx 3.3.2 and Python 3.12 Development.Embed,
so running either script here without an explicit IMAGE= dies at CMake
configure — which is a large part of why this fork reached M8 having never once
compiled (1633005bba). The volume defaults were fixed in that commit; the image
default was missed in both files.
docker-iter-build.sh also never mounted deps_src, so it could not have
configured even with the right image: root CMakeLists.txt:947 FATAL_ERRORs
when deps_src/pybind11/include/pybind11/pybind11.h is absent, and
src/CMakeLists.txt pulls semver/hints/imgui/imguizmo/hidapi from the same tree.
kernel-test.sh got that mount in 1633005bba; this is the same fix for the GUI
build path, needed before the Phase B GUI work can be ported here.
CadDocument.hpp: mate_kind comment, mirrored verbatim from snaporca dbaa104f62
so the header stays byte-identical across forks.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
206/206 targets built, 139 [CadDocument] cases / 1960 assertions passing —
identical to snaporca's suite. Until now this fork had never compiled at all:
CMake died at configure, so the M1-M8 "suite green" figures were snaporca's
alone and the ports rested on patch-apply plus byte-identical sources.
Three fixes here; the deps work is in the orcacad-deps image (see below).
1. kernel-test.sh mounts deps_src. pybind11 is vendored in-tree and CMakeLists
requires its headers; without the mount the container fell back to the
image's baked tree, which predates it.
2. tests/libslic3r/test_3mf.cpp: repair the upstream-merge conflict resolution.
Resolving it as a union dropped the three closing braces of our SCENARIO, so
upstream's SCENARIO opened inside ours ("a function-definition is not allowed
here", plus 12 cascading catch2 registry errors). Restored from the pre-merge
file; whole-file brace balance is now 0 and the case count reconciles as
5 (ours) + 8 (upstream) - 3 (shared) = 10, with both CAD recipe tests intact.
3. tests/libslic3r/test_caddocument.cpp: REQUIRE_CONTAINS / CHECK_CONTAINS.
Catch2 v2 (snaporca) spells substring-match Matchers::Contains; v3 (here)
spells it ContainsSubstring and gives Contains an incompatible meaning,
range-contains-ELEMENT, which fails to COMPILE against std::string. Four
sites had been hand-adapted long ago, but M2-M8 kept porting in un-adapted
Contains calls — 16 of them — and nothing objected because nothing compiled.
Both forks now use the same find()-based macros, so the assertion lines are
byte-identical again and future format-patch ports carry across unchanged.
Five orphaned `using Catch::Matchers::Contains;` lines removed with them.
The deps gap that blocked configure needed five additions on top of
snaporca-deps, built into image orcacad-deps: Eigen 5.0.1, Python 3.12.13
(exact, with Development.Embed), wxWidgets 3.3.2 (was 3.1.5), CGAL 5.6.3
(was 5.4 — mainline's own MeshBoolean.cpp calls CGAL::parameters::default_values,
added in 5.5), plus the pybind11 mount above. OCCT V7_6_0, Boost 1.84.0 and
OpenCV 4.6.0 are pinned identically in both forks and were reused as-is.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
BUILD_VOL defaulted to snaporca_buildcache — snaporca's volume — so running
this fork's kernel test wrote into the other fork's build cache. Defaults to
orcacad_kerneltest now.
docker-iter-build.sh had the identical defect and was fixed to
orcacad_buildcache; this script was missed at the time.
Observed rather than theorised: an orca_cad run under a different deps image
overwrote snaporca_buildcache's CMakeCache.txt, after which snaporca's own
kernel test failed to configure ("Cannot find NLopt library 'nlopt_cxx' in
.../lib/cmake/nlopt/lib") because it inherited the foreign cached paths. No
foreign object files were written — the run died at configure — but the cache
was poisoned, and the volume had to be wiped and rebuilt clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Catches the fork up from 449a4cf9fc (2026-06-28) to d6cb667b89 (2026-07-24).
Upstream touched 2326 files; 17 of them overlap the 164 this branch touches.
16 of the 17 auto-merged, including all three CMakeLists.txt, the build_all.yml
CI workflow, and every GUI file. The CAD core never conflicts: CadDocument,
SketchEngine, SketchSolver, McpControl and test_caddocument are files this fork
adds, so upstream does not touch them.
The one conflict, tests/libslic3r/test_3mf.cpp, was purely additive in all three
hunks and is resolved as a union: our test pinning that store_bbs_3mf embeds the
CAD recipe as Metadata/SnapOrca_cad.bin, upstream's multi-nozzle plate-metadata
round-trip tests, and both sets of includes. All three were verified present
after resolution rather than assumed.
NOT BUILD-VERIFIED, for a reason that predates this merge and is not caused by
it: this fork cannot be configured on nativedev at all. Its CMakeLists has
required Eigen3 5.0.1 since before the merge (line 592 pre-merge), while the
only deps image on the machine is snaporca-deps, built for snaporca's
find_package(Eigen3 3.3). CMake fails at configure, so nothing compiles.
That means this fork's Catch2 suite has never run. Every "suite green" figure
recorded for M1-M8 was snaporca's suite; the ports were verified by patch-apply
plus the CAD sources being byte-identical to snaporca's. Building an orca_cad
deps image with Eigen 5.0.1 is what would finally close that gap.
Pre-merge state is preserved at branch cad-mainline-pre-upstream-2026-07-25
(30d54f0074).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
check_interference(min_volume) reports every pair of solid bodies whose
intersection encloses more than min_volume, as {body_a, body_b, volume}. It
reports only — no geometry is mutated, so calling it cannot disturb mates or
placements. Read-only at the MCP surface too: no checkpoint, no recompute.
Sheet bodies are skipped up front: an intersection involving one encloses no
volume, so the boolean would be wasted work. Bodies that merely touch share a
face and enclose nothing, so face-to-face contact is not an interference.
A boolean that fails on one pair must not lose the report for every other pair,
so each pair is guarded — and OCCT raises Standard_Failure, which is not a
std::exception and would otherwise escape.
No new serialized fields, no recipe bump: this reads `bodies`, which is
recompute output and was never serialized.
No separate MCP listing for instances and mates: describe_scene already emits
the feature tree, and Mate has rendered there correctly since M8a fixed
feature_type_name.
Tests assert the exact overlap volume (20*20*4 = 1600 mm^3), both negative cases
(clearly apart, and exact face contact), that sheets are skipped, that the
min_volume gate silences a real overlap, and that a clash created by a Fastened
mate is detected — which ties the M8b placement work to this report.
Suite 139 cases / 1960 assertions green. McpControl.cpp is reviewed but not
compiled by kernel-test.sh, which builds only libslic3r_tests.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Each kind constrains the DOFs it owns and PRESERVES the rest from the body's
current pose, following the pattern Planar established in M8a. Resolved instead
as "Fastened with a parameter", all three would have been geometrically
identical to Fastened — relabelling rather than behaviour.
Revolute fixes position on the axis line; rotation about it survives
Slider fixes orientation and perpendicular position; axial position survives
Cylindrical fixes the axis line only; rotation and axial position both survive
No new serialized fields, no recipe bump, no fixture regeneration: mate_kind is
already an int and mate_offset / mate_angle already exist.
The minimum-rotation z-alignment (including the antiparallel 180 deg case fixed
in M8a) is now a shared make_z_align lambda rather than a second copy.
Fixes a rotation-about-pivot bug found by the no-op tests: R_full was built as a
rotation about the origin with a translation to oB appended, instead of a proper
rotation about oB (translation = oB - R*oB). It moved bodies that were already
correctly placed, and accounted for three of the seven initially failing cases.
Testing notes, both of which cost real debugging time here:
- Mates are defined on connector FRAMES, but the convenient thing to measure is
CentreOfMass(), and the two coincide only when the body is symmetric about its
connector. Five expectations in this milestone asserted the centroid while
meaning the connector. These tests assert on the mated face's centroid.
- A CoordSys built from a face ALONE takes its z from the face normal (which
follows the body) but its x from coordsys_x_hint, a world constant. Such a
frame cannot see rotation about its own normal, so no mate can correct or
preserve a spin it does not encode. The Slider and Cylindrical rotation tests
pin coordsys_edge to an edge of their own body; without that both passed
vacuously, one of them for a wrong implementation.
The Cylindrical rotation test was verified to fail when its mate kind is mutated
to Slider, and the Slider test failed at axis_aligned == 2 before the connectors
were edge-pinned. Neither is green by accident.
Known wart: mate_angle is silently ignored for Slider, whose rotation is fully
constrained. Defensible but undiagnosed at the API surface.
Suite 134 cases / 1927 assertions green. McpControl.cpp is reviewed but not
compiled by kernel-test.sh, which builds only libslic3r_tests.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
An assembly is a multi-body document. An instance is already expressible as
Transform with xf_copy=true, and a mate connector is already a CoordSys
feature, so this adds exactly one feature type: Mate.
No constraint solver. A mate rigidly transforms the body carrying connector B
so that B's frame lands on connector A's, applied in feature order like every
other feature. Chains resolve by composition; closed kinematic loops do not
converge (last mate wins) and are out of scope.
The vendored SolveSpace in src/libslic3r/slvs/ was evaluated for 3D extension
and rejected: it is built and linked but has zero callers, and SketchEngine's
solver is hand-rolled. Extending it would mean adopting a dependency to write
more code than the alternative.
- CadFeatureType::Mate appended; six fields (mate_kind, mate_cs_a, mate_cs_b,
mate_offset, mate_angle, mate_flip) appended at the END of both cereal lists
- SNAPORCA_CAD_RECIPE_VERSION 2 -> 3; v2 blobs are rejected, as by design there
is no migration path. Golden fixture renamed to cad_recipe_v3.bin and
regenerated once, extended with two CoordSys + one Mate so the new fields are
tripwired by the field-order assertions
- datum_frame() extracted from resolve_datum_coordsys() so a mate can resolve
its connectors against the in-progress bodies vector during replay
- apply_mate dispatched early-return, so Mate is deliberately absent from
starts_new (unreachable for that dispatch style)
- Planar: the degenerate branch splits on the sign of zB.z_target — antiparallel
needs a 180 deg rotation about a perpendicular axis, which an earlier revision
silently skipped, leaving the body's normal inverted
- MCP: mate command, named bare to match the other 38 methods
Drive-by: feature_type_name() was missing Mirror, ThickenSurface, SurfaceOffset,
SurfaceLoft and SurfaceFill, which reported as "Unknown" to MCP clients.
Suite 122 cases / 1741 assertions green. Note that kernel-test.sh builds only
libslic3r_tests, so McpControl.cpp is reviewed but not compiled here.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The two remaining ways to create a sheet body. SurfaceLoft skins 2+ profile
sketches without end caps via a new SketchEngine::make_loft_surface — a
sibling of make_loft with the ThruSections solid flag false, so no existing
call site changes. SurfaceFill patches a single closed boundary wire into a
smooth face with BRepOffsetAPI_MakeFilling, adding each boundary edge as a
C0 constraint.
Purely additive: two enum values appended to CadFeatureType, reusing the
existing loft_profile_refs/loft_ruled and sketch_ref fields. No new cereal
fields, recipe stays v2, golden fixture unchanged (30773). MCP
surface_loft/surface_fill added as pure additions. Suite 107 cases / 1553
assertions green, including a test contrasting the open skin against the
solid loft of the same profiles.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Two features bridging sheet bodies back to solids and to other sheets:
ThickenSurface feeds a whole sheet shell to MakeThickSolidBySimple (the same
OCCT recipe the face-level Thicken already uses) and appends the result as a
solid; SurfaceOffset offsets a sheet's shell along its normals via
MakeOffsetShape::PerformBySimple, keeping it open. Both refuse a non-sheet
target with a clear error.
Purely additive: two enum values appended to CadFeatureType, reusing the
existing target_body / thicken_thickness / thicken_flip / plane_offset
fields. No new cereal fields, recipe stays v2, golden fixture unchanged
(30773). MCP thicken_surface/surface_offset added as pure additions.
Suite 103 cases / 1520 assertions green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Two body-producing features that emit an open shell instead of a capped
solid: SurfaceExtrude (prism of a sketch wire, no end caps) and
SurfaceRevolve (revolve of a wire about an in-plane axis, no caps). Each
appends a new sheet body whose TopoDS_Shape has no TopAbs_SOLID.
Purely additive: two enum values appended at the end of CadFeatureType,
reusing existing serialized fields (sketch_ref/distance,
revolve_angle/revolve_axis). No new cereal fields, recipe stays v2, golden
fixture unchanged. is_sheet_shape() derives sheet-ness from the OCCT shape
type (bodies are not serialized). MCP surface_extrude/surface_revolve added
as pure additions. Suite 99 cases / 1474 assertions green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Add named document variables (CadDocument::variables) and per-feature
expression bindings (CadFeature::expr, field-name -> expression). On
recompute(), variables are evaluated topologically (cycle detection), then
each feature's expr entries are evaluated and written into its numeric fields
before geometry runs. Self-contained shunting-yard evaluator (+ - * /, parens,
unary minus, sqrt/abs/sin/cos/tan(deg)/min/max, pi). assign_field allow-lists
the 33 dimension fields + pattern_count; unknown names error loudly.
Additive: recipe stays v2 (fields appended to both cereal lists, golden
fixture regenerated). MCP set_variable / set_feature_expr are pure additions.
Suite 95 cases / 1440 assertions green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Extend CadFeatureType::Pattern (no new enum) with a curve mode: when
pattern_curve_sketch >= 0 it takes precedence over linear/circular. The guide
entity is sampled at equal-parameter points via a file-local sample_entity_2d()
(Line lerp, Arc angle-lerp, cubic-BSpline Bernstein, p0->p1 fallback), and each
seed copy is translated by (P_i - P_0) and fused. Two serialized fields
(pattern_curve_sketch/pattern_curve_entity) appended to both symmetric cereal
lists (version stays 2, golden fixture regenerated 30269->30517). MCP:
pattern_on_curve. 3 new [CadDocument][pattern] tests; suite 89/1395.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
New CadFeatureType::Rib (appended). A straight open Line entity in a sketch
is offset ±thickness/2 along its in-plane perpendicular into a thin rectangle,
extruded rib_depth along the sketch-plane normal, and fused to the target body.
Line-only for now (ponytail; polyline/arc ribs are a later extension) — a
non-line entity fails cleanly at recompute. Four serialized fields
(rib_sketch_ref/rib_entity/rib_thickness/rib_depth) appended to both symmetric
cereal lists (version stays 2, golden fixture regenerated 29525->30269). MCP:
rib. 3 new [CadDocument][rib] tests; suite 86/1376.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Extend CadFeatureType::Hole (no new enum value) with a style flag
(simple/counterbore/countersink) and the matching geometry: a coaxial
shallow cylinder cut for counterbores, a cone-frustum cut for countersinks.
A file-local hole_std_lookup() resolves screw designations (ISO 273/4762/
10642 metric M3–M10 + common ANSI unified) into clearance/cbore/csink dims;
add_hole_standard() fills the feature from it, add_hole_styled() takes them
explicitly. Six serialized fields appended to both symmetric cereal lists
(recipe version stays 2, golden fixture regenerated 28161->29525). MCP:
hole_styled, hole_standard. 4 new [CadDocument][hole] tests; suite 83/1351.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
New CadFeatureType::DeleteFace: removes a set of global face ids from target_body
and heals the gap via BRepAlgoAPI_Defeaturing (TKBO, already linked), mirroring the
Shell/Draft body-modifying pattern. delete_faces appended to both symmetric cereal
lists (recipe version stays 2, golden fixture regenerated 27913->28161). MCP
delete_face method (pure additions). 3 new [CadDocument][deleteface] tests: remove a
fillet face restores the sharp-box volume, bad index fails safely, round-trip.
Full kernel suite green (79 cases, 1309 asserts).
Move-face / replace-face deferred to snaporca-3c4 / snaporca-tc6 (no clean shipping
OCCT direct-modeling primitive; need research, and replace-face depends on M7 surfaces).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Adds SketchEngine::make_bridge (4-pole cubic Bezier, G1-tangent to Line/Arc
endpoints, straight-line fallback for other types) emitted as the existing
BSpline SketchEntity — no new geometry type, no serialized-field change, golden
recipe fixture untouched. CadDocument::add_bridge appends it (non-parametric,
index-validated, throws on bad refs). MCP `bridge` method mirrors action_project.
4 new [CadDocument][bridge] tests; full kernel suite green (76 cases, 1280 asserts).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The construction flag was already honored (excluded from the extrude wire in
SketchEngine.cpp, participates in the solver, and serialized) but the existing
"construction line excluded" test was a false tripwire: its construction line
ran corner-to-corner inside the square, so the bbox was unchanged whether or not
the line was excluded.
- Strengthen that test: the construction line now runs (-30,0)->(30,0) outside
the profile, so an exclusion regression breaks the closed wire / bbox.
- Add a serialize/deserialize round-trip test asserting construction survives.
- Lock the flag on-disk: add Sketch_Ctor to the golden fixture with a real edge
+ a construction edge, and assert both flags survive the binary recipe.
Test-only; no kernel change. Recipe version stays 2 (construction was already a
serialized field). Suite: 72 cases / 1246 assertions green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Onshape-style "Use / Convert entities": pick edges (or a whole face) of an
existing solid and get sketch geometry projected onto a target plane, then
extrude/revolve/edit it like any sketch. Parametric: apply_project re-derives
the feature's entities from the source body on every recompute, so editing the
source updates the projection.
Line edges -> Line entities (exact); circles/arcs whose plane is parallel to
the sketch plane -> Circle/Arc (exact); everything else (incl. non-parallel
circles that project to ellipses) -> sampled Line chain.
Append-only: new enum value Project + project_source_body/project_edges/
project_face fields at the end of save/load; recipe version stays 2. Recompute
loop made non-const solely so apply_project can write back f.entities.
Suite 67->71 cases, 1178->1229 assertions, RC=0. Fixture 25493->26835 B.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Split-by-face reuses the existing Cut feature rather than adding a new type:
two appended fields (cut_face_body, cut_face) let apply_cut derive the cut
plane from a picked face via SketchPlane::from_face when cut_face >= 0,
otherwise it keeps using the base `plane`. cut_offset / cut_flip still apply
along the derived normal, so the same square-wire split machinery handles
both cases. add_split_by_face() is the convenience entry point; MCP gains a
`split` method (body / face_body / face / keep_upper / keep_lower).
Serialization stays append-only — cut_face_body, cut_face appended to
save/load, recipe version unchanged at 2.
Tests: the previously-missing both-halves plane cut (keep_upper && keep_lower
=> two bodies whose volumes sum to the original), split-by-face via a
top-face plane offset into the interior, keep-upper-only, and a round-trip.
Golden fixture regenerated with a GoldenSplit cut-by-face feature and exact
field-value assertions. Suite 63 -> 67 cases, 1119 -> 1178 assertions.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Pick a face of an existing body, offset it by a wall thickness along its
normal, and append the resulting thin solid as a new body. Onshape-parity
Tier-2 item; the kernel had Shell (hollow a whole solid) but no way to turn
a single face into a plate.
Kernel: CadFeatureType::Thicken, add_thicken()/apply_thicken() as a
body-level op next to Transform/Mirror. The picked face is wrapped in a
TopoDS_Shell and offset via BRepOffsetAPI_MakeThickSolid::MakeThickSolidBySimple;
the result is orientation-normalised to positive volume (same convention as
apply_mirror). Serialization stays append-only — thicken_face,
thicken_thickness, thicken_flip appended to save/load, recipe version
unchanged at 2.
MCP: `thicken` method (body/face/thickness/flip) plus the missing
feature_type_name() case.
Tests: 6 new [CadDocument] cases (plate volume within 1%, flip direction,
bad face id, zero thickness, fuse-with-source, round-trip). Golden fixture
regenerated with a GoldenThicken feature and exact field-value assertions.
Suite 57 -> 63 cases, 1054 -> 1119 assertions.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Until now move and rotate lived only in the GUI as m_body_xform, a display
transform. That made them a correctness hole, not a missing tool: a moved body
recomputed and booleaned at its ORIGINAL position, and the move was not in the
recipe at all, so it vanished on save/reload. Only export_step consulted the
transform, which is why the discrepancy stayed hidden.
CadFeatureType::Transform makes it a real feature: rotate angle_deg about
xf_axis through xf_pivot, then translate by xf_translate, applied to the target
body with BRepBuilderAPI_Transform. xf_copy=true keeps the source and appends
the transformed body instead of mutating in place, which covers Onshape's
Transform/copy in the same feature.
Rotation is composed before translation (trsf = tr * rot) so the pivot means
what a user expects — the point the body turns about, not a point that then
drifts with the translation. A rotation with a degenerate axis is refused
rather than silently skipped; a zero angle skips the rotation entirely so a
pure move needs no axis at all.
The decisive test is not the bbox arithmetic but "moved body participates in a
later boolean at its new position": two coincident boxes, one moved to partial
overlap, fused. The fused volume must be strictly greater than one box (the
move took effect in the kernel) and strictly less than both (they still
intersect). With a display-only transform the first assertion fails.
Serialization stays append-only; recipe version unchanged at 2. Golden fixture
regenerated with a GoldenTransform feature carrying distinctive literals so a
field reorder shows up as obviously wrong values. Also fills in the Helix arm
of feature_type_name(), missing since the helix commit.
Kernel suite 51 -> 57 cases, 985 -> 1054 assertions, green.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Adds CadFeatureType::Helix — the missing input for Sweep. Sweep already
existed but could only follow a sketch, so springs, coils, augers and
non-standard-pitch threads were unreachable. Helix + the existing Sweep now
gives all of them with no further work.
Built the OCCT way: a 2D line on a Geom_CylindricalSurface (Geom_ConicalSurface
when helix_taper_deg != 0) turned into an edge and lifted to 3D with
BRepLib::BuildCurves3d — a true analytic helix, not a sampled polyline, so a
swept spring is smooth rather than faceted. The axis is the plane normal
through the plane origin, matching how Revolve and Plane already work.
Sweep's path resolution is widened to accept either a Sketch (unchanged
behaviour) or a Helix, and rejects anything else with a clear error. Helix
itself is skipped in route_feature and recompute — like the datum features, it
produces no body and exists to be consumed.
Invalid input is refused rather than approximated: non-positive radius or
pitch, negative height, a turn count above 10000 (which would hang OCCT), and
a taper that would drive the radius negative before reaching the top all fail
with a specific error.
Serialization: helix_radius/pitch/height/left_handed/taper_deg appended at the
very end of both save and load, identical order, after the coordsys block.
SNAPORCA_CAD_RECIPE_VERSION stays 2; Helix is appended to the end of
CadFeatureType. Golden fixture regenerated with distinctive literals and
field-value assertions; all pre-existing assertions pass unchanged.
Tests assert analytic values. The one that actually proves it is a helix and
not a circle or a spiral: arc length of r=5 pitch=2 height=10 measured with
BRepGProp::LinearProperties against 5*sqrt((2*pi*5)^2 + 2^2), WithinRel 1e-3.
Plus bounding box (2r in X and Y, height in Z), the conical top radius, the
left-handed winding compared at equal parameter, and the integration test:
a circle r=1.5 swept along a 5-turn helix gives one valid solid of ~1115 mm^3
(WithinRel 0.1 — pipe sweeping is not exact).
MCP: `helix` method registered in describe_tools().
Ported from snaporca 6cdc6b5459. Two fork-specific adjustments: the two new
error-message assertions use Catch2 v3's ContainsSubstring (v2's Contains does
not exist here), and the golden fixture is copied rather than regenerated
because this fork cannot be compiled locally — make_golden_doc_v1() is
byte-identical across both forks, so the blob is provably the same.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Adds CadFeatureType::Axis and ::CoordSys — reference geometry that produces
no solid, modelled on the existing Plane datum feature.
Axis construction methods (AxisType): TwoPoints, FaceNormal,
CylinderCenterline, PlaneIntersection, AlongEdge. The centreline case is the
useful one: it gives a real axis through an existing hole or boss.
Coordinate systems (CoordSysType): PointWorld and FaceAndDirection. The
latter Gram-Schmidts the picked references, so the stored frame is
orthonormal even when the user's X hint is not perpendicular to the face
normal; the third axis is derived by cross product rather than stored, so it
cannot drift out of sync.
Revolve and pattern are deliberately NOT rewired to consume these — this
commit adds the reference geometry only and changes no existing behaviour.
Serialization: all axis_*/coordsys_* fields appended at the very end of both
CadFeature::save and load, identical order, after mirror_keep_original.
SNAPORCA_CAD_RECIPE_VERSION stays 2; Axis and CoordSys are appended to the
end of CadFeatureType so existing type ordinals are unchanged. Golden fixture
regenerated with distinctive non-default literals and field-value assertions
for every new field; all pre-existing assertions pass unchanged.
Tests assert analytic values: two-point axis direction exactly +Z with unit
length, cylinder centreline collinear with Z and on the true axis, parallel
planes fail cleanly, and the Gram-Schmidt frame is orthonormal to 1e-9 with
X x Y == Z. Degenerate input (identical points) fails with a non-empty error
rather than producing NaNs.
MCP: `axis` and `coordsys` methods registered in describe_tools().
Ported from snaporca 242d4efecb. The golden fixture is copied rather than
regenerated because this fork cannot be compiled locally (Eigen 5.0.1 vs the
build image's 3.3); make_golden_doc_v1() is byte-identical across both forks,
so the two fixtures are provably the same blob.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Adds CadFeatureType::Mirror: reflect a target body about a plane using
gp_Trsf::SetMirror + BRepBuilderAPI_Transform. BooleanMode::New keeps the
mirrored copy as its own body (mirror_keep_original decides whether the
source survives); BooleanMode::Add fuses it back into the source, so an
overlapping mirror does not double-count volume.
Serialization: mirror_keep_original is appended at the very end of both
CadFeature::save and load (append-only contract). The mirror plane reuses
the existing `plane` member and the body selector reuses `target_body`,
as Cut already does. Golden fixture regenerated at the current
SNAPORCA_CAD_RECIPE_VERSION = 2; the existing field-value assertions all
still pass unchanged, and the reorder tripwire was re-verified after
regeneration (swapping draft_face/draft_angle in `load` alone still
fails the golden test).
Tests assert analytic values: mirrored volumes equal (8000 each) with the
reflected centroid, Add on a non-overlapping asymmetric body gives exactly
2x volume, Add across an intersecting plane gives strictly less than 2x,
and an invalid body index fails cleanly with a non-empty error.
MCP: `mirror` method registered in describe_tools().
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Port of snaporca c2821af783. New GeometryEngine::mass_properties over BRepGProp
(separate VolumeProperties/SurfaceProperties), bounds-checked
CadDocument::body_mass_properties, and a mass_properties MCP method. Query-only:
no CadFeature, no serialization, no version change. Analytic tests
(WithinRel/WithinAbs): cube 8000/2400/COM(0,0,10)/inertia 533333, cylinder
500pi/300pi, hollow = solid-500pi, invalid index -> valid=false.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Port of snaporca 04c3d579a7. Bump SNAPORCA_CAD_RECIPE_VERSION 1 -> 2;
deserialize_recipe sets a user-facing error distinguishing too-new / too-old
/ corrupt instead of a silent bare false. No per-version migration by design.
Golden fixture regenerated at v2 (v1 retired), field-value reorder tripwire
unchanged. Fork adjustment: Catch2 v3 string matcher ContainsSubstring
(not v2's Contains) in the two new version-mismatch tests.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
CadFeature::save/load is append-only by contract, and ~20 planned features
each append fields. The existing roundtrip test cannot police that: it writes
and reads with the same code, so any self-consistent ordering passes. Only a
blob written by older code and stored on disk can detect that the format moved.
The first attempt at this test passed while the defect was present. I proved
it by swapping draft_face (int) with draft_angle (double) in both save() and
load() -- a genuine byte-layout change -- and it still reported 613 assertions,
exit 0. It asserted only derived geometry: body count, per-body volume, feature
types. The golden document had no Draft feature, so those fields sat at their
defaults, the reorder scrambled values nothing read, and the recomputed solids
came out byte-identical.
So the fixture now asserts the DATA, not what the data produces:
- make_golden_doc_v1() builds 22 features across 14 types (Draft, Shell,
Revolve, Pattern, Cut, Hole, Chamfer, Fillet, Extrude taper/symmetric,
Thread, Sweep, Loft, Boolean, Plane) with distinctive non-default literals
(draft_angle 7.25, shell_thickness 1.375, revolve_angle 217, pattern_count 5)
so a reorder produces visibly wrong values rather than swapped defaults.
- Layer 1 reads the committed blob with raw cereal and asserts field by field,
independent of recompute, so a geometry regression cannot mask a format break.
- Layer 2 keeps the geometry checks as a separate concern.
Verified to trip, twice, by deliberate breakage rather than by assertion:
draft_face <-> draft_angle -> draft_face reads 1075642368 (0x401d0000),
the high half of double 7.25
revolve_angle <-> revolve_axis -> revolve_angle reads 0.0, not 217.0
Both revert clean to 758 assertions / 30 cases.
Also scoped the "regenerate the fixture" hint to the feature-count check only.
It was in scope for every assertion in the block, so a detected reorder told
you to run [.regen] -- which would bake the corrupted layout in as the new
golden and permanently disarm the test. A guard must not advise disabling
itself.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Ported from snaporca f9e0f99bcb. The patch needed fuzz: this fork's copy of
test_caddocument.cpp carries a Catch2 v3 include, a `using Catch::Approx`, and
an extra [Deviation] case appended after line 1611 -- exactly where these hunks
land. Verified after applying: new symbols present, the [Deviation] case
intact, braces balanced, and only WithinAbs/WithinRel used (both exist in v3).
Compile and test verification here is CI, not local: this fork needs Eigen
5.0.1 while the local deps image ships 3.3.
Adds scripts/kernel-test.sh: build the libslic3r_tests target and run the CAD
kernel tags, exit code as the whole contract. Clean build 4m36s, incremental
18s, no display needed -- every [CadDocument] case builds a CadDocument,
recompute()s it and asserts on geometry.
Four things this had to get right, each found by it going wrong first:
- SLIC3R_GTK=3 and BUILD_TESTS=ON are mandatory. src/CMakeLists.txt turns
SLIC3R_GTK into 'wx-config --toolkit=gtk<N>', so omitting it asks for
toolkit "gtk", nothing matches, and configure dies with the thoroughly
misleading "Could NOT find wxWidgets" -- while wx-config sits right there
in the deps prefix, working. BUILD_TESTS=ON is what creates the target.
- Configure runs unconditionally. Guarding on "CMakeCache.txt exists" is
wrong because a FAILED configure writes that file too, after which the
guard skips reconfiguring forever and every later run silently reuses the
poisoned cache, ignoring corrected flags.
- A new build volume is always built clean. Cloning a warm cache from
another tree is a correctness trap: rsync preserves source mtimes, ninja
compares them against foreign object timestamps, concludes everything is
current and relinks stale objects. That produced a binary containing NO
[CadDocument] tests at all -- while exiting 0. A green run that tests
nothing is worse than a red one.
- --host builds where the deps image already lives, staged per volume so
parallel workers never share a tree.
The two [known-broken] tags: "entity constraints: tangent/midpoint/symmetric/
angle" aborts inside the vendored solver (slvs/dsc.h FindById, "Cannot find
handle"), and SIGABRT is fatal to the Catch2 process -- that single case took
the suite down at 12 of 31, so a green baseline was unreachable. The thread
groove case is a plain pre-existing assertion failure. Both are excluded from
the dev loop's default filter ONLY; ctest in CI still runs and reports them,
so neither bug is hidden. Baseline is now 603 assertions / 29 cases green,
which is what makes "I broke nothing" a meaningful statement.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Enumerated from the source rather than from recollection: CadFeatureType and
add_* in CadDocument.hpp, Tool in DesignPanel.hpp, Mode in
DesignSketchTool.hpp, SketchConstraintType + SketchEntity::Type in
SketchEngine.hpp, and the JSON-RPC dispatch in McpControl.cpp.
Findings worth stating up front:
- The 2D sketcher is at or near Onshape parity -- 19 constraints, every entity
type including B-splines and elliptical arcs, trim/extend/offset/mirror and
both array kinds. Very little is missing there.
- The gaps are all breadth beyond sketching: assemblies/mates, surface
modelling, sheet metal, drawings, and variables/configurations.
- The most defensible criticism is the absence of variables and expressions.
Every dimension is a literal double, so the feature tree is parametric in
structure but not in value -- "change one number and the model updates" is
only half delivered. It is also the cheapest Tier 1 item to close.
The doc separates platform capabilities (version control, FeatureScript, FEA,
rendering, cloud PDM) into their own tier rather than counting them as missing
tools: that is Onshape-the-platform, not Onshape-the-modeller, and holding a
slicer tab to it would not be a fair comparison.
One entry is a correctness gap rather than a missing feature: move/rotate body
(m_body_xform) is display-only and never enters the B-rep, so a moved body
exports and booleans at its original position while the viewport shows it
moved.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The 2026-06-21 assessment was written before the persistence work landed and
got two load-bearing facts wrong. Both are corrected here against the branch
itself rather than from recollection:
1. It called OCCT "a dependency mainline OrcaSlicer has never carried" and
built its whole conclusion on that. False: deps/OCCT/ exists at the
merge-base, and upstream already links it from Format/STEP.cpp,
Format/svg.cpp and Shape/TextShape.cpp. The real dependency diff is one
line -- BUILD_MODULE_ModelingAlgorithms OFF -> ON -- costing a measured
3.77 MiB of Windows DLLs (TKFillet + TKOffset; TKBool already arrives
transitively via DataExchange).
2. It described the vendored SolveSpace solver as LGPL. False:
src/libslic3r/slvs/LICENSE is GPL-3.0. Harmless for us, but a licence
must not be misstated in a document aimed at upstream.
It also claimed no changes to Model, which stopped being true when 3MF
recipe persistence added a std::string there.
The rewrite replaces prose estimates with counted figures: 138 new files,
23 modified upstream files at +457/-75, nothing deleted, 99.3 % of the diff
in new files. That reframes the ask from "adopt a CAD kernel" to "widen a
build flag you already carry", which is the argument that actually has a
chance upstream.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Selection failures on a real desktop kept looking identical from the UI
whether the ray missed the solid, the click was rejected as a drag, or the
press never reached the tool at all. The status line added in edb1adbfa3
reports WHICH body/face was picked, so it distinguishes "picked" from
"silence" and nothing finer -- not enough to tell those three apart.
This narrates the whole press -> release -> ray path on stderr, one distinct
line per failure mode:
down x= y= the press reached the tool
up with no pending press the press was eaten upstream
up ... drift=N / rejected the click-vs-drag threshold decided
ray tris=N -> body= face= t= the ray reached the solid (or missed)
no solid data (bodies= mesh=) the pick pointers were never wired
Gated on getenv("SNAPORCA_PICK_TRACE"), cached in a function-local static,
so a normal build pays one load and prints nothing. It ships enabled-on-
demand because the failing environment is a real X session with a real
mouse, which the headless rig cannot reproduce.
First use retired a wrong theory of my own: the trace showed drift=0 on
three consecutive picks, so the widened threshold in edb1adbfa3 was not
what fixed anything. The reported symptom is best explained by a stale
pre-orcawidgets binary left running alongside the current one.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Ports snaporca-cad f5c7e74e9b.
The LeftUp pick discarded anything moving more than 4 px total since the
press. A hand-held mouse drifts that much during an ordinary click, so
real clicks were thrown away as drags and it read as "selection does not
work". Use 8 px per axis, GTK's own drag threshold.
Also report the pick on the status line (body / face / edge). A solid
pick previously set no text at all, so its only feedback was the viewport
highlight, and a pick that registers but draws faintly looked identical
to one that never fired.
DesignSketchTool.cpp copied verbatim (identical between the forks apart
from this change). DesignPanel.cpp took the status hunk only, since this
fork keeps mainline's Item-based DropDown in feat_dropdown/ToolFlyout.
Not confirmed on hardware yet, on either fork; this fork remains
uncompiled (needs Eigen 5.0.1).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Ports snaporca-cad c741fb9677.
SketchInlineEditor::open() clamped its position with
wxGetClientDisplayRect(), which describes only the PRIMARY monitor. On a
multi-head desktop (the reporting machine runs 5760x1080 across screens
at +0, +1920 and +3840) a field anchored on the left or right screen was
clamped onto the middle one and left invisible, while m_awaiting_length
made the sketch tool consume every mouse event until it was answered:
orbit and pan died after any sketch, with Enter the only way out. Clamp
to the display the anchor is actually on instead.
Also let drags and the wheel through that freeze, so a field that lands
somewhere unexpected degrades to odd placement rather than a dead
viewport.
Both files were byte-identical between the forks apart from this change,
so they are copied verbatim. Verified on the snaporca side by the user;
this fork is still uncompiled (needs Eigen 5.0.1, which the available
deps image does not provide).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Mirrors snaporca-cad a016dffb48 + 4b828c4798 so the two forks do not drift:
- Sidebar on Orca's widget set: ComboBox for all 25 pickers, StaticBox
frames for the tool cards / feature tree / bodies list, framed double
spins, and no stack of full-width action buttons (Prepare's left panel
is parameters only).
- Toolbar: document group + undo/redo + mode-gated tools, commit far
right, ordered via keyed slots, at Prepare's 40 px / 4 px geometry.
- Polygon options and a new Move/Rotate card (distance, axis, angle) in
the sidebar instead of inline in the toolbar row.
- DesignSketchTool: stop consuming the LeftDown over a solid. Orca starts
a rotate drag on the press, so swallowing it killed orbit/pan whenever a
body was on screen. The pick now resolves on LeftUp within 4 px.
DesignSketchTool.{cpp,hpp} were byte-identical across the forks and are
copied verbatim. DesignPanel.cpp needed one hand-merge: mainline's
DropDown is Item-based (DropDown::Item{text,tip,icon}, DropDown(items&))
where snaporca passes parallel vectors, so feat_dropdown/ToolFlyout keep
the mainline form. Checked field-by-field against this fork's
Widgets/DropDown.hpp.
scripts/docker-iter-build.sh: mount the root CMakeLists.txt and cmake/
rather than inheriting the baked copies, which silently drops the
SLIC3R_CAD gate, and stop defaulting to snaporca's build volume - sharing
it made the two forks overwrite each other's cache and binary.
NOT COMPILED. This fork needs Eigen 5.0.1 (find_package(Eigen3 5.0.1
REQUIRED)) while the available snaporca-deps image supplies 3.3, so it
needs its own deps build to verify. The behaviour above was verified only
on the snaporca side.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The big upstream merge replaced the old per-idle unconditional plate-selector toolbar
refresh with a dirty flag (Plater::mark_plate_toolbar_image_dirty), which the
geometry-change sites already set. PartPlate::update_slice_result_valid_state flips a
plate's slice-ready / IMEX blocked-plate ("naughty plate") state without any geometry
change and did not set the flag, so after the merge nothing repaints that plate's
thumbnail or warning badge — the old per-idle refresh used to mask it. Mark the toolbar
image dirty when the state actually changes, matching the established pattern.
Cosmetic; no effect on slicing. Surfaced by the merge review of 58b4a68a10.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Resolved 4 conflicts:
- PrintConfig.hpp / Preset.cpp: upstream restored calib_flowrate_topinfill_special_order
(removed in the prior merge, now re-added and re-registered); kept it alongside our
IMEX options (imex_parallel_mode, imex_head_filament_map).
- Plater.cpp: kept our <sstream> include plus upstream's <optional> and plugin includes.
- tests/fff_print/test_gcodewriter.cpp: both sides appended disjoint scenarios after a
shared base. Reconstructed as upstream's full file (base + its 3 toolchange/H2C
scenarios + the boost/filesystem include its helper needs) followed by our 10
set_pressure_advance / set_temperature scenarios. 22 scenarios total, no duplicates.
Design drew raw OS wxCheckBoxes (grey square + inline text) where Prepare shows
Orca's teal check, one of the most visible reasons the two panels looked
unrelated. The six sidebar checkboxes (extrude flip, hole through, thread
internal, revolve flip, boolean keep-tool, loft ruled) now use Widgets/CheckBox:
the widget carries no text, so each label moves into the row's left column,
which is also Prepare's row idiom and matches the label/control grids.
CheckBox is a wxBitmapToggleButton, so its per-control Binds and the panel-wide
preview refresh listen for wxEVT_TOGGLEBUTTON as well; boolean/loft get a
label+control row instead of a bare full-width control. Checkboxes align to the
left edge of the control column, as Prepare aligns its own.
The two sketch-toolbar checkboxes are left alone — they live in the top toolbar,
not the left panel.
Verified on :10: Flip direction renders as a teal check with a white tick and
toggles correctly.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Second half of the Prepare alignment. The 14 tool forms were plain 2-column
wxFlexGridSizers whose control column never grew, so every control sat at its
natural width right next to its label — nothing lined up, and it looked nothing
like Prepare's "label ......... [value]" rows.
- two_col_form() builds the grid with a growable control column; all 14 forms use
it and their 56 controls are added with wxEXPAND, so controls fill one aligned
column at the panel edge.
- The sketch-session Plane row is a box sizer, not a grid, so it gets a stretch
spacer for the same effect.
- The DoF readout collapses when empty, and the Bodies block starts hidden — both
reserved a blank line on a fresh document, leaving dead space above the action
buttons that Prepare does not have.
Verified on :10: Extrude edit shows Extrude dist / End / 2nd dist / Taper /
Result as aligned label+control rows; Plane row right-aligns; no Bodies box on an
empty document.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The two tabs used different idioms for the same concepts, which reads as two
different apps: Design was 264 px against Prepare's ~467 (the canvas edge jumped
on every tab switch), used bare micro-labels where Prepare uses icon + Head_14
card headers with a rule, hung its row actions on a loose strip under the tree
instead of in the section header, and drew raw OS-default wxButtons next to
Prepare's Orca-styled ones.
- Width now tracks Prepare's sidebar at runtime (sync_sidebar_width() reads the
live width on tab activation) rather than being hardcoded, so the two cannot
drift apart if Orca changes its sidebar.
- "Feature tree" and "Bodies" use the card_header() helper the panel already had
(icon + Label::Head_14) plus a wxStaticLine, exactly as the tool cards do.
- The six row actions moved into the Feature tree header, Prepare-style, at
header weight (24 px) instead of 36 px control weight.
- Buttons are Orca Buttons (ButtonType::Expanded, full width); Commit to Plate
gets ButtonStyle::Confirm as the tab's primary action.
- Margins/spacing come from SidebarProps (ContentMargin/TitlebarMargin/
ElementSpacing) instead of hardcoded 12/6/4.
Verified on :10: sidebars are the same width and share the header idiom.
Still to do: label-left/control-right rows inside the tool dialogs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Bodies were appended INSIDE the feature tree, after the features, and only when
there was more than one body. Two consequences:
- With a single body — sketch + extrude, the common case — no body row existed
at all, so the solid could not be selected from the tree. That also blocked
Move (it requires a selected body), the show/hide eye and every body-targeted
op; the viewport was the only way to select.
- With several features the Bodies group was pushed past the tree's auto-sized
height (capped at 9 rows) and clipped out of view, so bodies became
unreachable as history grew.
Bodies now live in their own list under the feature tree, mirroring Onshape's
Features + Parts split that the rest of the tab already follows. The list is
hidden while empty, sizes to its content (scrolls past 6), keeps the selected
row across a recompute, and greys hidden bodies as before. Selecting in either
list clears the other, so only one thing is ever "the target".
Verified on :10: single body -> Body 1 listed, selectable, Move opens the gizmo
on it (previously impossible); two imported solids -> both listed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The Design gizmo already had both move arrows and rotation rings, but drew them
at a fixed 70 px arm regardless of the body: on a 40 mm cube everything
collapsed into a ~100 px tangle buried inside the solid, so the rings were
effectively invisible and the tool read as "move only, no rotate".
Orca's Prepare gizmos size themselves from the selection's bounding sphere
(GLGizmoRotate3D: m_radius = Offset + sphere radius) so the handles always clear
the object. Same rule here: DesignPanel passes the body's bounding-sphere radius
(scale-aware) into the gizmo, and move_gizmo_arm() returns
max(70 px, 1.25 * radius) — the screen-space floor keeps it grabbable on a tiny
body or when zoomed far out. Ring radius and BOTH hit-tests derive from that one
helper, so picking cannot drift from what is drawn.
Verified on a 40 mm cube: rings now encircle the body, arrow drag moves with a
live mm readout, ring drag rotates live.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Resolved conflicts in PrintConfig.hpp and Preset.cpp. Both were adjacent to upstream's
removal of calib_flowrate_topinfill_special_order (superseded by the new
top/bottom_surface_fill_order options); took upstream's removal from the option class
and the print-options list while keeping our new IMEX options (imex_parallel_mode,
imex_head_filament_map) and the IMEX enums (ImexToolLayout, ImexVizTheme) alongside
upstream's new SurfaceFillOrder enum. The option remains in PrintConfig's legacy
ignore-set so old projects still load.
query_topology: indexing a body face-by-face was quadratic — face_by_index
re-walks the explorer and edge_by_index rebuilds the whole indexed map on every
single call. On a 15.7k-face / 25.6k-edge imported solid this blew past the
MCP 15 s main-thread timeout with the UI frozen throughout. GeometryEngine
gains faces_of()/edges_of(), which enumerate once in the very same order (ids
stay interchangeable with the _by_index accessors, so fillet/up_to_face targets
are unaffected). Measured on that body: 15 s timeout -> 0.46 s.
Feature ops: every commit-time m_doc.recompute() (fillet, cut, shell, boolean,
extrude, ...) now goes through recompute_guarded(), which runs the rebuild on a
worker thread. Live-preview/drag paths stay inline on purpose — yielding inside
a drag would be worse than the stall.
run_off_ui_thread(): the progress dialog is now created only after 300 ms, so a
fast op does not flash a dialog, while input stays blocked (wxWindowDisabler)
for the whole operation since the worker owns the document.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Viewport: GLCanvas3D::on_idle() cleared m_dirty even when
_refresh_if_shown_on_screen() rendered nothing because the canvas was not on
screen yet — a frame requested while the notebook was still showing a page got
silently swallowed and the viewport stayed blank until some later event dirtied
it again. _refresh_if_shown_on_screen() now reports whether it rendered, and
on_idle keeps the canvas dirty when it did not. Covers Design/Prepare/Preview.
Freeze: a big STEP (17.8 MB) spent ~40-50 s inside OCCT on the UI thread
(read_step_solids + recompute), so the window stopped repainting and the
compositor marked the app unresponsive. Both now run on a worker thread behind
an app-modal pulsing progress dialog: the UI keeps painting and the document
cannot be touched while the worker owns it. OCCT's Standard_Failure is not a
std::exception, so the worker catches it explicitly — an escaping exception
would terminate the process.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
request_repaint() only invalidates the canvas (Refresh()) on the hardware-GL
path and relies on a wxEVT_PAINT to follow. When the notebook re-shows the
Design page, that invalidation is issued mid-show and dropped: no paint event
arrives, the canvas never renders, and the pane stays blank until another tab
switch forces an expose. Software GL renders directly, so it never showed there.
force_repaint() defers past the show, then Refresh() + Update() for a
synchronous paint. Called from on_tab_shown().
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Follow-up to the cross-gantry mirror axis change, from an adversarial review of it.
GCodeViewer builds its own copy of the zone grid to place the sequential-preview
carriage markers, and it must reproduce PartPlate::calc_imex_zones exactly or the
markers drift away from the ghosts they are meant to track. It did not, in two ways,
because sizing the grid and placing a cell answer different questions:
- Sizing: calc_imex_zones counts every Copy/Mirror tool's OWN column, including the
non-representatives of an aggregated (Span) gantry -- they still donate a column.
GCodeViewer only ever saw the representative, so on an aggregated gantry it could
count fewer columns than the plate and lay its markers out against wider strips.
- Placement: calc_imex_zones PINS an aggregated cell to the primary's column, because
that row-strip spans the full bed and has no column of its own. GCodeViewer used the
representative's own column, which put the marker a strip away from the ghost
whenever the representative was not column-paired with the primary.
Track the two sets separately: grid_tool_ids sizes the grid from own columns, eff_col_of
pins only aggregated tools when placing. Out-of-grid tool indices are deliberately left
unfiltered -- calc_imex_zones drops them while calc_imex_ghosts keeps them, so no policy
here can agree with both, and a comment says so rather than pretending otherwise.
Also:
- The mirror-axis rule lived in three copies (two PartPlate lambdas plus an inline
re-derivation here). Hoist it to imex_mirror_axis_for() so the ghosts and the markers
cannot drift apart, and unit-test it, including degenerate tools_per_gantry.
- Drop imex_head_transform's mirror_axis default. A defaulted axis silently hands a
forgetful caller the X reflection, which is wrong for every cross-gantry tool and
fails silently -- exactly how a stale test kept certifying the old rule.
- Replace that stale test: it asserted a diagonal mirror "flips X only", the rule this
work overturned, and stayed green only because of the default.
- A secondary sharing the primary's gantry now takes the primary's Y box facing. The
box shows the side a tool could be hit from, and two tools on one beam can only be
hit by the same other gantry. No-op on the rear-* layouts, where the hardcoded value
already matched; on front-* layouts it pointed the box away from the only tools that
could reach it.
- Correct two comments that described the aggregated X-frame substitution as a
reflection plane. It is not one: it exists to zero gantry_offset.x, and removing it
would push aggregated ghosts a column off their strip.
Verified: 385/385 tests; CLI slice of the IMEX regression project is byte-identical to
the previous commit's G-code apart from the timestamp, confirming this is
visualization-only and cannot affect sliced output.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
catch_discover_tests registers each Catch2 test with ctest by name, and ctest then
re-invokes the binary passing that name back as a -# filter. The IMEX test names
contain em dashes (U+2014). They survive discovery, but on Windows the round trip
through the console codepage mangles them, so the filter matches nothing:
Filters: "imex_head_transform G-- copy mode is pure translation"
No test cases matched -> No tests ran -> exit != 0 -> ctest: Failed
All 85 IMEX tests were reported as failures on both Windows x64 and arm64 without a
single one of them ever executing. Linux and macOS are UTF-8 end to end and were
unaffected, which is why this went unnoticed since the names were introduced in
77c32a2e15.
These were the only non-ASCII test names in the whole tests/ tree. Renaming them to
plain hyphens costs nothing and keeps the suite portable.
Test names only -- no assertion, no logic, no comment is touched.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A Mirror tool reflects across the boundary it shares with the primary's zone, and
which boundary that is depends on where the tool sits:
- Same gantry: the tools are side by side along X, so the shared boundary is
vertical and the reflection negates X. This is what single-gantry IDEX does,
and it was the only case the code modelled.
- Different gantry: the zones are stacked along Y (front strip vs rear strip),
so the shared boundary is horizontal and the reflection negates Y. The part
that comes off gantry 1 is a Y-reflection of the tool directly behind it.
imex_head_transform() hardcoded diag(-1, 1, 1) for every mirror, as its own TODO
acknowledged. Lift the axis to a caller-supplied ImexMirrorAxis; PartPlate picks it
from the tool's gantry row. Both reflections keep det = -1, so a mirrored part stays
a true mirror image rather than a 180-degree rotation, which would print the
primary's part merely turned around.
The correct axis removes two workarounds. Both ghost paths special-cased aggregated
mirrors to "drop the X reflection, translate 1:1 and bake the flip into the mesh"
because reflecting X pushed the ghost off-bed as the primary was dragged. With a Y
reflection the X translation is already zero for aggregated tools, so that falls out
for free and the special cases are deleted.
Preview markers follow the same rule, which also fixes two placement bugs:
- Mirrors reflected across a Copy tool's zone edge, falling back to the primary's
column when a row had no Copy. In iq-mirror (0:P,1:C,2:M,3:M) the front row has
no Copy, so t2 and t3 both fell back and computed the identical X — both drawn
on top of each other in t3's zone. A mirror now reflects within its own zone.
- The toolhead footprint box flipped to the far side of the nozzle for any mirror
right of the primary. That only holds for an X-axis mirror, which reverses the
carriage's orientation; a cross-gantry mirror keeps the X orientation of the
tool behind it, so its box stays on the same side.
Tests cover the cross-gantry and diagonal cases, that the axis is caller-supplied
rather than inferred from the offset vector, and that both axes are reflections
(det = -1) rather than rotations.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Print::apply() derives physical_extruder_map from printer_extruder_id and writes
it into new_full_config, but does so after m_ori_full_print_config is snapshotted.
The selector write-back path rebuilds m_full_print_config from that snapshot, so
m_full_print_config kept the unexpanded default while every subsequent apply
re-derived the expanded map. An unchanged config therefore diffed on
physical_extruder_map on every re-apply, and since that key is not handled by
invalidate_state_by_config_options() it fell through to the catch-all and
invalidated every step — forcing a full reslice on each apply.
Mirror the derived map into the snapshot so the two stay consistent.
Caught by the upstream test 'Selector write-back expands migrating filaments and
survives re-apply', which asserts a re-applied unchanged config is not
APPLY_STATUS_INVALIDATED.
Conflicts were all co-located additions rather than design collisions:
- GCode.cpp: adopt upstream's toolchange(filament_id, nozzle_id) signature and
per-variant set_config_index() while keeping the IMEX bare-T<n> suppression;
rebase the second-layer temperature loop's non-IMEX branch onto upstream's
get_filament_config_index() resolution.
- Preset.cpp / PresetBundle.cpp / PrintConfig.cpp: keep both sides' option-list
and enum-map entries.
- GLCanvas3D.cpp: upstream's printable_heights argument plus the IMEX ghost pass.
- PartPlate.cpp: keep <set> (still used).
- test_gcodewriter.cpp / test_3mf.cpp: keep both sides' test cases.
Resolves two conflicts:
- GLCanvas3D.hpp: keep both the IMEX ghost render declarations and
upstream's _render_wireframe_overlay().
- test_gcodewriter.cpp: both sides appended test cases to the same
region; keep upstream's origin/machine-limit tests alongside the
pressure-advance and temperature scenarios.
Opening an STL/OBJ in the Design pane now rebuilds it into a real OCCT B-rep
solid that the face/edge feature tools can operate on, instead of a print mesh.
GeometryEngine::mesh_to_brep is a native C++ port of mesh2step
(github.com/tommasobbianchi/mesh2step): vertices and edges are shared across
triangles at construction time (vertex cache by deduped index, edge cache by
unordered index pair), so no BRepBuilderAPI_Sewing pass is needed to rebuild the
topology afterwards, and watertightness falls out of the edge-usage counts for
free. An open mesh is returned as a shell and reported as such — never dressed up
as a fake solid.
It runs in-process on the OCCT kernel libslic3r already links, so no STEP file is
written or re-read. That is not an optimisation but the whole point: a faceted
STEP of a 62k-triangle mesh is ~149 MB and OCCT's STEPControl_Reader takes >300 s
to parse it back, so routing this through a file would hang the GUI.
Coplanar neighbours are merged (ShapeUpgrade_UnifySameDomain, 5° default) so the
body arrives with pickable CAD faces rather than one face per triangle — on the
20,656-triangle test part that is 20,614 faces down to 4,784. Without it the
import is technically a solid but nothing you can meaningfully fillet or extrude.
- Design pane: "Import mesh" button + Shift+M; warns above 50k triangles.
- MCP: import_mesh {path, tolerance, merge_angle_deg}, returning the full
conversion stats so a caller can tell an honest solid from an open shell.
- Catch2: cube round-trip (exact volume, 12 faceted faces, 6 after merge), open
mesh stays a shell, and the scale-independent sliver rule that a naive
area < tolerance^2 test would get wrong.
Verified end-to-end on the real 20,656-triangle ir3v2 hotend STL: reproduces
mesh2step's Python run exactly (20,614 kept, 42 degenerate, 0 boundary edges,
2 non-manifold edges, not watertight) and the resulting body's bbox matches the
one FreeCAD reports for the same part.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Gate the entire parametric Design/CAD subsystem behind a single CMake
option so the fork can be built with or without it. With SLIC3R_CAD OFF
the build is behaviour-neutral against upstream OrcaSlicer; this is the
"parallel build" for upstream integration discussion.
Gated surface:
- option(SLIC3R_CAD) + add_definitions(-DSLIC3R_CAD)
- deps/OCCT/OCCT.cmake: BUILD_MODULE_ModelingAlgorithms=${SLIC3R_CAD}
(OFF matches upstream OCCT exactly; ON adds only TKFillet + TKOffset)
- TKFillet/TKOffset link, add_subdirectory(slvs) + libslvs
- CAD kernel + GUI sources, GLGizmoPrimitive (needs GeometryEngine),
CAD Catch2 tests
- 11 upstream C++ hook sites (GLCanvas3D, MainFrame, GLGizmosManager)
- TabPosition and gizmo EType enums switched to implicit numbering so
OFF reproduces upstream indices exactly
Verified on behemoth both ways: OFF and ON link snapmaker-orca +
libslic3r_tests (exit 0); ON shows the Design tab and passes 8 [design]
+ 1 [Deviation] tests, OFF omits them and shows upstream's tab layout.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The previous guard only caught an empty glob. That is the wrong invariant.
The glob ships whatever the deps prefix holds, which is not the same as what
libslic3r links.
This fork sets BUILD_MODULE_ModelingAlgorithms=ON; upstream OrcaSlicer sets it
OFF. Upstream's DataExchange module pulls in most ModelingAlgorithms toolkits
transitively, but not TKFillet and TKOffset -- those two exist only because we
turned the module on. So a source build against a deps tree carried over from
upstream has 40 of the 42 OCCT DLLs. The glob copies all 40, the build
succeeds, and the slicer dies at launch with "error 126 (dependency not
found)". Reported by SoftFever, who named exactly those two DLLs.
CI is unaffected: the deps cache key is hashFiles('deps/**'), so flipping the
OCCT flag invalidated it and every shipped artifact has all 42.
Publish OCCT_LIBS from libslic3r as the single source of truth and check each
toolkit has a DLL, so the list can never drift from what we link. Verified
against a simulated stale prefix: the configure fails naming exactly
TKFillet.dll;TKOffset.dll, and passes when both are present.
Windows-only: every call site of the copy function is inside if (WIN32).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The glob that replaced the hand-maintained OCCT DLL list traded a loud
failure for a silent one. The old code named each DLL explicitly, so an
unpopulated deps prefix made file(COPY) error out at configure time. A glob
that matches nothing instead yields an empty list, copies nothing, and leaves
the install manifest without a single TK*.dll -- producing a package that dies
at launch with "error 126 (dependency not found)".
This only bites source builds whose CMAKE_PREFIX_PATH lacks bin/occt (OCCT's
install layout varies by version); CI populates it, so every shipped artifact
is intact. Restore the loudness rather than ship a slicer without OpenCASCADE.
Windows-only: every call site of the copy function is inside if (WIN32).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Romanian was absent from this fork and from upstream, so there was no
catalog to reuse. Machine-translated with a local qwen3.6 model against the
union of both forks' .pot files; 99.0% coverage (5472 singular + 9 plural).
Header credits it as machine translation pending native review.
Registering wxLANGUAGE_ROMANIAN in supported_languages[] is what actually
exposes the language: the Preferences dropdown filters the installed
dictionaries against that allowlist, so a valid .mo alone would ship a
language nobody could select.
Translations that failed to preserve their printf/boost placeholders were
left empty rather than shipped, since msgfmt --check-format is fatal in
run_gettext.sh and would break the build on all three platforms. Those 57
strings fall back to English. Verified: run_gettext.sh exits 0 and the
compiled catalog has 0 placeholder mismatches.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The portable/installer hardcoded an OCCT DLL list that drifted from what
libslic3r actually links (OCCT_LIBS) + their transitive OCCT deps. The
CAD build links TKFillet/TKOffset/TKBool (and pulls TKFeat/TKBin*/TKIGES/
TKRWMesh/TKSTL/TKVRML/TKXDEIGES/TKXml* transitively) which were absent
from the shipped zip -> OrcaSlicer.dll failed at launch with error 126
(dependency not found). Replace both hardcoded lists (the file(COPY) and
the install manifest) with a glob over the built occt/ dir so the package
can never drift from the build again. Windows-only: OCCT is Shared solely
on Windows (deps/OCCT/OCCT.cmake); macOS/Linux static-link it. Reported by Marc.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Hole: with no face explicitly picked, the tool fell back to the XY datum at
z=0 (the model's underside), so placing a hole from a top view read parallax-
shifted. Default to the solid's top face (top_face_index_of) so the footprint
sits on the surface being viewed; the XY/XZ/YZ dropdown still overrides.
Internal thread: the bore was re-cut at the nominal radius, which coincides
with an existing hole's wall — the coincident faces fouled the groove boolean
so it removed ~nothing (invisible thread). Cut the bore at the minor diameter
(radius - depth) instead: strictly inside any existing wall, leaving it clean
for the groove; on solid stock it forms the tap-drill.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Windows minwindef.h defines legacy 'near'/'far' as empty macros, so MSVC
mangled `auto near = ...` (C2513) and every near(...) call (C2679/C2678 on
Vec2d). gcc/clang were unaffected, so only the Windows build failed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
gcc reads "\x10cad..." as one escape (c/a/d are hex digits → 0x10CAD > 255).
Split the string literal so the \x10 escape terminates. clang let it slide;
gcc (Linux/Windows CI) errored and blocked the OrcaSlicer build.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
- Keyboard shortcuts (Onshape-style, three scoped layers): feature tools on
Shift+letter, 2D sketch tools on single letters (in-sketch), view/nav on
single letters (out-of-sketch). Dispatched via the DesignPanel CHAR_HOOK.
- View toggles: P = origin planes, A = world axis triad (render_view_helpers).
- Section view (non-destructive): a single horizontal clip that hides half the
model to inspect inside, showing only the solid remaining half (no ghost).
Left-panel "Section View" button or X toggles it; PageUp/PageDown move the
plane; "Flip Section" button or F shows the opposite half. Never a body/Cut.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Six fixes to the Design tab, all live-verified on :10:
- Datum-plane re-pick: wxEVT_CHOICE on m_draw_plane re-planes the live sketch
- Delete-feature dismisses its lingering settings card (on_delete_feature)
- Revert rotation-orbit regression (no feed_bodies/reload on close_tool)
- Sketch undo/delete via focus-independent CHAR_HOOK (Ctrl+Z/Y, Delete)
- Fix undo hang: reset_autoedit() clears dangling auto-edit sequence
- Titled value fields: each inline dim field shows its role (Width, Height,
Radius, Angle, Length, Side, Distance, Major, Minor)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Closes the gizmo-parity gap (bd snaporca-4h5): the five solid features that were
card-only now give in-canvas feedback like their siblings.
Draft — angle-arc drag gizmo (clone of the Revolve gizmo): once a side face is
picked, a cyan arc anchored at the face centroid shows the taper angle; drag the
tip or type the angle, the live ghost tapers with it. Axis = world +Z (the neutral
pull direction), clamped [-89, 89].
Cut — plane offset-arrow + cutting-plane rectangle (clone of the Shell arrow, adds
a wire rectangle in the cut plane sized to the target body bbox). The arrow drags
the signed offset along the plane normal; the rectangle rides at the cut position;
the ghost splits live. (Also covers bd snaporca-1gh / snaporca-mmr.)
Boolean / Sweep / Loft — operand highlighting (new by-index highlight infra):
- Boolean tints the target body teal-green and the tool body orange (per-index
body tint added to the DesignCanvas GLVolume colour loop + set_operand_bodies).
- Sweep tints the profile sketch cyan and the path sketch magenta.
- Loft tints every selected profile sketch green.
Sketch tints reuse the DisplaySketch overlay via a feature-index -> colour map
(sketch_hl_color); DisplaySketch struct unchanged. All self-gate by active tool
and clear on close_tool.
Wiring mirrors the existing gizmo pattern 1:1 (m_*_active / render_* / set_* /
clear_* / update_* in refresh_preview / DesignCanvas passthroughs / render dispatch
+ on_mouse drag branch). Both forks; DesignSketchTool.{cpp,hpp} + DesignCanvas.hpp
+ DesignPanel.hpp byte-identical across forks. Built clean on both. Draft, Cut and
Boolean highlight live-verified on :10; Sweep/Loft sketch tint is code-complete and
build-clean (visual check pending — needs hand-drawn profile/path sketches).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Sketch dimension labels — now identical to the Prepare/Preview Measure gizmo:
- draw_text repurposed to draw_dim_label: white ImGui text in a translucent-white box,
mirroring GLGizmoMeasure::render_dimensioning exactly (push_common_window_style sets the
text colour, BringWindowToDisplayFront, imgui_internal.h).
- ROOT-CAUSE FIX: world_to_screen_px multiplied two Eigen Transform3d objects
((proj * view).matrix()); a projection is not affine so Eigen mangled it -> garbage screen
coords, so labels never appeared. Now proj.matrix() * view.matrix() like Measure. (That
helper was previously [[maybe_unused]] dead code, never exercised.)
- Leaders: offset clear of the sketch line (no longer coincident with the geometry), single
point-to-point dimension line + arrows, neutral colour, width 0.6 -> 0.2.
- dim_text appends mm/in on linear dims (angles keep the degree sign).
New Design + delete:
- New "New Design" button wipes the whole document (confirm dialog) — the clear-all the
per-row Delete can't give. CadDocument::clear() now also clears bodies + display_body_meshes
(it left them stale, so solids lingered after a clear).
- on_delete_feature: a Body-row selection now shows a helpful hint (bodies are recomputed
results with no directly-removable feature) instead of silently doing nothing.
Feature tree: auto-fits its content (refresh_tree clamps height 1..9 rows, scrolls past),
instead of a fixed 140px block.
i18n (Design tab pinned English, per the UX contract):
- Restore the lost #undef _L / #define _L(s) wxString::FromUTF8(s) override atop DesignPanel.cpp;
wrap all ~54 dropdown options in _L so the single lever governs them. feature_type_name left
untranslated (machine-facing MCP JSON).
UX: per-card Value Confirm/Cancel buttons removed — the single ribbon action bar owns value
confirm/cancel via an m_value_cont guard in tool_confirm/tool_cancel. "needs a body" status
messages unified.
Both forks; DesignSketchTool.{cpp,hpp} + CadDocument.cpp byte-identical across forks. Built
clean; New Design / feature-delete / rotation / tree auto-fit live-verified on :10.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Dimensions (uniform with Prepare/Preview):
- Repurpose DesignSketchTool::draw_text -> new draw_dim_label that renders each
sketch dimension as the exact Prepare "Measure" gizmo label: white ImGui text
in a translucent-white box, positioned via the existing world_to_screen_px
projection inside the active ImGui frame. All 29 label call sites convert with
no churn; the bespoke Hershey vector font is retired.
- dim_text appends mm/in on linear dims (angles keep the degree sign).
- Placed-dimension leaders simplified to a single point-to-point line + arrows in
a neutral colour (no extension lines), matching the Measure look.
i18n (Design tab pinned English, per the UX contract):
- Restore the lost "#undef _L / #define _L(s) wxString::FromUTF8(s)" override atop
DesignPanel.cpp so one lever de-translates the whole tab, ending the half-EN/IT
state. Wrap all ~54 dropdown options in _L so the single lever governs them.
- feature_type_name left untranslated (it feeds the MCP JSON, machine-facing).
UX consistency:
- Remove the per-card Confirm/Cancel buttons from the Value card; the single ribbon
action bar now owns value confirm/cancel via an m_value_cont guard in
tool_confirm/tool_cancel (one confirm surface, per contract).
- Unify the eight divergent "needs a body" status messages to one template.
Both forks; DesignSketchTool.{cpp,hpp} byte-identical across forks. Built clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
orca_cad's test_geometry.cpp is upstream Catch2 v3 with no CAD suite, so the
surface_deviation check lives alongside the CAD tests here. Built + ran green
in orca_cad's kernel-test docker: All tests passed (4 assertions). Both forks
now carry the test.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Reconstruction is no longer box-only. Five new actions, all thin
wrappers over existing CadDocument::add_* (the same kernel the GUI
calls): revolve (sketch -> add_revolve), fillet/chamfer targeted at a
measured edge id from query_topology, hole (add_hole), and boolean
(union/subtract/intersect over two bodies). extrude and revolve also
accept an optional closed profile=[[x,y],...] -> add_sketch_profile,
the Measure->Build bridge: feed a measured/sliced contour straight back.
Shared helpers plane_from / bool_from / profile_from. Each action is
transactional (checkpoint -> add -> recompute -> undo on failure ->
refresh) and returns post-state. Live-verified: fillet, chamfer (clean
edge), hole, profile-extrude (L), revolve (ring), boolean union (3->2
bodies). MCP method catalogue now 13.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Add the Build-input and Validate ends of the reverse-engineering loop.
import_step loads a STEP as native B-rep bodies (the reference part to
measure), reusing the GUI Import path (read_step_solids -> Import
features -> recompute). validate_against compares a body to a reference
({step:path} | {body:id}) and reports volume delta %, bbox delta, and
centroid offset via OCCT GProp + Bnd_Box — the RE skill's actual
acceptance metric (the "scarto %"); surface-deviation heat-map is the
upgrade path.
With this the full Understand -> Measure -> Build -> Validate loop is
live: 8 methods (describe_tools/describe_scene/query_topology/measure/
slice_body/import_step/validate_against/extrude). Verified: import a STEP
then validate body vs same STEP = 0.0%% delta / 0.0mm offset.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
The "evidence" half of the reverse-engineering loop — read-only
perception that turns a body into measured numbers. query_topology
lists faces (centroid/normal, cylinder radius+axis when round) and
edges (length, circle radius); measure returns distance (and angle when
both refs have a direction) between two {face|edge|point} refs;
slice_body cross-sections a body by a base plane at an offset and
returns world polylines (sections-as-evidence). All reuse GeometryEngine
topology accessors; slice_body adds one BRepAlgoAPI_Section call.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Zero-dependency Python stdio MCP server that exposes the app's control
socket as first-class MCP tools. Builds the tool list live from the
app's own describe_tools reply (introspection drives the schema), and
forwards tools/call to the Unix socket. Falls back to the slice-1 tool
set and reports a clear error when the app socket is down — never
crashes. Registered via a workspace .mcp.json (ssh stdio to the build
host); activates on the next session.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Predispose the Design tab to run under an external MCP agent. New
McpControl.{cpp,hpp} embeds a line-delimited JSON-RPC 2.0 server over a
Unix domain socket, off unless env SNAPORCA_MCP is set. Requests are
marshalled onto the wx main thread and run through the SAME CadDocument
kernel the GUI uses, via two thin DesignPanel hooks (mcp_doc /
mcp_after_change) — no parallel engine.
Slice-1 methods: describe_tools (introspection -> the bridge builds tool
schemas), describe_scene (feature tree + per-body bounding boxes), and
extrude (centred rectangle sketch -> new solid). Every op returns its
full post-state. POSIX-only; Windows compiles to a no-op.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Fusion 360 .f3d uses Autodesk's closed ShapeManager kernel — no offline
reader exists and it is input-only even in Autodesk's own cloud API, so
native import/export is infeasible. Intercept .f3d at priv::load_files
(the single funnel for drag-drop and File > Open) and show a dialog
pointing the user to export STEP from Fusion, then load any other files
in the batch.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Add an "Export STEP…" button to the Design panel that writes every body to a
.step file as native B-rep (not mesh).
- CadDocument::export_step: compound all bodies (applying their per-body Move
display transform so the STEP matches what Commit ships) and write via OCCT
STEPControl_Writer (AsIs). Full error handling incl. OCCT Standard_Failure.
- DesignPanel::on_export_step: bake any open preview, wxFileDialog save, export
at the displayed body positions, status feedback.
Kernel write path verified with a standalone OCCT box->STEP->readback check
(1 solid, non-null) against the same OCCT build.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Boolean features can now be re-edited from the feature tree (op, target/tool
body, keep-tool, fuzzy tolerance), funnelling through the same replace_feature
path as every other editable feature.
- on_edit_feature: route CadFeatureType::Boolean to the Boolean card.
- load_feature_into_dialog: populate the Boolean card from the saved feature.
- populate_body_choices(as_of_feature): when re-editing, list the bodies as they
existed just before the boolean (replay the recipe truncated to that slot) so a
consumed tool body still appears and the saved target/tool selections round-trip
instead of collapsing to one entry.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Port of the snaporca CAD work to the mainline fork.
- Onshape default planes (XY/XZ/YZ) at the bed centre (transparent, labelled);
modeling origin unified to the bed centre (CadDocument::modeling_origin);
world-axis triad moved to the bed centre on the Design canvas only.
- Datum plane: clickable ghost-plane base pick + draggable offset arrow.
- Slot: dims reassessed to inter-centre distance / radius / angle; fixed the
duplicate cap-arc radius quote.
- Hole: 3D cube move-handle on the face; binds to the face on the first click;
decluttered side-distance construction lines.
- Thread: derive the M spec (diameter/pitch/depth) from a picked cylindrical
surface or circular edge (GeometryEngine::circle_of_edge); fuse the helical
ridge onto the existing body; MakePipeShell fixed-binormal sweep (uniform, no
twist) + self-intersection/param guards.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Mirror of snaporca-cad 75045ff. Re-editing an Extrude spawned a NEW misplaced
box: build_candidate() rebuilt the feature from live tool state, but GUI
extrudes store their profile as `entities` (sketch_ref = -1) which the edit
card never restores, so the candidate had an empty profile and replace_feature
swapped in a degenerate extrude. Fix: when editing, seed the candidate from the
feature being edited and skip add-time structural re-derivation (profile source,
up-to-face, target body); the card still overrides scalar params.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Mirror of snaporca-cad a82cb0d. snaporca-88g — three [CadDocument] tests
proving recompute() replays the whole timeline, so editing any feature (not
just the last) rebuilds downstream: mid-timeline extrude edit -> fillet
rebuilds; first-feature sketch edit propagates the chain; the same survives
serialize_recipe()/deserialize_recipe(); a sketch->extrude->hole->chamfer
chain rebuilds hole+chamfer on a mid-edited extrude. All 3 pass (34 assertions,
Catch2 v3). on_edit_feature edits the tree-selected feature at any position
(13/16 types); Import/Boolean/Cut get a status message instead of a silent
no-op (full dialogs = follow-up snaporca-nu9).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Mirror of snaporca-cad f3b665b. DesignCanvas had ~38 call-sites doing
`set_as_dirty(); render()` directly, tagged "llvmpipe: force repaint" —
software-GL workarounds from the headless :10 test box. On hardware GL this
bypasses the dirty/refresh cycle and burns frames outside the paint event.
Funnel all 38 through a new request_repaint() helper that branches on the
cached GL renderer string: hardware GL gets set_as_dirty()+wxGLCanvas::Refresh()
(render() runs inside the paint cycle), software GL keeps the direct render(),
undetected backend takes the safe direct path. Behaviour on the llvmpipe :10
box is byte-identical; hardware GL invalidates the canonical editor way.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Mirror of snaporca-cad 0f53a09. DesignPanel.cpp pinned every _L() to
wxString::FromUTF8 via a TU-local #define, hardcoding English and bypassing
the gettext catalog — a hard review-blocker for upstream (OrcaSlicer ships
~20 locales). Remove the override so the 463 _L("...") call-sites route
through the real Slic3r::GUI::I18N::translate (wxGetTranslation). All args
are string literals; untranslated strings fall back to the source msgid, so
English is byte-identical while other locales now translate when .po entries
exist. xgettext scans source text, so .pot extraction is unaffected.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Mirror of snaporca-cad ad822bf. The earlier mirror (2c0140afb9) implemented
persistence in the PrusaSlicer 3mf.cpp, but the GUI saves/loads projects via the
BBS-native backend (store_bbs_3mf / load_bbs_3mf), so the recipe was never
written to nor read from GUI-saved projects.
- bbs_3mf.cpp: BBS_CAD_RECIPE_FILE = "Metadata/SnapOrca_cad.bin"; writer
_add_cad_recipe_file_to_archive (called after layer-height in store_bbs_3mf);
reader branch in _load_model_from_file's metadata dispatch loop (iterate +
iequals on m_filename — mz_zip_reader_locate_file does not work on these
archives).
- Plater.cpp: load_files carries the Model-level cad_recipe onto q->model().
- test_3mf.cpp: [3mf] test asserting store_bbs_3mf embeds the recipe entry
byte-for-byte (read back via miniz, Catch2 v3).
Verified on behemoth: libslic3r_tests + the new [3mf] BBS test pass; orca-slicer
GUI links clean. Round-trip verified live on snaporca-cad (identical code path).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
DesignPanel's per-body colour button references icon "color_palette" but the
asset was never ported with the CAD subsystem. On mainline OrcaSlicer the
missing bitmap throws during MainFrame construction → segfault at startup
(create_scaled_bitmap "Could not load bitmap: color_palette"). Copied from
snaporca-cad resources; runtime-loaded, no rebuild needed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Catches the iXex/IDEX branch up to upstream main (44 commits). One
content conflict resolved:
- src/libslic3r/Print.hpp: kept upstream's default-initialized
m_origin {0,0,0} alongside our m_imex_slice_offset member.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Dual-fork mandate: port the parametric-recipe 3MF persistence from
snaporca-cad commits ed19eac+147e1c4 so a saved project reopens with the
editable CAD feature tree, not just the baked mesh.
Shared kernel/format/GUI (identical to snaporca-cad):
- CadDocument serialize_recipe/deserialize_recipe (cereal BinaryArchive,
versioned) + CadFeature split save/load + imported_solid<->BRep string.
- Model::cad_recipe carried through 3MF zip entry Metadata/SnapOrca_cad.bin
(writer + binary-verbatim reader branch).
- DesignPanel on_commit() stamps the recipe; on_tab_shown() rehydrates a
loaded project via load_recipe() (deserialize -> feed_bodies + refresh_tree).
Mainline-only adapters (no snaporca-cad counterpart — Catch2 v3 vs v2):
- tests/libslic3r/test_caddocument.cpp: <catch2/catch_all.hpp> +
`using Catch::Approx;` (v3 scopes Approx under Catch::).
- tests/libslic3r/CMakeLists.txt: register test_caddocument.cpp (the
original CAD port had left it out of the test build).
Verified on behemoth (snaporca-deps toolchain): libslic3r_tests clean;
[CadDocument] 17/18 (only the pre-existing tangent-to-circle SIGABRT fails,
identical to snaporca-tkz); K1 serialize round-trip + version-reject pass
(13 assertions); new [3mf] "CAD recipe blob survives a 3mf save/load cycle"
passes byte-for-byte; orca-slicer GUI links clean (186/186, DesignPanel.cpp
compiled). Interactive :10 click-through pending (no Design-tab automation).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Catches the iXex/IDEX parallel-printing branch up to upstream main
(102 commits). Two content conflicts resolved:
- src/libslic3r/Preset.cpp: s_Preset_printer_options — kept upstream's
new "use_3mf" key and our iMEX printer-capability/mode keys.
- tests/fff_print/test_gcodewriter.cpp: upstream revived the disabled
suite (#14196), dropping the obsolete [.]-tagged lift() test and its
config_lift_unlift.ini; kept their set_speed + z_hop tests and appended
our 10 per-firmware set_pressure_advance/set_temperature scenarios.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
* fix tree support brim
* treesupport3d part 1: more diagnostic logging. (todo once things are fixed: remove this / gate it properly)
* make area under Z=0 in rotated slice pipeline not solid
* fix solid Z=0 layer for belt printers
* fix renderer
* clean up logging
* final review pass
# Description
<!--
> Please provide a summary of the changes made in this PR. Include
details such as:
> * What issue does this PR address or fix?
> * What new features or enhancements does this PR introduce?
> * Are there any breaking changes or dependencies that need to be
considered?
-->
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
<!--
> A guide for users on how to download the artifacts from this PR.
-->
[How to Download Pull Requests Artifacts for
Testing](https://www.orcaslicer.com/wiki/how_to_download_pr_artifacts)
The Cartesian designed-view preview over-extended the toolpaths past the model
shell by a height-proportional amount (up to ~20mm tall parts), most visibly on
long multi-part prints; compact parts like a calibration cube looked fine.
Two coupled causes:
- Belt start G-code that primes with a Z advance and a 'G92 Z0' reset leaves a
constant machine-Z origin in the GCodeProcessor, so move positions are stored as
gcode_Z + origin. The linear back-transform mixes that constant with the
gantry-Y term, leaving a per-move designed-Y error that min-corner anchoring
cannot cancel when an elevated move (e.g. a bridge) happens to cancel it at the
bbox minimum. Expose GCodeProcessorResult::belt_z_origin (the m_origin[Z] left by
the start G-code) and subtract it before the back-transform.
- Elevated features (bridges/overhangs) are mis-mapped by the linear inverse to
outside the model body; build the anchor bbox only from moves within model_bb +/-
10mm, with a fallback to the full bbox when the clip would drop the bulk (object
placed away from the belt entry) so the gross-offset case still anchors.
Preview-only; G-code output is unchanged.
The belt designed (upright) preview back-transforms the machine-frame G-code
into model space with the linear belt inverse. That inverse recovers the
print's shape and orientation, but not the per-object placement/lift
translation: the object's position on the belt, the BeltSliceStrategy min-Z
lift, and the centering pre-translate are applied OUTSIDE
build_forward_transform() (see PrintObjectSlice.cpp), so its linear inverse
cannot undo them. The result was a constant offset (~20 mm on the belt-advance
axis) of the toolpaths from the model shell, on every model.
Recover the missing translation generally — independent of the offset's exact
source or the axis remap — by anchoring the back-transformed object body
(extrusions on layer_id >= 1, i.e. excluding the layer-0 prime/skirt) onto the
upright model bounding box, the same space the shells render in, and folding
that translation into the belt inverse before converting to libvgcode.
Replaces the previous Y=0 anchoring in LibVGCodeWrapper, which pinned the
toolpaths to the belt entry rather than to the model and so left the offset in
place for any object not sitting at the origin.
On a belt printer the emitted G-code is in the machine frame (45-deg sheared,
axis-remapped, scaled), so the toolpath preview shows the print as a sheared
slab floating off the bed. Map each toolpath vertex back to model/Cartesian
space for the "designed" view.
The back-transform is the inverse of the full G-code forward pipeline
(BeltGCodeWriter::to_machine_coords):
model = [BeltForward^-1 if !gcode_back_transform] . AxisRemap^-1 . MachineFrame^-1
built from config, so it handles any rotation / shear / scale / axis-remap
combination, not just plain 45-deg belt slicing. Computed in load_as_gcode()
from print.config() and applied per-vertex inside libvgcode::convert (display
position only; layer_id, times and the volumetric/flow math keep the raw
machine values, so the layer slider and stats are unaffected).
- Toggle with the existing "Show designed view" checkbox / hotkey B; off shows
the raw machine-frame G-code (useful for debugging the transform itself).
Defaults to on.
- Belt printers skip the same-result-id load cache so the upright view applies
and the toggle takes effect even when the G-code is unchanged.
- The object extrusions (layer_id >= 1) are anchored to the belt entry to drop
the constant machine-origin offset (start-G-code belt advance) that the linear
back-transform alone does not capture; start-G-code prime lines are excluded
so they don't steal the anchor.
Physical max-volumetric-speed test (belt #62 v4 asset) on the IR3 V2 with eSUN
PLA white: the wall stayed clean up to ~100 mm/s = ~20 mm3/s before
under-extrusion. The shipped cap of 10 mm3/s was ~half the real ceiling and
was silently throttling infill.
- eSUN PLA @IdeaFormer IR3 V2: filament_max_volumetric_speed 10 -> 20
- 0.20mm Standard @IdeaFormer IR3 V2: sparse_infill_speed 200 (~18 mm3/s at the
new cap, no longer throttled). Outer wall (45), PA (0.12), accel (1000)
unchanged — accuracy preserved.
- IdeaFormer.json version bump for profile-cache refresh.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Belt printers can't slice a tall vertical temperature tower. This adds a
belt-specific temperature-tower model — a row of discrete, individually
engraved provini laid along the belt, each printed at one temperature via
custom per-layer M104. Each provino is an inverted-L overhang that stresses
print quality, so the operator reads the best temperature off overhang
quality rather than a continuous ramp.
It is offered as a "Test model" choice in the temperature calibration dialog
(mirroring the Cornering test's selector), so users keep Joe's counter-rotated
sectioned tower as "Standard" and can pick this one as "Overhang":
- Calib_Params::test_model (existing field) carries the choice.
- Temp_Calibration_Dlg gets a Standard/Overhang radio.
- Plater::calib_temp belt branch: test_model 0 -> _calib_temp_belt_sectioned
(unchanged Standard path), 1 -> the discrete-provini Overhang path.
Assets: belt_temp_provino_unit.stl + belt_temp_tower_<start>_<end>.stl (6
ranges) + gen_belt_temp_tower.py (manifold engraving). Based on
belt/generic-calibrations. The Overhang path is HW-validated on the IdeaFormer
IR3 V2 (discrete M104 + engraved numbers); not re-validated since the rebase.
Enables supported printing of standard Orcaslicer calibration profiles.
* Build 2 Checkpoint
* fix support generation wedge, ghost layers
* flip cornering tests 180 deg to waste less supports
* fix row spacing on the flow ratio calibrations
* more testing, this didn't fix anything
* switched rotation tools, same issue
* fixed Z-offset issues
* add rest of PA features, may look a bit weird on a belt
* make temp towers work
* re-enable spiral on calibrations that want it
* Final cleanup pre-PR and community testing
The IdeaFormer IR3 V2 End G-code ran `G28 ; home all`, which homes the
Z (belt) and Y (gantry) axes. On a belt printer Z is the conveyor, so
homing it runs the belt all the way back to origin, dragging the finished
part back under the gantry that G28 has just lowered — the head knocks the
print (reported by an IR3 V2 user; the `G1 Y50` lift came after the G28,
too late).
Replace the end sequence with a belt-safe one: switch to relative mode
(G91), lift the gantry for clearance, advance the belt forward one full
machine-depth (Z676, the 676 mm product depth) to eject the part and cycle
the belt surface clean, then home X only — never the Z/belt axis.
collect_layers_to_print() warns (CRITICAL) when an extrusion layer sits above
the previous one with an empty gap below — the fixed-bed assumption that
material with nothing under it is floating and unprintable. On a belt printer a
*leading* empty range (the gap starts at Z=0, no prior extrusion layer) is not
floating: it is the conveyor lead-in, and the part rests on the advancing belt
as the first material is laid down well above Z=0. A part not designed for a
belt (e.g. a flat test model tilted into the belt frame) then trips this as a
false "Object can't be printed for empty layer between 0 and N" error.
Suppress only the leading case (belt_printer && last_extrusion_layer == null);
genuine internal gaps are still flagged, since on a belt those can be an
over-angle overhang printing into air. Non-belt output is unchanged.
The original PR skipped the max-print-height check entirely on belt printers
because the sliced (virtual) Z is belt travel, not build height. As the reviewer
noted, that removed the only working height guard. Restore a correct guard:
- Print::validate: on belt printers, compare the upright object height
(max over instances of the scene-space bbox) against printable_height directly.
printable_height is the usable VERTICAL clearance above the belt: the gantry
travels up the tilted plane (reach = height/cos(tilt)) and its axis range is
sized for that (IR3 V2: ~354 mm gantry travel = 250 mm vertical at 45deg, and
printable_height = 250). Hardware-confirmed 250 mm vertical clearance, so no
cos(tilt) factor is applied.
- BuildVolume::set_belt_printer: drop the diagonal Z scaling; the build-volume Z
already equals printable_height, keeping the live 'outside build volume'
highlight in agreement with validate().
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* Add IdeaFormer IR3 V2 belt printer profile
Self-contained vendor profile for the IdeaFormer IR3 V2 (45 deg belt printer):
machine (0.4 nozzle) + 0.20mm process + Generic PLA/PETG filaments, with the
belt machine-frame transforms set explicitly on the machine preset
(belt_printer, belt_slice_rotation x/45/global, build_plate_tilt_x=45,
gcode_remap_x/y/z, gcode_shear_z=pos_tan, gcode_scale_y=inv_cos).
The vendor bundles its own machine/process commons (fdm_belt_common,
fdm_klipper_common, fdm_machine_common, fdm_process_common) on purpose:
OrcaSlicer resolves system-preset inheritance per-vendor, so a profile that
inherits the Custom vendor's commons cross-vendor fails to resolve its parent
and the whole IdeaFormer vendor silently fails to load. Bundling the commons
(and listing them in IdeaFormer.json in dependency order) keeps the vendor
self-contained, matching how every other vendor folder is structured.
Machine limits, bed temperature (75 C for belt PLA) and start/end G-code are
taken from a working IdeaFormer IR3 V2.
* feat(belt/profile): eSUN PLA @IdeaFormer IR3 V2 — HW-calibrated belt filament
Add an eSUN PLA belt profile for the IR3 V2, inheriting Generic PLA @IdeaFormer
IR3 V2 (self-contained: parent is in the same IdeaFormer vendor, registered
after it in filament_list). HW-calibrated on the IR3 V2:
- nozzle_temperature 200/200 (temp-tower calibration)
- pressure_advance 0.12 (PA calibration)
- filament_max_volumetric_speed 10 mm³/s (max-vol-speed calibration: wall
failed at 126 mm/s → 126 × 0.0798 mm³/mm ≈ 10 mm³/s)
Ooze prevention currently treats every toolchange as a transition between
independent hotends — the OLD filament's extruder cools to a standby
temperature (via standby_temperature_delta or the filament's
idle_temperature) before the change, and the NEW filament's extruder
ramps back up afterward. This is correct for IDEX/toolchanger setups
where the parked nozzle would drip otherwise.
For AFC/MMU lane swaps where the SAME physical extruder stays selected
(only the loaded filament changes), the cool-down → re-heat round trip
is pointless: same nozzle, same heater, just a different filament feeding
it. In-print this costs 30+ seconds per lane swap, and the AFC tip-form
sequence ends up running on a cooling extruder.
Gate the pre_toolchange call on physical_extruder_map: when both the old
and new filament index map to the same physical extruder, skip the
standby cool-down. post_toolchange is left untouched — its M109 to the
new filament's print temp is still emitted, so per-lane temperature
differences (e.g. PLA → PETG on the same AFC manifold) are still handled.
Note on pem sizing: the guard requires physical_extruder_map to be sized
to the filament count for the per-filament lookup to succeed. The option's
registered default is a single-entry [0], which is shorter than the filament
count on any multi-filament setup, so the bounds check fails and ooze runs
as before. The fix fires only on profiles that explicitly author pem to
match filament count (the AFC/MMU/toolchanger configs that actually encode
same-physical routing).
Behavior matrix:
| Config | pem | Ooze behavior |
|---|---|---|
| Single-extruder + SEMM | (any) | `init_ooze_prevention` already disables ooze. No change. |
| Single-extruder, no SEMM | (any) | Single filament, no toolchanges. N/A. |
| IDEX / Toolchanger | `[0,1,…]` per-filament | Cross-physical → ooze runs as before. |
| Vanilla AFC, SEMM=true | (any) | `init_ooze_prevention` disables ooze. No change. |
| Vanilla AFC, SEMM=false | `[0,0,…]` per-filament | Same physical → **ooze SKIPPED** (the fix). |
| Toolchanger + AFC | per-filament | AFC swaps skip, cross-physical swaps run. |
| Default pem `[0]` (1 entry) | shorter than filament count | Bounds check fails for filament index ≥ 1 → ooze runs as before. No-op for profiles that haven't authored a per-filament pem. |
| Empty pem | `[]` | Guard returns false → ooze runs as before. |
The guard depends only on physical_extruder_map; no machine-class check.
Any printer whose pem is sized to filament count and indicates multiple
logical slots routed to the same physical hotend benefits.
The firmware-managed-zones writer offset puts prim_pos and the preview
toolpaths in a frame shifted relative to the bed bounds the zone math
uses, so secondaries would have rendered off the plate. Skipping the
secondary computation when firmware-managed is on keeps the centered
preview honest.
Addresses Felix14-v2 feedback on PR #13086.
124 upstream commits including CrealityPrint integration (added include
in Plater.cpp alongside the existing IMEXHelpers include), profile fixes
and version bumps (#14084, #14085, Polymaker), CI artifact publishing,
test refactors (arachne walls test added; test_3mf/test_config/
test_gcodewriter content moved/removed upstream — IMEX test file
preserved as it is feedback-only), translations.
The firmware-managed-zones writer offset centers the slice at bed origin,
so prim_pos and the preview toolpaths sit in a shifted frame relative to
the plate-local bed bounds the secondary-marker zone math uses. Computing
secondaries against unshifted bed bounds puts them off the build plate.
Skip the secondary-marker / toolhead-box computation when the
firmware-managed-zones option is enabled. The firmware physically fans the
centered toolpath out into the zones at print time, so the honest preview
is the single centered toolpath with no secondaries.
Reported by @Felix14-v2 on PR #13086.
* fix: restore BuildVolume bounds when toggling belt mode
set_belt_printer() mutated m_bboxf when enabling but never restored
the original extents on disable or when switching infinite_y true->false,
leaving stale max.y/max.z values that broke collision and object_state
checks. Recompute m_bboxf from m_bed_shape + m_max_print_height at the
top of each call, then apply belt-specific adjustments on top.
Addresses Copilot review comment on PR #12998 (BuildVolume.cpp:196).
* chore: drop [BELT-DEBUG] to_machine_coords log to trace
Was emitting at warning level once per 0.2mm Z bucket during every belt
print export, polluting default user logs. Trace level matches the rest
of the belt diagnostics and is silent in production.
Addresses Copilot review comment on PR #12998 (BeltGCodeWriter.cpp:86).
* chore: drop [BELTRACE] make_perimeters/support logs to trace
Eight warning-level traces around make_perimeters and
generate_support_material were emitting on every call/exit during normal
slicing, cluttering default logs. They're concurrency-debug breadcrumbs
not user-facing diagnostics, so drop them to trace.
Addresses Copilot review comment on PR #12998 (PrintObject.cpp:438).
* perf: gate BeltSliceStrategy diagnostic bbox tracking behind compile flag
apply_to_trafo() walked every model vertex twice (once for min_z, once
for per-volume mesh/slicer bboxes) and emitted seven trace logs per
call. The bboxes and logs are diagnostic only; min_z is the load-bearing
output. Wrap the bbox accumulation, logging, and supporting headers in
SLIC3R_BELT_DIAGNOSTIC_LOG so production builds do the bare min_z scan.
Addresses Copilot review comment on PR #12998 (BeltSliceStrategy.cpp:95).
* fix: apply part_cooling_fan_min_pwm to first-layer plane fan crossings
apply_first_layer_plane_fan_eval emitted band-crossing M106 commands
through GCodeWriter::set_fan() without the per-printer PWM floor that
every other set_fan call in CoolingBuffer applies. On printers with a
non-zero part_cooling_fan_min_pwm, fans could fail to spin up at low
requested speeds near the belt surface.
Addresses Copilot review comment on PR #12998 (CoolingBuffer.cpp:1227).
Secondary carriage markers now read the tool layout enum correctly,
aggregate to one active marker per gantry in Span modes, and mirror
across the bed centerline so they render on-bed.
Brings the standalone PrintApply dedup-fix commit and the prior upstream
merge into staging. Their content is already present here (the dedup fix
was re-applied during the latest upstream sync), so this only reconciles
the commit ancestry — no tree changes expected.
- Layout enum read via opt<ConfigOptionEnum<ImexToolLayout>>() did a
dynamic_cast that fails for the ConfigOptionEnumGeneric type enums load
as from presets, silently defaulting to FrontLeft. This desynced flip_y
from PartPlate::calc_imex_zones, computing secondaries against the wrong
primary zone and rendering them off the bed. Use option<>() (type-checked
static_cast), matching how PartPlate reads the same key.
- Span modes now collapse each non-primary gantry to one active marker via
group_imex_active_tools_by_gantry() (the same aggregation PartPlate uses
for zones), since only one tool prints per zone at a time. Previously every
Copy/Mirror tool got its own marker.
- Aggregated mirror markers reflect across the bed centerline rather than a
column edge; the aggregated strip spans full-X, so edge reflection would
push the marker off the bed.
Includes nozzle-diameter guards for printers without nozzle info (#13255,
now landed upstream) and manual-calibration nozzle mismatch fix (#13882).
# Conflicts:
# src/slic3r/GUI/GCodeViewer.cpp
# src/slic3r/GUI/GLCanvas3D.hpp
# src/slic3r/GUI/Tab.cpp
* initial commit
* fix upper bounds for assemblies
* significantly less Z shift issues, still not quite tamped down yet though
* add instrumentation to logs
* finally found the issue
* update printer defaults
Upstream #13360 moved `m_ori_full_print_config = ...` and
`update_values_to_printer_extruders_for_multiple_filaments(...)` into a
new `if (!extruder_applied)` block. The 3-way merge with our IMEX
`physical_extruder_map` derivation block (sitting between the old and
new positions) silently kept both copies, running the per-filament
reshape twice and miscounting filaments for the multi-temperature
compatibility check. Removed the obsolete second copy at the old
position. Resolves `p1s_multicolor.3mf` regression test failure.
Brings in upstream sync + AFC extruder_index integration:
- physical_extruder_map auto-population from AFC lane data
- verified end-to-end against xplorer (7 lanes, pem [0,0,0,0,1,2,3])
* initial commit
* fix upper bounds for assemblies
* significantly less Z shift issues, still not quite tamped down yet though
* add instrumentation to logs
* finally found the issue
* update printer defaults
* clean up UI elements
* further cleaning
* final cleanup for first round of settings UI streamlining
* update generic belt printer settings
* fix generic again
Reconciles the belt-printer branch with upstream PRs through #13723. Six
files had conflicts; three additional files needed manual follow-up fixes
where the auto-merge produced code that referenced upstream-renamed fields
or changed function signatures.
Notable reconciliations:
- TreeSupport.cpp: kept belt-floor early-exit branches around HEAD's
drop-down logic, folded upstream's `(distance_to_top > 0 ? 1 : 0)`
formula into the non-belt-floor path (upstream PR #11812). Dropped dead
`roof_enabled`/`force_tip_to_roof` locals.
- TreeSupport3D.cpp: combined upstream's safety-offset + remove_small
changes with HEAD's belt-floor clip in the per-slice trim loop. Dropped
HEAD's `else` block (superseded by upstream's rewritten bottom-contact
propagation) and re-added the belt-floor clip into the new propagation
loop. Gated the propagation on belt printers to prevent OOM when
belt-floor clipping produces empty initial slices.
- TriangleSelector.{cpp,hpp}: merged both new `select_patch` parameters
(HEAD's `up_direction` and upstream's `select_partially`); body uses
`dot(up_direction)` for the overhang angle check and forwards
`select_partially` to `select_triangle`.
- SupportMaterial.cpp: `slicing_params.soluble_interface` →
`zero_gap_interface_bottom` in HEAD's `detect_belt_floor_bottom_contacts`,
matching upstream's same-purpose rename at line 2495.
- Custom.json, GCodeWriter.cpp: simple additive merges (kept entries /
includes from both sides).
Verified by building OrcaSlicer (RelWithDebInfo) after a full deps
rebuild (Eigen v5.0.1, libigl v2.6.0 are now managed deps) and slicing
a scaled Benchy on the NORMALIZER belt-printer profile without OOM.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* minor logic swap
* first attempt, has a race condition
* fixed the offset issue
* found a solution, I think things work now (at least once I quash this race condition)
* still chasing down race conditions
* add manual shear / scale order strategy swap
* tweak manual shear, fix ui uninitialization crash
* fix z height / g-code desync issue
* fix shear then scale cutoff planes
* getting closer
* fix support termination planes
* fix incorrect offsets in shear-then-scale mode
* test - fix overextrusion due to model/layer scale
For single-gantry IMEX setups, front/rear is meaningless — only left/right
matters. Collapse the imex_tool_layout dropdown to two items ("Left" / "Right",
mapped to front-left/front-right internally) when imex_gantry_count == 1, and
show all four corner items otherwise. Normalize stored rear-* selections to
their front-* equivalents on the transition so the displayed selection always
matches the persisted value.
Also rename the field label "Tool 0 Corner" → "Tool 0 Position" so it reads
correctly in both single- and dual-gantry contexts, with a tooltip that
explains the IDEX/IQEX distinction.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds the `imex_firmware_managed_zones` printer-config key (default off) for
IDEX/IQEX printers whose firmware applies its own copy/mirror offsets in
non-primary modes (e.g. RepRapFirmware IDEX duplication mode, Flashforge
Creator Pro 2/3 Pro). For these printers the slicer needs to emit a single
centered slice at bed origin and let the firmware fan toolheads out from there;
the previous slicer-managed iMEX rendering would draw a print at the primary
zone's world position (off-bed for the firmware-fan-out paradigm).
When the flag is on and the active mode is non-primary, the slicer subtracts
the primary zone's plate-local center from the gcode emission frame. The
writer offset is augmented but the gcode-processor offset stays at plate_origin
so the gcode-preview visualizer renders the centered slice at the bed center
rather than at the prepare-view zone placement. translate_to_print_space is
augmented too so first_layer_print_min/max placeholders (consumed by user
start_gcode like Felix's M118 header) reflect the centered frame.
Slice handoff lives in PartPlate::refresh_imex_slice_offset, called from both
update_slice_context (plate switch) and Plater::priv::update_background_process
(every-slice path — reslice() goes through here with switch_print=false so the
plate-switch hook alone wouldn't fire on mode toggle).
calc_imex_ghosts early-returns in firmware-managed mode: the existing
imex_head_transform math places ghosts at primary_zone_center + gantry_offset
(slicer-managed semantics), which renders off-bed when the toolpath is being
emitted in a centered frame. Proper firmware-managed ghost rendering (showing
where copies/mirrors will actually print after firmware fan-out) is deferred.
When the flag is off, all the new code paths reduce to no-ops byte-identical to
prior behavior. Layer 1 unit tests in test_imex_helpers cover every gating path
of compute_imex_slice_offset; full ctest suite passes (247/247).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Conflicts resolved in src/libslic3r/GCode.cpp and src/slic3r/GUI/GUI_Factories.cpp.
GCode.cpp: combined upstream's air-filtration per-extruder gating
(activate_air_filtration_during_print / _on_completion), the new
extrusion-role-change gcode lambda, ZAA's path.z_contoured arc-fit
disable, raft-aware slow_down_layers branch, and Vec3d/Line3 ZAA
plumbing with the local belt-printer changes (path_on_first_layer,
effective_layer_index_for_point, should_disable_arc_fitting). All
auto-merged m_writer.X() calls converted to m_writer->X() to match
the local unique_ptr<GCodeWriter> refactor.
GUI_Factories.cpp: inserted brim_flow_ratio in the Support category
list and renumbered around the local build_plate_tilt_x/y entries.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
`check_outside` (placement check) calls `ensure_imex_zones()` to make sure
zone geometry is current before deciding if an instance is in-bounds. In
the GUI path that's fine, but the CLI / headless 3MF-load path also reaches
this through `PartPlateList::load_from_3mf_structure -> reload_all_objects
-> add_instance -> check_outside`, and CLI mode has no GUI_App initialized.
`build_imex_cache_key` and `calc_imex_zones` both dereference
`wxGetApp().preset_bundle` — without a GUI_App, `wxGetApp()` returns memory
that segfaults on member access, killing the slicer with SIGSEGV before
any G-code is produced.
Latent since 461c69c83e (Apr 9), surfaced now that upstream's main carries
the headless regression-test CI step (#13353) that exercises CLI slicing
on every PR build.
Fix: short-circuit `ensure_imex_zones()` when `m_plater` is null (already
the GUI/CLI marker used by `calc_imex_ghosts`). Also tighten the existing
`build_imex_cache_key` null check to consult `m_plater` first as defence
in depth, so the function stays safe if reached from another headless
caller.
Verified locally against the upstream regression suite — klipper /
p1s_multicolor / toolchanger_4_color all slice cleanly within the
20% baseline tolerance.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The aggregated-mirror ghost was X-flipping about the mesh's local origin,
which shifted the ghost sideways for models whose local origin sits at a
corner (calibration cubes, calicat, most STL imports anchored at the
min corner). Visible as a constant left-X offset between the primary's
position and the ghost's position.
Pivot on `mo->raw_mesh_bounding_box().center()` instead, applied through
the instance transform so rotated objects flip about the rotated bbox
center too. Same correction applied to the live-drag update path.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Introduce ImexRole::Span as a 5th tile cycle state that declares "this
tool is the multicolor partner of Primary on the same gantry." Encoded
as the `S` role suffix in `imex_mode_active_tools` (e.g. `0:P,1:S,2:M,3:M`).
Disambiguates paired-gantry mc-mirror from 4-independent-copies — both
share the same active_tools shape sans the marker.
Span drives:
- Multicolor block rule: now requires Span on primary's gantry to allow
multi-color slicing in a parallel mode. Replaces the prior "≥2 tools
on primary's gantry" check; pre-existing 4-tool multicolor configs
need T1 flipped to Span.
- Ghost aggregation: one ghost per non-primary gantry when Span is
present, using the column-paired representative. Aggregated-mirror
drag tracks primary 1:1 in X (gantries don't share an X rail) with
X-flip baked into mesh-local frame so geometry still reads as mirrored.
- Zone aggregation: one full-X row strip per non-primary gantry instead
of per-tool quadrants.
- UI: 5th button in IMEXModesCtrl. Cycle Off→P→C→M→S→Off, only offered
on multi-gantry printers and only on tiles sharing primary's gantry row.
Single source of pairing truth: group_imex_active_tools_by_gantry in
IMEXHelpers, consumed by ghost factory and zone calculator.
Also fixes the carriage collision strip's X-boundary check, which lacked
the row constraint its Y-boundary counterpart already had — paired-gantry
mc-mirror was drawing a spurious right-edge strip from T3 sitting
diagonally from primary.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
_render_imex_ghosts and _picking_pass_imex_ghosts both walked every
plate in the partplate list and rendered/hit-tested all ghosts found —
which meant background plates' ghosts bled through into the active
scene whenever GL state was shared (most visibly when entering paint
mode), and clicks could land on a ghost that belonged to a plate the
user wasn't actually looking at.
Switch both paths to read get_curr_plate() and skip the per-plate loop.
Per-plate ghost volumes still live on each PartPlate so 3MF
round-trips work and switching plates picks up the new active plate's
ghosts cleanly; we just don't draw or hit-test the ones whose plate
isn't the active one.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Catches the case where a non-primary IMEX mode's active tools all sit on
one gantry — e.g. mode "0:P,1:C" on a 2x2 IQEX where T0 and T1 share
gantry 0. The "Copy"/"Mirror" label is decorative there: nothing actually
parallel-prints, but the user's mode_gcode still fires and emits firmware
setup that doesn't apply, while the slicer treats it as a multi-color
parallel print. Conceptually it's just a regular multi-tool single-gantry
print and belongs in Primary mode.
Changes:
- imex_multicolor_block_reason now collects the set of distinct gantries
spanned by the active tools and short-circuits with a clear "single
gantry — not a parallel-print scenario" message before falling through
to the existing within-gantry-swap check.
- Drops the redundant pre-slice "Multi-material objects detected" soft
warning from collect_imex_warnings — the slice-time block surfaces a
more specific message at the right moment, and the soft warning was
vague handwaving in front of it. Bed-temp + filament-type checks stay.
- New unit test covering the single-gantry block.
Behavior matrix on a 2x2 IQEX with multi-color:
"0:P,1:C" single gantry -> BLOCK (new)
"0:P,2:C" dual gantry, 1 each -> BLOCK (existing within-gantry-swap)
"0:P,1:C,2:M,3:M" dual gantry, 2 each -> ALLOW
multiple filaments to same physical via pem -> BLOCK (existing MMU sharing)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The mode-type tag (imex_mode_types config + imex_mode_type_for helper +
Split sentinel) was added in 0bb1cef as scaffolding for the Split rendering
work that landed in 4370cca and then got reverted in d9be71b. With Phase 2-4
gone, this scaffolding is now unused dead code — and the design we settled
on instead is to leave topology entirely implicit (parsed from active_tools_str)
rather than carrying a per-mode type tag the user would otherwise have to
manage explicitly.
The multi-color slicing block + bare T<n> suppression that were the actual
substance of the safeguards work stay in place:
- imex_multicolor_block_reason still allows multi-color exactly when 2+ tools
are active on the primary's gantry — Felix's hypothetical IQEX paired-gantry
case works through this path, no new mode type required.
- Slicer-side: bare T<n> stays suppressed at print-start in IMEX parallel modes;
mid-print T<n> emits naturally for the legitimate IQEX 4-tool-active scenario.
Removed:
- ConfigOptionStrings imex_mode_types (PrintConfig.hpp/cpp + Preset.cpp key list)
- imex_mode_type_for helper + kImexModeType{Primary,Copy,Mirror,Split} sentinels
- mode_type parameter on imex_multicolor_block_reason and the Split short-circuit
- mode_type plumbing in Print::validate and PartPlate::has_imex_multimaterial_conflict
- Three unit tests for imex_mode_type_for + two Split-specific multicolor block tests
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Completes the Split-type IMEX mode plumbing started in 0bb1cef. Phase 1
landed the config option, helper, validator, and plater badge updates;
this commit lands the visualization side and the editor surface so users
can actually create and use Split modes.
Zone aggregation (PartPlate::calc_imex_zones)
- When mode_type == "split" and there's row separation between primary
and secondaries, collapse copy_cells and mirror_cells per-gantry
(each cell gets primary's column) and force has_col_sep = false.
make_boxes then takes its full-X-row branch and renders one zone
covering each non-primary gantry's full Y band, instead of per-tool
quadrants.
Ghost aggregation (PartPlate::calc_imex_ghosts)
- Pre-scan the active mode's tool list to pick a canonical head per
non-primary gantry (the tool whose physical column matches primary's;
fallback to first-seen). For Split modes, the per-head emission loop
skips non-canonical tools so each non-primary gantry gets exactly one
aggregate ghost rendered at the canonical's mirrored position.
- update_imex_ghost_transforms unaffected — it iterates the already-
filtered ghost set.
Modes editor Type column (IMEXModesCtrl in Tab.cpp)
- Each non-primary mode row gains a wxChoice dropdown selecting Copy /
Mirror / Split. Primary row gets a static "Primary" label.
- Header row picks up a "Type" column header with sized spacer that
aligns with the dropdown.
- get_mode_data() now returns a fourth tuple element (types vector).
load_from_config() reads imex_mode_types via imex_mode_type_for so
legacy presets without the new option still infer types from mode
names. on_change writes imex_mode_types back to config.
- matches_config(), snapshot_row(), row_differs_from_parent(), and
reset_row_to_parent() all extended to track type alongside name/
tools/gcode — per-row reset arrows reflect type-only changes, and
reset restores the parent's type via imex_mode_type_for fallback.
- Type-choice dropdown fires notify() on change so the dirty/save flow
catches it like any other row edit.
End-to-end: switching a row to Split type in the modes editor, then
selecting that mode on a plate, allows multicolor slicing (the gantry-
pair check is bypassed in imex_multicolor_block_reason) and the bed
visualization shows a single aggregate zone + one ghost per non-primary
gantry instead of per-tool clutter.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Felix14-v2 PR review feedback (https://github.com/OrcaSlicer/OrcaSlicer/pull/13086#issuecomment-4323696312)
plus the slice-time validation work that follows from his bug list.
UI polish:
- Capitalize "Primary" in display (right-click mode menu, plate tooltip).
Sentinel stays lowercase for wire compatibility.
- Pre-slice-warnings checkbox now uses Orca's ::CheckBox so it matches the
green toggle style of the rest of the IDEX/IQEX configuration page.
- DPI-scaled the IMEXModesCtrl (modes editor) sub-panels, button grid,
text wraps, gcode textarea, and the ghost-tooltip swatch (imgui.scaled).
Legend swatches sized to body-text height for visual balance.
- Primary mode tool buttons in the modes editor are now disabled (read-
only): cycling roles on the IMEX-off mode is a no-op and confusing.
- Modes editor sub-panels now explicitly inherit the app's window-default
dark colour so chromeless ScalableButtons don't render with a visible
light box around their icons on GTK dark themes.
- Per-mode-line reset arrows in the modes editor: each row gets a small
reset bitmap that snaps that row's name+tools+gcode triplet back to the
saved preset's value. matches_config() guard on the page-level reload
prevents the textbox-being-typed-into from being destroyed mid-keystroke.
- New View menu item "Show IDEX/IQEX Toolhead" — toggles the per-carriage
footprint boxes during G-code preview playback. Gated to Preview tab +
IMEX printer; backed by app_config so it persists.
Coordinated config migrations:
- imex_tool_layout and imex_viz_theme migrate from coString to coEnum
(ImexToolLayout / ImexVizTheme). Existing wire format preserved so
saved presets and 3MFs deserialize unchanged. Side-benefit: both now
pick up standard Field rendering and so finally show reset arrows.
Slice-time safeguards (the validation half):
- imex_suppresses_bare_toolchange(parallel_mode, count): suppresses the
slicer's bare T<n> at print-start in any IMEX parallel mode (the user's
imex_mode_gcode + machine_start_gcode owns tool activation there). Mid-
print T<n> emits normally — Print::validate blocks the configurations
where mid-print T<n> wouldn't make sense. Applied to both code paths
inside GCode::set_extruder (the long multi-extruder path AND the
single-extruder path that fires when multiple_extruders=false).
- imex_multicolor_block_reason(): hard-stop validator returning a user-
facing reason string when the active IMEX configuration can't physically
support multi-color. Catches IDEX (1 tool/gantry), 2-tool-active IQEX
(no within-gantry swap topology), and any MMU/AFC lane sharing among
used filaments. Wired into Print::validate as a slice blocker, and into
PartPlate::has_imex_multimaterial_conflict so the plater badge agrees
with the slice block (no more false positives where the badge warns but
the slice goes through).
- New imex_mode_types config option (parallel array to imex_mode_names)
and imex_mode_type_for() helper. Mode-type tag drives behaviour: zones,
ghosts, and validation interpret modes differently per type. Initial
types: "primary", "copy", "mirror", "split". Split modes are explicitly
designed for paired-gantry IQEX multi-color and bypass the gantry-pair
check in imex_multicolor_block_reason (MMU sharing still blocks them).
Test coverage:
- New unit tests cover imex_suppresses_bare_toolchange (4 cases),
imex_multicolor_block_reason (8 cases including IDEX, IQEX 2-/4-tool-
active, MMU sharing, missing primary, Split type), and imex_mode_type_for
(3 cases including legacy fallback). Total: 186 IMEX assertions across
71 test cases, all passing.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
AFC now publishes an extruder_index field per lane. This lets Moonraker
auto-derive physical_extruder_map without user configuration:
- Add extruder_index field to AmsTrayData (-1 = not provided)
- Parse extruder_index from AFC lane JSON using contains() check since
0 is a valid extruder index (safe_json_int returns 0 for missing fields,
which would be ambiguous)
- After build_ams_payload, if all trays have extruder_index, derive
physical_extruder_map and write it to the printer preset config via
CallAfter so IMEX PA and temperature emission use the correct physical
extruder qualifier
AFC lanes sharing a carriage all report extruder_index=0; independent
direct-drive tools on separate carriages report their carriage index.
PrintApply.cpp's auto-derive from printer_extruder_id is skipped when
physical_extruder_map has more than 1 element, so the AFC-populated map
takes precedence.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Three of the four IMEX physical-vs-logical fixes landed earlier on this
branch (fbc58d2a1d, fa048babeb, a38b95bf45, e11e7d46df) used inline
lambdas / direct loops to translate physical extruder indices to logical
filament slots. No test coverage existed for the specific composition,
even though the underlying primitives (resolve_filament_for_head,
first_filament_for_physical_head) were tested.
Pull two patterns out of Plater.cpp and GCode.cpp into IMEXHelpers as
named helpers, then unit-test them:
imex_primary_logical_from_objects(used_slots_1b, pem, primary_physical)
Walks the plate's used filament slots (1-based) and returns the
first one whose pem entry maps to the primary's physical extruder.
This is what the warning's `logical_for_primary` now delegates to —
moves the "look at object assignments, not pem first-routed default"
behavior introduced in fa048babeb out of the lambda and into a
separately-testable function.
imex_secondary_logical_slots(active_physicals, primary_physical,
plate_head_filament_map, pem)
Iterates IMEX active physicals, skips the one matching primary,
resolves each remainder via resolve_filament_for_head (per-plate
override + first-routed fallback), deduplicates, drops -1 entries.
Replaces the inline loop in GCode::_do_export's is_extruder_used
marking (a38b95bf45 + e11e7d46df).
10 new test cases in test_imex_helpers.cpp cover the cases that
correspond directly to bugs hit:
imex_primary_logical_from_objects:
- AFC primary picks the object's slot (the user's specific bug)
- Multi-color AFC primary returns first input-order match
- Direct extruder primary unambiguous (no MMU)
- No object routed to primary returns -1
- Empty inputs (no objects, empty pem)
imex_secondary_logical_slots:
- Copy mode skips primary, falls back to first-routed
- IQEX 4-mode enumerates all three secondaries
- Per-plate override wins over first-routed
- Drops unrouted physicals + deduplicates
- Only-primary-active returns empty
All [IMEX] + [Variant] + [3mf] regression: 166 assertions / 61 cases.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
a38b95bf45 translated physical -> logical for IMEX active tools but used
resolve_filament_for_head for *all* active tools including the primary.
For the primary, that falls back to first_filament_for_physical_head —
which returns the FIRST logical slot routed to the primary's physical
extruder, not the slot the user actually assigned to the printing object.
On the user's Neo XP 0.6 (pem [0,0,0,0,1,2,3,3]) printing in copy mode
with the object on slot 2 (PLA dark grey):
- tool_ordering correctly marks slot 2 (object's filament).
- The IMEX-marking loop then *also* marked slot 0 (ABS) as the primary's
"first-routed" slot — wrong: slot 0 isn't loaded, slot 2 is.
- Start-gcode template emitted EXTRUDER=260 EXTRUDER2=235 EXTRUDER4=235;
the ABS-temp emission for slot 0 was harmless noise in this macro
design but conceptually bogus.
The primary's filament is already correctly covered by
tool_ordering.all_extruders() — that lists the slots the objects on the
plate are assigned to. Skip the primary in the IMEX-marking loop using
the same pattern as the IMEX PA emission path at GCode.cpp:3265
(translate initial_extruder_id -> physical via pem, skip that physical).
Result on the user's setup post-fix:
- is_extruder_used: slot 2 (object), slot 4 (T1's filament in copy).
- Start-gcode emits EXTRUDER2=235 EXTRUDER4=235 — exactly two temps,
one per active heater, with no spurious ABS bookkeeping.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
is_extruder_used is a logical-filament-slot indexed bool array — start-gcode
templates use it as `is_extruder_used[N]` where N is a logical filament slot
(matches how the rest of the codebase consumes per-filament arrays like
filament_settings_id, nozzle_temperature_initial_layer, etc.).
`tool_ordering.all_extruders()` returns LOGICAL slots and was correctly
marking those. But the IMEX-secondary marking added in 5aa624b025 was
writing PHYSICAL extruder indices straight into the logical array, mixing
index spaces. On any printer with physical_extruder_map size > 1 (MMU/AFC),
this marks the wrong slots and misses the right ones.
Symptom on the user's Neo XP 0.6 (pem [0,0,0,0,1,2,3,3]) in copy mode
[0:P,1:C] with object on slot 2 (PLA):
- tool_ordering marks slot 2 (correct: object's filament).
- IMEX active = [0, 1] (physical T0, T1) → mistakenly marks logical
slots 0 (ABS) and 1 (ASA), neither of which is used.
- Slot 4 (PLA on physical T1, the actual filament that loads on the
secondary in copy mode) is NOT marked.
- Start-gcode template emits EXTRUDER=ABS_temp EXTRUDER1=ASA_temp
EXTRUDER2=PLA_temp; no EXTRUDER4.
- PRINT_START macro reads t4=0, skips heating extruder1 — T1 stays
cold during the print.
Translate physical → logical via resolve_filament_for_head before marking
(per-plate imex_head_filament_map override consulted, with first-routed
fallback when no override is set). This matches what the firmware actually
loads on each carriage during the parallel-mode print, and what the rest
of the IMEX hot path (PA emission, layer-change temperature) already does.
Result on the user's setup post-fix:
- is_extruder_used[2]=true (object), is_extruder_used[4]=true (T1 in copy).
- Start-gcode emits 2 temps for the actually-used filaments.
- extruder1 heats correctly to slot 4's temp before the print begins.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The pre-slice warning's primary-tool filament lookup (added in
fbc58d2a1d) used resolve_filament_for_head, which returns the FIRST
slot routed to the primary's physical extruder via pem. That's the
right rule for *secondaries* (no object owns them in copy/mirror mode;
filament comes from the per-plate imex_head_filament_map override) but
wrong for the *primary*: the primary prints the actual objects on the
plate, and the slot it uses is whatever the user assigned to those
objects — not whatever happens to be at the head of the AFC manifold.
Symptom on the user's Neo XP 0.6:
- pem = [0,0,0,0,1,2,3,3] (slots 0-3 share AFC manifold on physical T0)
- filaments: slot 0 ABS, slot 2 PLA, slots 4-6 PLA, etc.
- object assigned to slot 2 (PLA), IMEX mode "copy" (T0 primary, T1 copy)
- User picks slot 5 PLA for T1 via the IMEX ghost picker (writes a
per-plate imex_head_filament_map).
- Warning reads slot 0 (ABS) for primary because that's
first_filament_for_physical_head(pem, 0). Fires "T0 ABS vs T1 PLA
type mismatch" even though the actual print uses slot 2 (PLA) for
the primary — slicer and warning disagree.
Split the lookup:
logical_for_primary(physical_idx)
Walks plate->get_extruders(true) (1-based slots used by objects on
this plate), returns the first slot whose pem entry maps to
physical_idx. Falls back to first_filament_for_physical_head if no
object on the plate routes to this physical extruder (defensive).
logical_for_secondary(physical_idx)
Unchanged behavior: per-plate imex_head_filament_map override with
first_filament_for_physical_head fallback.
The user-facing "T%d" labels still display the physical extruder index
(carriage identity); only the filament-info lookups change.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
collect_imex_warnings parses imex_mode_active_tools to get the active tool
indices. Those are PHYSICAL extruder indices (one per carriage). It then
used those same indices to look up filament_presets[tool_idx] and
bed_temps[tool_idx] — but both arrays are indexed by LOGICAL filament slot.
For MMU/AFC layouts where multiple logical slots feed one physical
extruder (e.g. AFC manifold: 4 lanes on physical T0), the warning would
report the wrong filament: a secondary on physical T1 would be named with
filament_presets[1] (= AFC lane 2) instead of the actual filament on T1.
Symptom: in IMEX parallel mode on the user's IQEX-AFC printer (pem
[0,0,0,0,1,2,3]), the multi-extruder warning called the secondary tool
"T1" but reported the filament type for logical slot 1 (an AFC lane),
not the actual filament 4 routed to physical T1.
Translate physical → logical via effective_physical_extruder_map (with
per-plate imex_head_filament_map override) before indexing into
filament_presets and bed_temps. The displayed "T%d" still shows the
PHYSICAL extruder number — that's the carriage identity the user sees
on hardware. Only the filament-info lookup is changed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Three IMEX bugs of the same shape have surfaced over the lifetime of
this feature:
- Inline pem lookup duplicated at two PA emission sites
- Pre-slice warnings indexing filament_presets by physical index
- Ghost color resolution with stale default pem handling
Each was a place where a per-filament array got indexed by what the call
site had on hand (a physical T-number) without translating through
physical_extruder_map. On non-MMU/non-AFC printers the indices coincide
and nothing breaks; on AFC layouts the slicer reads the wrong filament
preset for a carriage with no error or log line.
Add a header comment block to IMEXHelpers.hpp describing the two index
spaces, when each is used, how to translate, and a list of the bugs we
hit so future contributors can recognize the pattern.
The constant kImexPrimaryMode and the helper declarations follow this
block; readers searching for pem helpers will land on the guidance first.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Brings in the IMEX test coverage (PA per-firmware, Temperature per-firmware,
[Variant] expansion, 3MF round-trip, imex_pem_tool_for helper + tests, and
the cherry-picked variant-truncation regression test).
Resolution notes:
- Two GCode.cpp call sites for set_pressure_advance had divergent edits:
* tests/imex-coverage rewrote them to use the new imex_pem_tool_for
helper (commit c2492ccc47), eliminating the inline parallel-mode
check entirely.
* feedback replaced the literal "primary" with kImexPrimaryMode in the
same lines (commit 085f5ccec8).
Resolution: keep the helper-call form. The kImex change is moot on lines
the helper replaces, and imex_pem_tool_for in IMEXHelpers.cpp is also
updated to use kImexPrimaryMode for consistency with the rest of the
codebase.
- Test test_3mf.cpp updated for upstream's load_bbs_3mf signature change
(PR adds is_orca_3mf out-parameter between is_bbl_3mf and file_version).
All three call sites in the new IMEX 3MF round-trip tests pass &is_orca
in addition to &is_bbl.
Full regression post-merge:
libslic3r: 143 cases / 48,553 assertions (+10 cases from new tests)
fff_print: 24 cases / 245 assertions (+10 cases from new tests)
sla_print: 21 cases / 14,100 assertions
libnest2d: 14 cases / 488 assertions
slic3rutils: 3 cases / 3 assertions
All tests pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Follow-up to 085f5ccec8. Self-review of that commit's grep output missed
GCodeViewer.cpp. Three sites in the layer-preview multi-carriage marker
logic still compared against the bare "primary" literal:
- GCodeViewer.cpp:1538 — process-preset mode default
- GCodeViewer.cpp:1542 — per-plate mode override gate
- GCodeViewer.cpp:1551 — secondary marker computation gate
All three now use kImexPrimaryMode. The file already includes
IMEXHelpers.hpp (line 13) so no new include needed.
Verified by grepping the full IMEX-touching set: only IMEXHelpers.hpp
itself (the constant definition) still references the literal string,
which is correct.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The IMEX zone cache key drives ghost rebuild and zone-overlay
invalidation. Adding a printer config option that affects zone geometry,
ghost transforms, or collision strips without extending the key produces
a silent staleness bug: the cache thinks the zones are still valid and
ghost meshes / overlays don't refresh after the new option changes.
Document what currently feeds the key and pin the precision convention
(*10 scale on float values for 0.1 mm resolution) so future contributors
know the contract and where to extend it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
GCode.cpp's static get_imex_active_tools() inlined its own "phys[:role]"
tokenizer with subtly different semantics from IMEXHelpers'
parse_imex_active_tools — only the GCode version bounded against
MAXIMUM_EXTRUDER_NUMBER. Three other call sites (PartPlate zones,
GCodeViewer legend, Plater warnings) routed through parse_imex_active_tools
already.
Move the bounds check into parse_imex_active_tools so all consumers get
it consistently, then rewrite get_imex_active_tools to do only the
Print-extraction portion (active mode lookup, tools-string fetch) and
delegate token parsing to the helper. Keeps Print out of IMEXHelpers'
include set.
No behavior change for the non-pathological case (mode strings have always
parsed identically); for indices >= MAXIMUM_EXTRUDER_NUMBER (64) the three
older call sites silently filter them out now where previously they would
have accepted them — this matches what get_imex_active_tools already did.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Across 6 files the literal "primary" string was the sentinel for "no IMEX
parallel mode active" — used for short-circuiting in serialization, ghost
visualization, zone calc, popup-menu list construction, the IMEXModesCtrl
non-deletable first row, and several layer-of-export checks. A typo in
any one would silently bypass the guard.
Define kImexPrimaryMode in IMEXHelpers.hpp with a docstring describing
what equality with it means semantically, and route every call site
through it. No behavior change.
Touched: bbs_3mf.cpp, GCode.cpp, PartPlate.cpp, Plater.cpp, Tab.cpp.
The bbs_3mf and Tab files now include IMEXHelpers.hpp; the other three
already did.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Four quick-fix items surfaced by pre-PR self-review.
GCodeViewer.cpp:1607
Null-check get_curr_plate() before dereferencing. Other call sites in
the file already guard; this was the only unguarded one in the IMEX
layer-preview path. In practice m_plate_list always has a plate, but
the inconsistency is easy to fix and removes the only ungated deref.
PartPlate.cpp:build_imex_cache_key
Cache key for IMEX zone geometry truncated nozzle_clearance_x/y to int
before stringifying — a config change from 30.0 to 30.5 would not
invalidate the cache. Match the *10 precision pattern already used for
imex_carriage_margin so 0.1 mm steps invalidate correctly.
Plater.cpp:select_plate_by_hover_id (right-click popup)
Two issues:
1. Lambda captured `modes` by reference. PopupMenu() is synchronous
today so the reference outlived the menu's event handling, but the
pattern is fragile — anyone refactoring to async Popup() would
silently dangle. Capture by value.
2. Used wxID_HIGHEST + i for menu item IDs — standard wx anti-pattern
because it can collide with other handlers listening in that range.
Allocate per-item IDs via wxNewId() and look up the chosen mode by
finding the event ID in a parallel vector. The lookup becomes O(N)
instead of O(1) but N is small (mode count) and this is clicker
latency, not a hot path.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
af59501f4a ("feat: firmware-agnostic per-tool PA emission for IMEX
parallel modes") inadvertently deleted the BBL-specific PA emission for
initial_non_support_extruder_id while adding IMEX per-tool PA support.
That deletion was scope creep into core BBL functionality and not part
of the IMEX feature.
Restore the original block verbatim. The new IMEX-parallel-modes PA
emission (per-secondary-tool, gated on m_imex_parallel_mode) is left
untouched — that's legitimately IMEX scope. BBL printers in non-IMEX
mode now get back the pre-PR initial-PA behavior.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Self-review found two weaknesses in the preceding test commits:
1) The equal-size [Variant] scenario claimed to distinguish the truncation
guard's `cur > target` predicate from a regression to `cur >= target`,
but both paths yield identical child values in practice: when
extruder_variant names match, set_with_restore's variant_index is fully
populated (no -1 slots) and the merge path restores every position from
backup — producing the same {1.5, 2.5} output as the skip path. The
test passes in both guard states.
Rewritten to use mismatched variant names between child and parent.
variant_index then has -1 slots, and set_with_restore overwrites those
positions with parent values. Now the merge path yields {0.8, 0.8} and
the skip path yields {1.5, 2.5} — observably different. Verified:
- `cur > target` (correct): 4 scenarios pass, 15 assertions
- `cur >= target` (regressed): equal-size scenario fails with
"1.5 is within 0.000000001 of 0.80000000000000004"
- Guard removed entirely: child>parent + stride=2 both fail with
truncation ("1 == 2" / "2 == 4")
2) The [3mf][IMEX] round-trip only covered a single plate. A plate-
indexing regression (IMEX metadata landing on the wrong plate, or
bleeding across plates on reload) would not have been caught.
Added a multi-plate scenario: two plates with distinct mode and
head-filament-map values. Asserts both land on their respective
destination plates after reload. Load-bearing verified:
- With IMEX serialization intact: 3 scenarios pass, 45 assertions
- With IMEX serialization disabled: positive + multi-plate fail
(both "nullptr != nullptr"); primary-mode passes (expects nullptr)
- With primary-mode short-circuit removed: primary-mode scenario
fails ("0x... == nullptr") because primary modes now serialize
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Four scenarios cover the temperature emission surface that IMEX layer-change
handling routes through (Tier 1 of the deferred Target B test plan —
pure-function-only, no fixture).
- Per-flavor command routing: Marlin (M104), RRF (G10 — M104 is deprecated
on RRF), Mach3/Machinekit (P-prefix for value instead of S).
- Wait handling: Marlin emits M109, MakerWare/Sailfish silently drop wait
requests (the firmware doesn't support blocking waits), Teacup and RRF
both emit a separate M116 poll.
- Per-tool qualifier for IMEX secondary carriages: Marlin and Klipper
emit T<N>, RRF uses P<N> (same P override as its wait poll). This is
exactly the path that lets IMEX set secondary-tool layer temperatures
without a tool-change.
- Instance overload's multi-extruder gating: a tool index passed to a
single-extruder GCodeWriter is discarded (no spurious T0 on
single-tool printers), but a multiple_extruders writer passes it
through verbatim.
All 24 assertions in 4 cases pass under [Temperature].
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The physical_extruder_map translation used by IMEX per-tool PA emission was
inlined identically at two sites in GCode.cpp (tool-change and second-layer
transition). Extract to IMEXHelpers so the routing rule ("parallel mode AND
populated pem → physical index, else -1") is testable in isolation and the
call sites read as intent rather than re-deriving the conditional.
Production change is behavior-preserving:
- Same predicate (`!mode.empty() && mode != "primary"`)
- Same empty-pem short-circuit returning -1
- Same get_at() dispatch on hit
- Both call sites replaced with a single call
Four unit tests in [IMEX] cover the routing matrix:
- non-IMEX ("") and primary mode short-circuit
- parallel mode + empty pem short-circuits (defense-in-depth; get_at would
throw on empty values otherwise)
- identity pem (non-MMU IDEX) routes filament to itself
- MMU collapse routes multiple logical slots to one physical (7-slot profile
with 4-lane MMU on physical 0 and direct drives on 1/2/3)
All IMEX + Variant regression suites pass post-refactor (133 assertions / 48
cases under libslic3r, 25 assertions / 6 cases under fff_print [PressureAdvance]).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Validates that imex_parallel_mode and imex_head_filament_map survive a
full store_bbs_3mf → load_bbs_3mf cycle — the same silent-state-loss bug
class that produced the variant-vector truncation regression, applied to
IMEX plate state which rides the same XML metadata path.
- Positive round-trip: a plate with copy_mode + a non-trivial head
filament map ("1:2,2:3") is saved and reloaded; both options land on
the destination plate's config with the exact values preserved.
- Guard scope: a plate with mode="primary" and empty head-filament-map
does NOT emit metadata (per the serializer's short-circuit), and the
reload leaves both options absent from the destination config. If the
serializer ever regressed to writing primary-mode plates, the load
path would surface phantom "primary" strings on plates that shipped
clean — this catches that.
Both scenarios call set_temporary_dir to point the BBS exporter's backup
scaffolding at a writable per-process temp directory (by default it
resolves under root at runtime, which fails for non-root test
processes).
All 27 assertions in 2 test cases pass under [3mf][IMEX].
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Exercises the IMEX per-tool PA emission surface added in af59501f4a
("feat: firmware-agnostic per-tool PA emission for IMEX parallel modes").
Six scenarios cover the full routing matrix:
- Negative PA returns empty across all flavors (early-exit guard).
- Klipper: bare vs EXTRUDER=extruder vs EXTRUDER=extruderN. Asserts the
tool=0 case emits the unsuffixed extruder name (first Klipper extruder
is named "extruder", not "extruder0") — a subtle edge case easy to
regress.
- RRF: bare vs D0 vs DN. The D0 case matters: passing tool=0 explicitly
must emit `D0`, not the current-tool fallback.
- Marlin 2.x: bare vs T0 vs TN.
- Marlin Legacy: tool index is silently dropped — verifies the fallback
branch can't accidentally start emitting T qualifiers on firmware that
doesn't support them.
- BBL: flag wins over firmware flavor (Marlin 2 flavor + BBL flag emits
the BBL-specific `M900 K... L1000 M10`) and BBL never emits a per-tool
qualifier regardless of the tool argument.
All 25 assertions across 6 cases pass under [PressureAdvance].
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds three scenarios alongside the existing child>parent stride=1 regression
test for update_non_diff_values_to_base_config:
- stride=2 child>parent: machine_max_acceleration_x (size 4 vs 2) — confirms
the truncation guard fires for the (normal,silent)-pair stride=2 path, not
just stride=1. Catches a regression class the existing test would miss
because stride=2 routes through normalize_stride2_floats and a different
set_with_restore call site.
- equal-size (2=2): exercises the path the guard does NOT short-circuit;
asserts child per-extruder values survive set_with_restore's nil-restore
merge. Catches any future change that breaks the equal-size merge — the
fix's `cur > target ? skip` predicate could regress to `cur >= target` and
silently override child values otherwise.
- non-variant scalar: layer_height in `keys` and `different_keys` but absent
from printer_options_with_variant_1/_2. Hits the is_scalar() / "nothing to
do" branch and must remain untouched. Scopes the guard's blast radius.
All four scenarios in the [Variant] tag pass: 15 assertions, 4 test cases.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Adds a Catch2 scenario that builds a 2-extruder child DynamicPrintConfig
inheriting from a 1-extruder parent, calls update_non_diff_values_to_base_config
through the real printer_options_with_variant_1 / _2 key sets, and asserts
that printer_extruder_id, printer_extruder_variant, and retraction_length
retain their full size after the merge. Covers three distinct
set_with_restore<T> instantiations (Ints, Strings, Floats) and verifies
both size preservation and per-extruder value preservation.
Verified load-bearing: with the guard in update_non_diff_values_to_base_config
temporarily removed, the test fails with "1 == 2" on pe_id.values.size() and
retraction_length.values.size(); with the guard restored, all six assertions
pass.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
update_non_diff_values_to_base_config sizes variant_index to the parent's
(inherits-target's) extruder count, and set_with_restore then replaces
the child's vector with a parent-sized one. When the child preset has
more extruders than the parent (e.g. an IDEX preset inheriting from a
single-nozzle base), every key in printer_options_with_variant_1 is
truncated to the parent's size on project reload, destroying per-extruder
data.
Observable symptoms: objects render with the wrong color (often black)
after reopening the project, and the printer preset shows a permanent
dirty-asterisk that no save/reload cycle can clear.
The child's saved value is authoritative for its own extruder count, so
skip the parent-shaped merge for the variant-keyed branch when cur >
target. Confirmed by loading a 2-extruder IDEX preset inheriting from a
single-nozzle base: all 24 variant-keyed options previously truncated
from child_size=2 -> 1 are now preserved at size 2.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
_picking_pass_imex_ghosts is what resets m_hover_ghost_head, but _picking_pass
early-returns (mouse drag, mouse off-canvas, gizmo drag) skip that reset. If
the user switches from an IMEX printer to a non-IMEX one during such a window
the plate clears its ghost volumes while the stale head index survives,
producing an orphan tooltip anchored to nothing.
Validate the hover state against live ghost volumes before rendering the
tooltip and self-heal the indices when they no longer point at anything.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Replace imex_head_transform's fifth argument (Vec3d primary_origin) with a
Vec2d primary_zone_center and rewrite the Mirror branch as a true reflection
about the plane x = primary_zone_center.x + gantry_offset.x/2. Previous math
flipped about the primary's current origin, which:
* let the ghost drift out of the target zone as the primary moved, and
* made mirrored drag motion track 1:1 with the primary instead of reflecting.
The new transform places the ghost at the mirrored position within the target
zone (matching where the mirror tool actually prints) and reflects drag so
primary +X → ghost -X while Y tracks 1:1 — i.e. the ghost stays a true
mirror while the user drags. Off-row Mirror targets (e.g. T3 on a 2x2) still
reflect across the same X-plane as on-row peers.
PartPlate::calc_imex_ghosts and update_imex_ghost_transforms now feed primary_off
(the primary head's zone center) instead of an instance-space Vec3d.
Mirror tests rewritten against the new geometric contract: ghost origin at the
reflected position, primary drag deltas reflected across the zone-boundary plane.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Centralize the project→printer→printer_extruder_id fallback for the physical
extruder map. PrintApply, PartPlate (ghost color + cache key), Plater (tooltip
+ click gate) all previously open-coded the three-step lookup, and each handled
the "pem unset, derive from pei" case slightly differently — an IDEX printer
without an explicit pem could paint an UNPRINTABLE_COLOR ghost even though the
slicer would have derived a valid mapping.
- IMEXHelpers: add effective_physical_extruder_map(explicit_pem, pei) and a
PresetBundle overload that wraps the project→printer precedence.
- PrintApply: use the helper in place of the inline pei→pem normalization.
- PartPlate / Plater: call the PresetBundle overload at every ghost-color,
ghost-cache-key, tooltip, and click-gate site.
- Plater::format_imex_ghost_tooltip: when no filament resolves to a head,
surface an actionable message directing the user to extend the extruder
count in the Machine tab, instead of the generic "(no filament routed)".
- IMEXFilamentPickerPopover: hold m_pem by value so callers can pass a
stack-local derived pem without lifetime worries.
- Tests: 5 new cases covering explicit-wins, default-pem fallback, null
inputs, and the IDEX ghost-color regression that motivated this.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
PartPlate::calc_imex_zones, GCodeViewer::render, and Plater::collect_imex_warnings
each hand-rolled their own "phys:P/C/M" tokenizer with subtly different error
handling. Replace the three inline loops with parse_imex_active_tools +
imex_primary_tool_for_mode so the Primary/Copy/Mirror classification agrees
across zones, the G-code viewer legend, and slice warnings.
No behavior change: the shared helpers preserve the 1=Primary / 2=Copy /
3=Mirror encoding already consumed downstream and continue to accept the
legacy bare-index form as Primary.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Replaces the plater-icon popover with colored transparent ghost copies
of primary-head instances on the plate, one per secondary active head
under its Copy/Mirror role transform. Left-click on a ghost opens a
compact filament picker popover for the ghost's head (MMU lane override).
Ghosts track the primary through drag/rotate/scale/mirror and invalidate
on mode or pem changes.
Key pieces:
IMEXHelpers -- imex_head_transform (Primary/Copy/Mirror), shared role
parser, per-head filament resolution with X-axis Mirror anchor.
PartPlate -- ghost state, volume rebuild on mode/map/object mutation,
primary_origin plumbed for Mirror reflection across the primary-row
gantry plane.
GLCanvas3D -- ghost rendering with per-head filament color and
translucent blending; picking routed via volume composite id.
Plater -- ghost click + tooltip; plater icon left-click always cycles.
IMEXFilamentPickerPopover -- BitmapComboBox row for one secondary head,
writes imex_head_filament_map on selection.
bbs_3mf -- round-trip the per-plate imex_head_filament_map option.
PrintConfig -- add imex_head_filament_map as a plate option.
MMU/AFC routing for parallel modes relies on the printer profile's
physical_extruder_map (see prior commit for authoring format). Primary-
row heads and their per-plate filament overrides are resolved through
that map, so PA and temperature emission address the correct physical
extruder when multiple logical slots share one carriage.
Tests: IMEXHelpers coverage for Primary/Copy/Mirror transforms
including a 2x2 off-row regression guard for the X-axis reflection fix.
Auto-populate physical_extruder_map (0-indexed) from printer_extruder_id
(1-indexed) in Print::apply(). The guard only runs when the map is still
at default size (<= 1 element), so printer profiles that set it explicitly
are untouched.
All IMEX parallel-mode PA and temperature emission now routes tool slot
indices through physical_extruder_map before constructing firmware
extruder qualifiers (EXTRUDER=, M104 T, M572 D). This ensures AFC/MMU
setups where multiple slots share one physical extruder get the correct
qualifier -- e.g. T6 on physical extruder 3 emits EXTRUDER=extruder3
instead of EXTRUDER=extruder6.
Profile authoring for MMU/AFC printers:
Add physical_extruder_map to the printer profile JSON as a 0-indexed
string array, one entry per logical filament slot, whose value is the
physical extruder carrying that slot. The array size must be > 1 for
the explicit map to override the auto-derive. Example for a 7-slot
printer with a 4-lane MMU on extruder 0 and three independent direct
drives on extruders 1/2/3:
"physical_extruder_map": ["0","0","0","0","1","2","3"]
Non-MMU printers need no action -- printer_extruder_id already encodes
the 1:1 mapping and the auto-derive handles it.
Deeper integration (zone validation, collision detection, filament
assignment grouping, Moonraker agent auto-population) is deferred.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The "Pre-slice warnings" line in the IDEX/IQEX Configuration section
was a widget-only line (no options) with full_width left at the default
of 0. activate_line() only skips the option_set.front() call when
full_width=1; without it, the code falls through to:
bool is_legend_line = option_set.front().opt.gui_type == ...
Calling front() on an empty std::vector is undefined behavior. On
Windows/MSVC release builds this dereferences a null pointer and reads
at offset 0x30 (where ConfigOptionDef::gui_type lands), producing an
ACCESS_VIOLATION at 0x30. On Linux/GCC the same UB happens to be
harmless, so the crash is Windows-only and cannot be reproduced on
Linux.
Fix: set line.full_width = 1, matching the pattern used by the "Modes"
(IMEXModesCtrl) line. This takes the early-return widget path in both
append_line and activate_line, bypassing option_set.front() entirely.
Reported by tester: crash on clicking Multimaterial tab with any
printer (K3D VOSTOK confirmed), build af59501f.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extends set_pressure_advance() with an optional tool index (default -1,
preserving existing behavior for all non-IMEX call sites). Per-firmware:
- Klipper: EXTRUDER=extruder[N] when tool >= 0, bare command otherwise
- RRF: M572 D<N> when tool >= 0, bare M572 otherwise (no D0 fallback)
- Marlin 2: M900 K<X> T<N> when tool >= 0, bare M900 otherwise
- Marlin Legacy / fallback: M900 K<X> always
Adds m_imex_parallel_mode to GCode, set once per export from the active
plate mode. PA and layer-transition temperature tool-qualification are
gated on this being a non-primary parallel mode — primary mode prints
use regular tool-change PA exactly like any non-IMEX printer. Secondary
active tools in parallel modes receive explicit per-tool PA at print
start since they never go through a tool-change sequence.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replaces MessageDialog with RichMessageDialog to show a suppress
checkbox on both slice-plate and slice-all warning paths. Persists
the choice to app_config as imex_pre_slice_warnings=false. Adds a
re-enable toggle in Printer Settings → Multimaterial → IDEX/IQEX
Configuration so the warnings can be restored if suppressed accidentally.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
update_visibility() and render() both called front() on an empty vector
when a CtrlLine had no options (pure widget lines). Added an early-return
path in update_visibility() and a null-guard in render().
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Temperature emission (GCode.cpp, PrintConfig.cpp, ConfigManipulation.cpp):
- Set temperatures for all active tools in layer_change_gcode for iMEX
parallel modes (primary + secondary carriages)
- Fix filament temperature commands so Layer 1 temperatures only emit on
the first layer; subsequent layers use normal layer-change temperatures
- Remove stray temperature commands that fired outside the intended context
- Consolidate iMEX temperature handling into the second-layer transition;
clean up emission logic and naming throughout
Pre-slice warning dialogs (PartPlate.cpp/hpp, Plater.cpp):
- Collect per-plate IMEX warnings before slicing: multi-material conflict,
bed temperature mismatch between carriages (>5 °C delta), and filament
type incompatibility
- Show a dismissible Yes/No dialog from both "Slice Plate" and "Slice All"
actions; No returns to 3D view, Yes proceeds to slice
- Refresh plate thumbnails after the panel switch so previously-generated
thumbnails are not left black
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
In iMEX (IDEX/IQEX) parallel printing modes (copy/mirror), only the
primary extruder generates toolpaths. The firmware duplicates the
primary's movements for secondary carriages, so they don't appear in
tool_ordering.all_extruders().
This caused is_extruder_used[N] to return false for secondary tools
even though they're physically active and moving.
The fix adds logic to parse the active mode's tool assignments from
imex_mode_active_tools config and marks all assigned tools as used.
Changes:
- Added null checks and bounds validation for config options
- Skip empty tool strings to avoid unnecessary parsing
- Reordered bounds checks for defensive programming
- Added clarifying comments for exception handling
This ensures is_extruder_used[N] is true for all tools in a parallel
mode, allowing printer profiles to correctly enable heaters and
emit cleanup G-code for all active carriages.
Fixes: is_extruder_used[1] returns false in copy/mirror modes (#13086)
Related: Comment [28]/[30], Comment [18] (is_extruder_used in G-code header)
- Add warning badge (obj_warning.svg overlay) to the iMEX plate icon when
a parallel mode is active alongside multi-material objects on the same plate
- Add has_imex_multimaterial_conflict() using get_extruders(true) so only
filaments actually used on the plate are checked
- Move multi-material caution dialog from reslice() into on_action_slice_plate /
on_action_slice_all so it fires exactly once per user action and does not
disrupt GL thumbnail generation during Slice All
- Fix is_imex missing from p->config init key list so on_config_change()
diff detection correctly triggers refresh_imex_icons()
- Defer imex_changed handling until after set_bed_shape() so m_shape is current
- Replace plain remove button with ScalableButton (imex_remove.svg) in mode rows
- Add EditGCodeDialog launch button per mode row for placeholder browsing
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add BeltBackTransform class that inverts the shear/scale matrix and
applies it in GCodeWriter::to_machine_coords() so G-code outputs in
the machine's physical coordinate space, gated by new
belt_gcode_back_transform config option
- Extend belt floor clipping to all three tree support pipelines
(Prusa-style, Orca organic, TreeModelVolumes) with per-layer polygon
clipping, anti-overhang integration, and belt raft extension layers
- Fix tree drop_nodes() belt termination, organic support global Z
offset, collision calculation index bug, and first-layer brim/empty
layer checks for belt printers
two-shot - first build built but didn't plumb to UI. Woah.
add pre-slice axis remap, because Y needs to be Z
going to change tactic and move based on bbox min
switch to per axis snapping
per axis swap snap now per object
build plate tilt wasn't invalidating slicer settings
support upper bound now correct, need to get lower bound corrected
axis swapped support termination corrected
Z Shear works with and without pre-slice remap now
- Fix support clipping z-shift calculation by removing coordinate-space
mismatch and sync belt_floor_z_shift with global_z_offset; fix
invalidation so posSupportMaterial no longer resets slicing params
- Add belt floor polygon clipping to non-organic tree support
(slim/strong/hybrid) with collision surface integration in
TreeSupportData, belt extension layers, and first-layer brim
suppression
- Add belt floor clipping to organic tree support pipeline with virtual
belt raft layers, per-layer polygons in TreeModelVolumes, and
post-generation layer trimming; fix pre-existing processing_last_mesh
bug in calculateCollision()
Fix belt floor support clipping: z-shift, invalidation, and global offset
- Fix support clipping z-shift calculation by removing coordinate-space
mismatch (raw_bounding_box min.z vs trafo_centered m_belt_min_z) and
sync belt_floor_z_shift with global_z_offset in global shear mode
- Fix invalidation so posSupportMaterial no longer resets slicing params,
preventing the exact posSlice z-shift from being overwritten by the
bounding-box approximation on support-only setting changes
- Remove double-counting of global z_offset on support layers — support
already inherits the offset from object layers during generation
This Work Was Co-Authored-By Claude Opus 4.6 (1M context) <noreply@anthropic.com>
UI: gray out inactive belt sub-options, rename to mesh transforms, move to Advanced
Fix mesh clipping through build plate after belt shear/scale transform
Generalize G-code viewer designed-view toggle for full belt transform
Clip support layers to transformed belt floor plane
Supports below the tilted build plate (Z = shear_factor * from_axis - min_z)
are now clipped via half-plane intersection after generation. Belt floor
parameters stored in SlicingParameters and populated in both update_slicing_parameters()
and the static slicing_parameters() overload.
Make belt G-code viewer toggle more prominent, add B keyboard shortcut
- Add separator + teal "Belt Printer" header in legend panel
- Append [B] hint to checkbox label
- Add B key shortcut in GLCanvas3D to toggle designed/machine view
- Read belt_printer_angle from loaded G-code headers to enable belt view
Add per-axis global transform option for belt printer shear
New belt_shear_{x,y,z}_global bool configs. When enabled, shear incorporates
instance shift so objects at different bed positions get position-aware
transform (Z += factor * instance_shift_on_from_axis).
Fix global shear: use layer Z offset instead of mesh transform, add config invalidation
- Global shear offset applied as post-slicing layer print_z adjustment
instead of mesh transform (which was absorbed by min_z normalization
or shifted mesh out of slice range)
- Register all belt transform options in Print::invalidate_state_by_config_options
to trigger posSlice re-slicing (the fallback only invalidated Print steps,
not PrintObject steps — belt changes had no effect without manual re-slice)
- Belt gcode remap options added to steps_gcode (gcode-export only)
- Skip empty-first-layer check for belt objects with global Z offset
WIP: split instances for global shear, relative Z offsets, debug logging
- PrintApply: when belt global mode active, prevent instance grouping by
adding unique Z perturbation to trafo — each copy becomes its own
PrintObject with independent layers
- PrintObjectSlice: compute global Z offset relative to minimum Y shift
across all PrintObjects (lowest-Y object stays at Z=0)
- Debug logging (warning level) for belt global shift values and offsets
Known issues:
- Cached posSlice results cause stale offsets when mixing copies with
individually-added objects — need to compute min baseline outside slice()
- Supports still generate to Z=0 instead of object's global Z offset
Fix global shear for copied objects: disable shared-object layer optimization
When belt global Z shear is active, each object needs unique layer Z
values based on its bed position. The shared-object optimization was
causing copies to reuse the source object's layers (and its Z offset)
instead of computing their own position-based offset.
started work on getting supports to work properly
one step forward, one step back
this version didn't quite work. Getting somewhere though
about to add UI controllable tests
added configuration options for supports
tweak CLAUDE.md to be more aggressive for my machine. This commit should probably be pulled out before contributing upstream
still chasing down some bugs
moving objects between slices no longer results in improper Z-height because of caching
added more data to the debug logs
Z offset is getting more global again
still not quite there, I think there's a fundamental logic flaw?
hunting for bugs
finally have a functional fix
Add belt floor clipping to tree supports (organic and non-organic)
- Add belt floor polygon clipping to non-organic tree support
(slim/strong/hybrid) in draw_circles() and terminate nodes at the
belt surface instead of the horizontal build plate
- Add belt floor clipping to organic tree support pipeline with virtual
belt raft layers for sub-floor branch generation, per-layer belt
floor polygons in TreeModelVolumes, and post-generation layer trimming
- Fix pre-existing processing_last_mesh bug in TreeModelVolumes that
prevented m_anti_overhang (support blockers) from ever being applied;
skip empty first layer check for belt printers
Commits:
current approach: make a face surface to build supports to
closer!
supports now terminate on shear plane, now need to get shear plane to correct Z height
nearly there
chasing down logic issues still
committing for checkpoint, this still does not work
still got logic problems...
cull support clipping
stashing changes for now. Going to focus on getting the global shear OFF support generation dialed first.
beginning per object shear calcs
Local shear transform is on correct Z offset now
local shear finally works now and needs more testing
global shear works now, needs thorough testing
debugging non-45 degree angles
debugging part 2
supports at all angles work now
remove debug logging
Add belt floor collision to non-organic tree support pipeline
- Integrate belt floor as a collision surface in TreeSupportData so
branches route around the belt naturally, replacing the explicit
termination checks in drop_nodes()
- Add belt extension layers below the object after draw_circles() to
allow support geometry to extend to the diagonal belt surface instead
of terminating at a horizontal first layer
- Fix coordinate overflow in belt floor polygons (scale_(1e4) exceeds
int32), skip first-layer brim expansion for belt printers, and
extend empty first layer check bypass to all belt modes
add debug logging, Z translate for tree supports
still not seeing any cutoff surface yet
adding debug options
attempt #2 at trees
if hit Z buildplate stop but don't set to_buildplate true
getting closer
tree support almost there, just need to get rid of the circles at the beginning
getting closer
belt / shear plane clip works, need to figure out the buidlplate plane issues
more logic, added debugging logs
supports now extend somewhat below Z=0 in global shear mode
fix bad alloc, add 10mm below build plate
fully works now
shear transform + prusa tree support generation works now.
pull out debug logging
- Implement per-object global shear transform in PrintObject with
layer Z-offset calculation, config invalidation, and fix for
shared-object layer optimization breaking copied objects
- Clip support layers to the transformed belt floor plane and begin
work on tree support adaptation for sheared coordinate space
- Improve belt UI: gray out inactive sub-options, add B keyboard
shortcut for G-code viewer design-view toggle, fix mesh clipping
through build plate after shear/scale transform
y' = y + z·cot(α),
while x' = x and z' = z
getting closer to customizable variant
getting closer
X/Y/Z shear initial
clean up UI
add 1/sin(a) transform, idea taken from blackbelt cura plugin
Things work now (turns out I've been using the wrong set of transforms)
- Replace monolithic belt rotation transform with independent per-axis
shear controls (mode/angle/source-axis for X, Y, Z) and G-code axis
remapping, giving full flexibility to match any belt printer's
coordinate system
- Remove all rotation mode logic and intermediate type+axes dropdowns,
simplifying the pipeline to pure shear matrices while preserving the
default behavior (Y += Z*cot(45deg) with identity remap)
- Clean up GCodeWriter, GCodeProcessor, and GCodeViewer for the new
shear-only model; expose 12 new settings in printer UI via
Tab.cpp/Preset.cpp
Implement belt printer tilted slicing
Implement the core belt slicing pipeline that makes the slicer
tilt-aware:
Step 1: GCodeWriter::to_machine_coords() - R(+alpha, X) rotation
from slicing frame to machine frame
Step 2: PrintObject - belt-rotated object height calculation
(y*sin(a) + z*cos(a)) for correct layer count
Step 3: PrintObjectSlice - apply R(-alpha, X) rotation trafo so
horizontal slice planes correspond to belt-parallel planes,
with Z-shift computed from model volumes
Step 4: GCodeProcessor - machine-frame preview (no transform needed)
Step 5: 3DBed - rotate bed visualization about X by belt angle
Fix: belt surface IS the build plate, no mesh rotation
Currently still slicing perpendicular to the belt normal. Need to figure out why.
Fix G-code Z sign: use R(-alpha, X) so Z+ is away from belt
The previous R(+alpha, X) transform produced negative Z values
(-y*sin(a) term dominated). Changed to R(-alpha, X) which gives
machine_z = y*sin(a) + z*cos(a), always positive for points
above the belt surface. Z increases with each layer as expected.
reverting and changing slice methodology
Add pink slicing direction arrow from origin
Shows the effective slicing direction (gantry normal) as a pink
arrow from the origin. Shorter and wider than the gravity arrow.
Direction: R(+alpha, X) * Z = (0, -sin(a), cos(a)), which is
the layer stacking direction in the original mesh frame.
Fix slicing arrow visibility and add raw G-code toggle
- Disable depth test for pink slicing arrow so it renders on top of
the tilted bed geometry (was being occluded)
- Remove unnecessary 5mm Z-offset from arrow position
- Add m_belt_show_raw toggle to GCodeViewer
- Add "Show raw G-code (slicing frame)" checkbox in legend when
belt mode is active
Implement to_machine_coords inverse rotation for belt printer G-code
The slicing pipeline rotates the mesh by R(-alpha, X) and shifts Z to
start at 0. The G-code output now undoes this transform via
to_machine_coords: R(+alpha, X) * T(0,0,+z_shift), recovering the
original machine-frame coordinates where Y is horizontal and Z is
vertical.
Changes:
- GCodeWriter: implement to_machine_coords with inverse rotation + Z-shift
- GCodeWriter: add belt_z_shift member and setter/getter
- GCode.cpp: compute Z-shift from print objects (same logic as
PrintObjectSlice) and pass to writer; write z_shift to G-code header
- GCodeProcessor: parse belt_z_shift from G-code header
- GCodeViewer: store belt_z_shift from processor result
Wire raw G-code toggle to apply slicing-frame view transform
When "Show raw G-code (slicing frame)" is checked in the preview
legend, the view matrix is modified to apply R(-alpha, X) * T(0,0,-z_shift)
to the toolpath rendering. This shows the G-code as it was during
slicing: rotated part with horizontal layers.
Default (unchecked): machine-frame view — upright part with tilted layers.
Remove belt printer placeholder comment from GCodeProcessor
The preview now correctly displays machine-frame G-code with the
optional raw view toggle. No transform is needed in the processor.
- Implement core belt slicing pipeline: R(-alpha, X) mesh rotation in PrintObjectSlice with corrected object height calculation for proper layer count
Add to_machine_coords() in GCodeWriter to convert slicing-frame coordinates back to machine-frame, propagated through GCode,
GCodeProcessor, and GCodeViewer
Add belt-mode UI: tilted bed visualization, slicing-direction arrow, and raw G-code toggle to switch between machine-frame and slicing-frame views
This is a combination of 6 commits.
checkpoint 1: initial MVP. Slicing functions, but rotates instead of skews are happening and a lot of other stuff too
getting somewhere, getting to the point where I need to figure out how to verify this stuff
this appears to be a dead end.
getting somewhere I think maybe
I'm pretty sure we've completely lost the plot at this point and need to restart this process...
remove slice logic in preparation for new, more invasive plan
Register imex_mode (string), imex_mode_index (int), and imex_mode_gcode
(string) in OtherSlicingStatesConfigDef so they appear in the placeholder
search UI under Slicing State.
Set all three via placeholder_parser().set() before any script processing
in _do_export(). Process imex_mode_gcode first so {global} declarations
defined there flow forward into machine_start_gcode.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
IMEXModesCtrl:
- Primary mode is now a non-deletable first row stored as sentinel
"primary" in imex_mode_names; older configs load cleanly
- New rows default T0 → Primary when no tool assignment is stored
- Filter "primary" from plater popup/cycle list to prevent double entry
- imex_tools_per_gantry cap raised 2 → 4
Zone sizing:
- Zone width/height now based on active tool count only; inactive tools
donate their bed share to active neighbors (fixes 4-tool layout)
- Active col/row maps (col_to_zone/row_to_zone) applied consistently
across zone fills, collision strips, and primary zone box
GCodeViewer animation:
- Mirror position formula fixed: left-of-copy reflects across copy
zone's left edge; right-of-copy reflects across right edge
(T3 was rendering on top of T1)
- strip_width/row_strip_height use active counts, matching PartPlate
Stability:
- Early return in calc_imex_zones() when tool_states is empty; prevents
OOB crash on new printer with stale process-preset mode name
- is_imex toggle in on_config_change calls refresh_imex_icons() so the
plate mode icon appears without requiring a new project
GCode:
- Remove "primary" guard so Primary mode gcode field is emitted at
start of print
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add all_tool_states map to IXexModesCtrl Row, storing all tool
assignments including those not currently visible due to grid size
- active_tools_string() now serializes from all_tool_states so
off-screen assignments survive the round trip through a smaller grid
- Button clicks keep all_tool_states in sync with visible btn_states
- Display anchors to the gantry row containing the Primary assignment
so reducing gantry count keeps the meaningful row visible rather
than always defaulting to row 0
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add per-plate iXex mode icon to the plate toolbar (normal, hover, dark, dark-hover SVG variants)
- Left-click cycles through available modes; right-click shows a popup menu with all modes as radio items
- Mode changes are recorded in the undo/redo snapshot system
- Fix double context menu: suppress EVT_GLCANVAS_PLATE_RIGHT_CLICK and EVT_GLCANVAS_RIGHT_CLICK when the iXex icon popup was already shown
- Remove ixex_parallel_mode combo from Print Settings > Other > Special mode (superseded by per-plate icon)
- Remove dead code: refresh_ixex_mode_combo(), m_ixex_mode_combo member, related Tab reload hook
- iXex mode persisted in 3MF project files via existing plate metadata serialization
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add per-plate iXex mode icon to the plate toolbar (normal, hover, dark, dark-hover SVG variants)
- Left-click cycles through available modes; right-click shows a popup menu with all modes as radio items
- Mode changes are recorded in the undo/redo snapshot system
- Fix double context menu: suppress EVT_GLCANVAS_PLATE_RIGHT_CLICK and EVT_GLCANVAS_RIGHT_CLICK when the iXex icon popup was already shown
- Remove ixex_parallel_mode combo from Print Settings > Other > Special mode (superseded by per-plate icon)
- Remove dead code: refresh_ixex_mode_combo(), m_ixex_mode_combo member, related Tab reload hook
- iXex mode persisted in 3MF project files via existing plate metadata serialization
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Includes two feature sets on top of the base iXex implementation:
Auto-arrange iXex zone constraint:
Constrains placement to the primary zone when a parallel mode is active.
Collision strips registered as hard obstacles via m_unselected so the
NFP placer correctly excludes them.
Nozzle clearance rename + carriage box visualization overhaul:
ixex_carriage_width_x/y → ixex_nozzle_clearance_x/y (breaking rename).
Collision strip width now uses the literal clearance value, not half.
Carriage footprint boxes place the nozzle at the correct physical edge
for all printer types (IDEX, IQEX 2x2) and modes (copy, mirror).
Fixed GLModel::reset() bug causing stale mesh on config change.
Config key rename (breaking for saved profiles — call out in PR):
ixex_carriage_width_x/y → ixex_nozzle_clearance_x/y
Labels updated to "Nozzle Clearance X/Y" with consistent tooltips
describing the measurement as nozzle-to-collision-side-edge distance.
Strip math fix:
Previously halved the clearance value (× 0.5) under the assumption
the nozzle was centered in the carriage. The measurement is now the
literal nozzle-to-edge distance, so the × 0.5 factor is removed.
The collision strip width now equals the configured value directly.
Carriage box visualization (GCodeViewer):
- Add per-carriage box_offset_x/y so the nozzle marker sits at the
physically correct edge of the footprint box rather than centered.
- X: zone-based by default (nozzle at inner edge facing bed center).
Copy secondaries inherit the primary's X orientation (same movement
direction). Mirror secondaries use the collision-side edge.
- Y: always row-based regardless of copy/mirror mode. Gantry is always
behind the nozzle (high-Y); front-row primaries with a back-row
secondary override to place nozzle at the low-Y edge.
- Fix stale mesh bug: GLModel::init_from() is a no-op when already
initialized. Call reset() before init_from() so mesh rebuilds
correctly when nozzle clearance values change in config.
- Remove m_ixex_toolhead_box_dims (was the now-unnecessary cache key).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
params.excluded_regions flows into libnest2d PlacementConfig.m_excluded_regions
which is declared but never consumed by the NFP placer — the field is dead code.
m_unselected is the working path: it becomes fixeditems that are preloaded as
physical fixed obstacles in the NFP computation, same as the wipe tower.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Auto-arrange now injects the carriage collision zones as hard excluded
regions (is_virt_object) when an iXex parallel mode is active, using the
same BoundingBoxf3 data already stored for rendering and violation checks.
Objects will no longer be placed in the mirror-edge danger strips.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds std::optional<BoundingBoxf> m_ixex_primary_zone_box to PartPlate,
populated by calc_ixex_zones() alongside the existing secondary/collision
zone geometry. The new ixex_primary_zone() getter calls ensure_ixex_zones()
so callers always get fresh data. ArrangeJob::process() replaces the
full-bed bedpts with the primary zone corners when the getter returns a
value, so auto-arrange no longer drops objects into the bed center.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The thin rectangular outlines around copy/mirror zones caused visible
aliasing and appeared to flash during interaction. Removed the border
GLModel, its build code in calc_ixex_zones(), the render block, and
the unused border field from IXexTheme.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Same class of bug as the TabPrinter fix: clear_pages() destroys all
page widgets but didn't null m_ixex_mode_combo, causing a SIGSEGV in
refresh_ixex_mode_combo() when load_current_preset() ran after a
page rebuild (e.g. triggered by Plater::reset()).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Arrays cannot be captured by value in C++ lambdas; replaced with a
static local declared inside the event handler.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Introduces first-class parallel printing (copy/mirror modes) for printers
with multiple independent X-axis carriages. Branded iXex (independent X
extruder), targeting Klipper firmware with a firmware-agnostic design.
- PrintConfig: new printer options declaring iXex capability and geometry
(is_ixex, ixex_gantry_count, ixex_tools_per_gantry, carriage dims,
tool layout, and per-mode name/role/gcode arrays)
- Preset: iXex keys registered in printer and process preset option lists
- Tab: IXexModesCtrl visual grid editor in Printer preset tab; mode
dropdown in Process → Others tab; clear_pages() nulls iXex pointers to
prevent dangling-pointer crash on preset save
- PartPlate: 2D zone visualization (active/dimmed/dividers) and
placement-violation detection (has_ixex_placement_violations) that
blocks slicing when objects fall outside the primary zone
- Plater: violation detection wired into update_background_process so
the Slice button is disabled with an error notification on violation
- GCode: mode-activation G-code injected before machine_start_gcode
- GCodeViewer: multi-carriage toolhead markers in sequential preview,
filament legend annotated with active carriage count and mode name
- Add BeltBackTransform class that inverts the shear/scale matrix and
applies it in GCodeWriter::to_machine_coords() so G-code outputs in
the machine's physical coordinate space, gated by new
belt_gcode_back_transform config option
- Extend belt floor clipping to all three tree support pipelines
(Prusa-style, Orca organic, TreeModelVolumes) with per-layer polygon
clipping, anti-overhang integration, and belt raft extension layers
- Fix tree drop_nodes() belt termination, organic support global Z
offset, collision calculation index bug, and first-layer brim/empty
layer checks for belt printers
two-shot - first build built but didn't plumb to UI. Woah.
add pre-slice axis remap, because Y needs to be Z
going to change tactic and move based on bbox min
switch to per axis snapping
per axis swap snap now per object
build plate tilt wasn't invalidating slicer settings
support upper bound now correct, need to get lower bound corrected
axis swapped support termination corrected
Z Shear works with and without pre-slice remap now
- Fix support clipping z-shift calculation by removing coordinate-space
mismatch and sync belt_floor_z_shift with global_z_offset; fix
invalidation so posSupportMaterial no longer resets slicing params
- Add belt floor polygon clipping to non-organic tree support
(slim/strong/hybrid) with collision surface integration in
TreeSupportData, belt extension layers, and first-layer brim
suppression
- Add belt floor clipping to organic tree support pipeline with virtual
belt raft layers, per-layer polygons in TreeModelVolumes, and
post-generation layer trimming; fix pre-existing processing_last_mesh
bug in calculateCollision()
Fix belt floor support clipping: z-shift, invalidation, and global offset
- Fix support clipping z-shift calculation by removing coordinate-space
mismatch (raw_bounding_box min.z vs trafo_centered m_belt_min_z) and
sync belt_floor_z_shift with global_z_offset in global shear mode
- Fix invalidation so posSupportMaterial no longer resets slicing params,
preventing the exact posSlice z-shift from being overwritten by the
bounding-box approximation on support-only setting changes
- Remove double-counting of global z_offset on support layers — support
already inherits the offset from object layers during generation
This Work Was Co-Authored-By Claude Opus 4.6 (1M context) <noreply@anthropic.com>
UI: gray out inactive belt sub-options, rename to mesh transforms, move to Advanced
Fix mesh clipping through build plate after belt shear/scale transform
Generalize G-code viewer designed-view toggle for full belt transform
Clip support layers to transformed belt floor plane
Supports below the tilted build plate (Z = shear_factor * from_axis - min_z)
are now clipped via half-plane intersection after generation. Belt floor
parameters stored in SlicingParameters and populated in both update_slicing_parameters()
and the static slicing_parameters() overload.
Make belt G-code viewer toggle more prominent, add B keyboard shortcut
- Add separator + teal "Belt Printer" header in legend panel
- Append [B] hint to checkbox label
- Add B key shortcut in GLCanvas3D to toggle designed/machine view
- Read belt_printer_angle from loaded G-code headers to enable belt view
Add per-axis global transform option for belt printer shear
New belt_shear_{x,y,z}_global bool configs. When enabled, shear incorporates
instance shift so objects at different bed positions get position-aware
transform (Z += factor * instance_shift_on_from_axis).
Fix global shear: use layer Z offset instead of mesh transform, add config invalidation
- Global shear offset applied as post-slicing layer print_z adjustment
instead of mesh transform (which was absorbed by min_z normalization
or shifted mesh out of slice range)
- Register all belt transform options in Print::invalidate_state_by_config_options
to trigger posSlice re-slicing (the fallback only invalidated Print steps,
not PrintObject steps — belt changes had no effect without manual re-slice)
- Belt gcode remap options added to steps_gcode (gcode-export only)
- Skip empty-first-layer check for belt objects with global Z offset
WIP: split instances for global shear, relative Z offsets, debug logging
- PrintApply: when belt global mode active, prevent instance grouping by
adding unique Z perturbation to trafo — each copy becomes its own
PrintObject with independent layers
- PrintObjectSlice: compute global Z offset relative to minimum Y shift
across all PrintObjects (lowest-Y object stays at Z=0)
- Debug logging (warning level) for belt global shift values and offsets
Known issues:
- Cached posSlice results cause stale offsets when mixing copies with
individually-added objects — need to compute min baseline outside slice()
- Supports still generate to Z=0 instead of object's global Z offset
Fix global shear for copied objects: disable shared-object layer optimization
When belt global Z shear is active, each object needs unique layer Z
values based on its bed position. The shared-object optimization was
causing copies to reuse the source object's layers (and its Z offset)
instead of computing their own position-based offset.
started work on getting supports to work properly
one step forward, one step back
this version didn't quite work. Getting somewhere though
about to add UI controllable tests
added configuration options for supports
tweak CLAUDE.md to be more aggressive for my machine. This commit should probably be pulled out before contributing upstream
still chasing down some bugs
moving objects between slices no longer results in improper Z-height because of caching
added more data to the debug logs
Z offset is getting more global again
still not quite there, I think there's a fundamental logic flaw?
hunting for bugs
finally have a functional fix
Add belt floor clipping to tree supports (organic and non-organic)
- Add belt floor polygon clipping to non-organic tree support
(slim/strong/hybrid) in draw_circles() and terminate nodes at the
belt surface instead of the horizontal build plate
- Add belt floor clipping to organic tree support pipeline with virtual
belt raft layers for sub-floor branch generation, per-layer belt
floor polygons in TreeModelVolumes, and post-generation layer trimming
- Fix pre-existing processing_last_mesh bug in TreeModelVolumes that
prevented m_anti_overhang (support blockers) from ever being applied;
skip empty first layer check for belt printers
Commits:
current approach: make a face surface to build supports to
closer!
supports now terminate on shear plane, now need to get shear plane to correct Z height
nearly there
chasing down logic issues still
committing for checkpoint, this still does not work
still got logic problems...
cull support clipping
stashing changes for now. Going to focus on getting the global shear OFF support generation dialed first.
beginning per object shear calcs
Local shear transform is on correct Z offset now
local shear finally works now and needs more testing
global shear works now, needs thorough testing
debugging non-45 degree angles
debugging part 2
supports at all angles work now
remove debug logging
Add belt floor collision to non-organic tree support pipeline
- Integrate belt floor as a collision surface in TreeSupportData so
branches route around the belt naturally, replacing the explicit
termination checks in drop_nodes()
- Add belt extension layers below the object after draw_circles() to
allow support geometry to extend to the diagonal belt surface instead
of terminating at a horizontal first layer
- Fix coordinate overflow in belt floor polygons (scale_(1e4) exceeds
int32), skip first-layer brim expansion for belt printers, and
extend empty first layer check bypass to all belt modes
add debug logging, Z translate for tree supports
still not seeing any cutoff surface yet
adding debug options
attempt #2 at trees
if hit Z buildplate stop but don't set to_buildplate true
getting closer
tree support almost there, just need to get rid of the circles at the beginning
getting closer
belt / shear plane clip works, need to figure out the buidlplate plane issues
more logic, added debugging logs
supports now extend somewhat below Z=0 in global shear mode
fix bad alloc, add 10mm below build plate
fully works now
shear transform + prusa tree support generation works now.
pull out debug logging
- Implement per-object global shear transform in PrintObject with
layer Z-offset calculation, config invalidation, and fix for
shared-object layer optimization breaking copied objects
- Clip support layers to the transformed belt floor plane and begin
work on tree support adaptation for sheared coordinate space
- Improve belt UI: gray out inactive sub-options, add B keyboard
shortcut for G-code viewer design-view toggle, fix mesh clipping
through build plate after shear/scale transform
y' = y + z·cot(α),
while x' = x and z' = z
getting closer to customizable variant
getting closer
X/Y/Z shear initial
clean up UI
add 1/sin(a) transform, idea taken from blackbelt cura plugin
Things work now (turns out I've been using the wrong set of transforms)
- Replace monolithic belt rotation transform with independent per-axis
shear controls (mode/angle/source-axis for X, Y, Z) and G-code axis
remapping, giving full flexibility to match any belt printer's
coordinate system
- Remove all rotation mode logic and intermediate type+axes dropdowns,
simplifying the pipeline to pure shear matrices while preserving the
default behavior (Y += Z*cot(45deg) with identity remap)
- Clean up GCodeWriter, GCodeProcessor, and GCodeViewer for the new
shear-only model; expose 12 new settings in printer UI via
Tab.cpp/Preset.cpp
Implement belt printer tilted slicing
Implement the core belt slicing pipeline that makes the slicer
tilt-aware:
Step 1: GCodeWriter::to_machine_coords() - R(+alpha, X) rotation
from slicing frame to machine frame
Step 2: PrintObject - belt-rotated object height calculation
(y*sin(a) + z*cos(a)) for correct layer count
Step 3: PrintObjectSlice - apply R(-alpha, X) rotation trafo so
horizontal slice planes correspond to belt-parallel planes,
with Z-shift computed from model volumes
Step 4: GCodeProcessor - machine-frame preview (no transform needed)
Step 5: 3DBed - rotate bed visualization about X by belt angle
Fix: belt surface IS the build plate, no mesh rotation
Currently still slicing perpendicular to the belt normal. Need to figure out why.
Fix G-code Z sign: use R(-alpha, X) so Z+ is away from belt
The previous R(+alpha, X) transform produced negative Z values
(-y*sin(a) term dominated). Changed to R(-alpha, X) which gives
machine_z = y*sin(a) + z*cos(a), always positive for points
above the belt surface. Z increases with each layer as expected.
reverting and changing slice methodology
Add pink slicing direction arrow from origin
Shows the effective slicing direction (gantry normal) as a pink
arrow from the origin. Shorter and wider than the gravity arrow.
Direction: R(+alpha, X) * Z = (0, -sin(a), cos(a)), which is
the layer stacking direction in the original mesh frame.
Fix slicing arrow visibility and add raw G-code toggle
- Disable depth test for pink slicing arrow so it renders on top of
the tilted bed geometry (was being occluded)
- Remove unnecessary 5mm Z-offset from arrow position
- Add m_belt_show_raw toggle to GCodeViewer
- Add "Show raw G-code (slicing frame)" checkbox in legend when
belt mode is active
Implement to_machine_coords inverse rotation for belt printer G-code
The slicing pipeline rotates the mesh by R(-alpha, X) and shifts Z to
start at 0. The G-code output now undoes this transform via
to_machine_coords: R(+alpha, X) * T(0,0,+z_shift), recovering the
original machine-frame coordinates where Y is horizontal and Z is
vertical.
Changes:
- GCodeWriter: implement to_machine_coords with inverse rotation + Z-shift
- GCodeWriter: add belt_z_shift member and setter/getter
- GCode.cpp: compute Z-shift from print objects (same logic as
PrintObjectSlice) and pass to writer; write z_shift to G-code header
- GCodeProcessor: parse belt_z_shift from G-code header
- GCodeViewer: store belt_z_shift from processor result
Wire raw G-code toggle to apply slicing-frame view transform
When "Show raw G-code (slicing frame)" is checked in the preview
legend, the view matrix is modified to apply R(-alpha, X) * T(0,0,-z_shift)
to the toolpath rendering. This shows the G-code as it was during
slicing: rotated part with horizontal layers.
Default (unchecked): machine-frame view — upright part with tilted layers.
Remove belt printer placeholder comment from GCodeProcessor
The preview now correctly displays machine-frame G-code with the
optional raw view toggle. No transform is needed in the processor.
- Implement core belt slicing pipeline: R(-alpha, X) mesh rotation in PrintObjectSlice with corrected object height calculation for proper layer count
Add to_machine_coords() in GCodeWriter to convert slicing-frame coordinates back to machine-frame, propagated through GCode,
GCodeProcessor, and GCodeViewer
Add belt-mode UI: tilted bed visualization, slicing-direction arrow, and raw G-code toggle to switch between machine-frame and slicing-frame views
This is a combination of 6 commits.
checkpoint 1: initial MVP. Slicing functions, but rotates instead of skews are happening and a lot of other stuff too
getting somewhere, getting to the point where I need to figure out how to verify this stuff
this appears to be a dead end.
getting somewhere I think maybe
I'm pretty sure we've completely lost the plot at this point and need to restart this process...
remove slice logic in preparation for new, more invasive plan
description:Use when creating, modifying, reviewing or debugging OrcaSlicer FFF system profiles under resources/profiles, including printer/vendor/nozzle/material additions, bundle indexes and versions, preset renames, setting_id and filament_id, and moving settings that sibling presets repeat onto shared bases after fix-variant or while drafting. Also use for missing presets or vendors, ignored profile settings, ambiguous AMS filament matches, and failures from orca_profile_tool.py, check_profile.sh/.bat, OrcaSlicer_profile_validator or the Check profiles CI job.
---
# OrcaSlicer system profiles
This skill describes how OrcaSlicer system profiles are drafted and shaped: the rules, equations and
patterns a profile follows. Use it to draft new profiles, modify existing ones, fix profile issues and
review profile changes.
A bundle is the index `resources/profiles/<Vendor>.json` plus the folder `<Vendor>/`. The vendor id is
the filename stem (`BBL`), not the index's display `name` (`Bambulab`). The index is the loader's only
entry point: an unindexed preset never loads. `OrcaFilamentLibrary` is the shared filament bundle,
loaded first; `blacklist.json` is data, not a bundle.
## References
Read the reference for the task before editing; load others only when the task crosses into them.
Paths below are relative to this skill. Commands run from the repository root; on Windows use `py -3`
for `python3`.
| Task | Read |
| --- | --- |
| Add or tune a filament, brand or material; fix compatibility, alias shadowing or overlapping coverage | [filament-profiles.md](references/filament-profiles.md) |
| Add a printer or nozzle; change models, variants, assets or per-extruder vectors | [machine-profiles.md](references/machine-profiles.md) |
| Add or tune extruder variants (`extruder_type` Direct Drive / Bowden × nozzle volume type Standard / High Flow / TPU High Flow / E3D High Flow / Extra High Flow variants) on a printer, process or filament | [extruder-variants.md](references/extruder-variants.md) |
| Add a quality tier or tune a process | [process-profiles.md](references/process-profiles.md) |
| Draft several presets, or clean up after `fix-variant`: which base each shared value belongs on, when a new base pays off, proving nothing loads differently | [shared-bases.md](references/shared-bases.md) |
| Name a preset; check what a name must equal; base names, uniqueness, filenames | [naming.md](references/naming.md) |
| Create a vendor bundle; index, `version`, `inherits`, `include`; migrate preset names; diagnose why a bundle fails to load | [vendor-bundle.md](references/vendor-bundle.md) |
| Change ids; diagnose AMS identity | [ids.md](references/ids.md), then `docs/HLSD/filament_id.md` for identity changes |
| Run checks, interpret failures, test another tree or verify in the app | [validation.md](references/validation.md) |
| Review a profile diff | [review-checklist.md](references/review-checklist.md) |
## Rules
1.**Bump the `version` of every bundle you change**, `OrcaFilamentLibrary.json` included when affected.
Increment the last component and carry `.99` into the third (`02.04.00.99` → `02.04.01.00`). The
updater installs only a strictly newer version, and CI does not check the bump.
2.**Register every preset, bases included, parents before children.**`update-index` writes the four
`*_list` arrays from the files on disk; `check` fails unless the index equals its output. Each index
entry's `name` must equal the file's `name`.
3.**Generate ids; never invent or copy them.** Keep existing ids during ordinary tuning. New presets
normally omit them until `generate-id`; bases carry no `setting_id`. BBL's own `setting_id`s and a wrongly
inherited `filament_id` need the explicit handling in [ids.md](references/ids.md).
4.**A name is an identity; preserve shipped selectable names.** Every reference (`inherits`,
`compatible_printers`, `default_*`, `printer_model`) is the exact, case-sensitive `name`. Renaming or
deleting a shipped selectable preset, or flipping its `instantiation` from `"true"` to `"false"`,
needs `renamed_from` (a `;`-separated string) on a selectable successor
12.**Run the full checks before reporting completion.** A `--vendor` run is only a development loop.
Review also covers version bumps, assets, non-default processes and hardware tuning, which CI cannot
establish.
## Names
| Type | Shape | What the loader uses |
| --- | --- | --- |
| `machine_model` | `<Model>` (`Bambu Lab X1 Carbon`) | the exact string, named by each variant's `printer_model`; also the `<Model>_cover.png` stem |
| `machine` | `<Model> <nozzle> nozzle` | the exact string, named by `compatible_printers`; `printer_variant` holds the nozzle token (`0.4`, [rules](references/machine-profiles.md#printer_model-and-printer_variant)) |
| `process` | `<lh>mm <Quality> @<target>` | the exact string when referenced or selected; `@<target>` is a label, compatibility comes from the preset's list or condition |
| `filament` | `<Product> @<target>` | text before the first `@` is the alias (shadowing, `filament_id`); the rest is a label, with reserved targets `@base` and `@System` |
The shapes are convention; `check` enforces only uniqueness. Per-type conventions, base names and
filename rules are in [naming.md](references/naming.md).
## Creating or modifying a profile
1.**Inspect the diff and the neighbouring presets.** Read their `name`, parent chain and children:
edits to a base, or to a leaf that others inherit, propagate. Match the bundle's structure and write
only overrides. New files use tab indentation, LF and a trailing newline; preserve unrelated
formatting in existing files. Match filename case exactly and use
6. **Verify the changed behaviour.** Slice newly added non-default processes and filaments
[explicitly](references/validation.md#checking-a-copy-of-the-tree), and
[test in the app](references/validation.md#testing-in-the-app) for selection or UI behaviour. Report
the checks actually run, their failures and skips, and any hardware tuning still unverified.
## Symptom → first look
| Symptom | Start here |
| --- | --- |
| A vendor disappears | the app's log or the `validate_system` log; [failure scopes](references/vendor-bundle.md#failure-scopes) |
| A setting has no effect | key spelling or a legacy name (`PrintConfigDef::handle_legacy`), value type, or a config key placed on a `machine_model` |
| A preset exists but is not selectable | index registration, `instantiation`, whether it is installed (chosen in the setup wizard, or listed in the model's `default_materials`), compatibility |
| A filament is missing, duplicated, or matches the wrong spool | [compatibility and alias shadowing](references/filament-profiles.md#compatible_printers), [ids](references/ids.md) |
| Presets differ only by colour, or an all-printer library preset lacks `@System` | [colour is a runtime property](references/filament-profiles.md#colour-is-a-runtime-property) |
| High Flow (or a second extruder) slices with Standard (or extruder 1) values; a variant switch is missing; a variant's tuned values never arrive | [extruder variants](references/extruder-variants.md#variant-strings), [widths](references/extruder-variants.md#widths), [variant names](references/validation.md#variant-names) |
| Values land on the wrong extruder or mode after a variant was added | [inserting a variant](references/extruder-variants.md#adding-a-variant-inserts-its-values-at-its-variant-index), [padding and composition](references/extruder-variants.md#padding-truncation-and-composition) |
| A resolved value matches neither the file nor its `inherits` parent; an array is not the width the file wrote, or a child got only a base's first value | [composition order and `include`](references/vendor-bundle.md#inherits-and-include) |
| A bed temperature is ignored | [the twelve plate keys](references/filament-profiles.md#bed-temperature-is-twelve-keys-not-one) |
| A change is absent from the running app | version bump and [installed profile location](references/validation.md#testing-in-the-app) |
| A check fails | [error → remedy](references/validation.md#error--remedy) |
## Source of truth
When this skill and the checkout disagree, the checkout wins: `scripts/orca_profile_tool.py` for the
tool and its flags; `src/libslic3r/Preset*.cpp` for loading and compatibility;
`src/libslic3r/PrintConfig.cpp` for keys, types, nullable options, the variant key sets and legacy
handling; `src/dev-utils/OrcaSlicer_profile_validator.cpp` and `.github/workflows/check_profiles.yml`
for validation coverage; `docs/HLSD/filament_id.md` for filament identity. The wiki's
Use this when a printer's hotend or extruder can be in more than one hardware configuration that
needs different settings: a Standard and a High Flow nozzle, a TPU High Flow, E3D High Flow or Extra High Flow nozzle, or
a Direct Drive and a Bowden extruder on one machine. Variants let one printer, process and filament preset carry a
separate value per configuration; the user picks the configuration in the sidebar, and slicing uses
the matching values. They are not for nozzle **diameter**: that stays one `machine` preset per
`printer_variant`.
## Variant strings
- A **variant string** is `"<extruder_type> <nozzle volume type>"`: the extruder's `extruder_type`
value, a space, and its nozzle volume type (the project's `nozzle_volume_type`, seeded by the
printer preset's `default_nozzle_volume_type`), e.g. `"Direct Drive High Flow"`.
- A **variant** is one entry of a preset's variant list, named by its variant string (plus an extruder
id for printer and process lists). A variant-aware key holds **one value per variant** (a
(normal, silent) pair for the `machine_max_*` limits), and each preset declares its variants in
that list. Slicing picks, for each extruder, the variant whose
variant string equals the current `extruder_type` + nozzle volume type.
- Matching is an **exact string compare** of the whole string (plus the extruder id for printer and
process lists) against the preset's own resolved list, in any order. With no match the **first
variant** is used, silently: variant index 0 for printer and process keys (extruder 1's first
variant, whichever extruder asks), the filament's own first variant for filament keys. Printer and
process keys are matched only on a printer with several extruders or whose
`extruder_variant_list` offers several variant strings (`support_different_extruders`); on any
other printer their variant index 0 is read whatever it names. Nothing rejects a profile for a
variant mismatch; mistakes surface only as wrong values in the G-code.
- Variant index and array length follow [Widths](#widths).
The complete enum is `s_keys_map_ExtruderType` and `s_keys_map_NozzleVolumeType` in
`src/libslic3r/PrintConfig.cpp` (`grep -A5 s_keys_map_NozzleVolumeType` there to confirm):
| Part | Values |
| --- | --- |
| Extruder type | `Direct Drive`, `Bowden` |
| Nozzle volume type | `Standard`, `High Flow`, `TPU High Flow`, `E3D High Flow`, `Extra High Flow` (`Hybrid` exists but is runtime-only) |
So the ten legal variant strings are the two extruder types × the five writable nozzle volume types,
and this table is the whole test of legality: a string's presence in a shipped profile is no evidence
for it. `Hybrid` (an extruder with several sub-nozzles) is never a variant: a filament on it reads the
variant of its own nozzle volume type from the project's `filament_volume_map`, printer and process
keys get one variant per nozzle volume type the extruder holds (`extruder_nozzle_stats`), and any
other lookup reads `Standard`. Never write it in a variant string. The same per-filament and
per-type reading applies on any extruder once `extruder_nozzle_stats` lists more nozzle volume
types than there are extruders.
Every other string (a nozzle volume type name from another slicer, a typo, a variant copied from a
shipped profile) is a **dead variant**: nothing selects it, and since lookup is by string it does not
shift the variants beside it; it still counts toward the variant length when arrays are sized.
`orca_profile_tool.py check` reports it as an error in every bundle, BBL included
([variant names](validation.md#variant-names)). Legacy names are errors too: in the four variant
lists, `default_nozzle_volume_type` and `nozzle_volume_type`, the loader still rewrites `Normal` →
`Standard` and `Big Traffic` → `High Flow`, so a ported `Direct Drive Normal` variant would load, but
`check` rejects the spelling and names the enum name to write. `DirectDrive` is rewritten only in
`extruder_type`, so `DirectDrive Standard` in a variant list is dead. Write the enum names;
`normalize` does not convert them.
### A nozzle the enum does not name
The nozzle volume types are fixed by the engine, and a profile cannot add one. A nozzle the table does
not name needs the type added in code first, which is outside profile work; until then any string
for it is a dead variant. Once the engine has the type, its variant string is the enum name after
the extruder type, e.g. `"Direct Drive <name>"`, and `check` accepts it with no tool change, since it
reads the enum from `PrintConfig.cpp`. Existing arrays are unaffected
([slice time](#slice-time-and-existing-users)).
## When to use variants
| Hardware | Do |
| --- | --- |
| One extruder, one nozzle type | Nothing. Without a variant list the printer has one variant ([variant length](#widths)) and no printer-key lookup takes place; a filament with several variants still gets the one for the printer's variant string. |
| Bowden-only printer | Nothing either. A list-less printer matches no string, so every printer-key lookup reads variant index 0 whether that variant is called `"Bowden Standard"` or `"Direct Drive Standard"`; naming it is cosmetic while it is the printer's only variant. A multi-extruder Bowden printer that declares the layout writes `"Bowden Standard"` variants: `"Direct Drive Standard"` entries match none of its extruders, so every extruder reads variant index 0 (and `check` reports them under [Printer rule 3](#printer-machine)). A filament with a `"Bowden Standard"` variant does get that variant there. |
| Nozzle types the user swaps (Standard / High Flow / TPU High Flow / E3D High Flow / Extra High Flow) | The printer lists them as variants; tune the keys that really differ per nozzle volume type. |
| Extruders of different types on one machine | One `extruder_type` per extruder, each extruder listing its own variants. |
| Several extruders that need different values in a variant key (retraction, z-hop, `nozzle_volume`, the `machine_max_*` limits) | Declare the layout even with a single nozzle volume type: `extruder_variant_list` with one `"<type> Standard"` per extruder, and the flattened pair. Without it the arrays still hold one value per extruder ([variant length](#widths)), but the loader keeps only extruder 1's value of a list-less printer's arrays, so every extruder prints with it. |
A preset without variant keys keeps working on a variant printer: its single variant is applied to
every extruder. So adding variants to a printer does not break existing processes or library
filaments; it only makes per-variant tuning possible.
## Printer (`machine`)
```json
"extruder_type":["Direct Drive","Direct Drive"],
"extruder_variant_list":["Direct Drive Standard,Direct Drive High Flow",
"Direct Drive Standard,Direct Drive High Flow,Direct Drive TPU High Flow"],
"printer_extruder_id":["1","1","2","2","2"],
"printer_extruder_variant":["Direct Drive Standard","Direct Drive High Flow",
"Direct Drive Standard","Direct Drive High Flow","Direct Drive TPU High Flow"],
1.`extruder_variant_list` has **one entry per extruder**; each entry is the `,`-joined variants that
extruder supports. It is the per-extruder menu the sidebar offers. It is in no
[variant set](#the-four-key-sets), so the variant-length resize leaves it alone. Extruder 1's
first variant, variant index 0 of the flattened pair, is the fallback of every extruder whose
variant is missing when the arrays are collapsed for slicing
([slice time](#slice-time-and-existing-users)); it is not necessarily the configuration the user
sees, which is `default_nozzle_volume_type` (rule 4).
2.`printer_extruder_variant` is that list **flattened** extruder-major, one entry per variant, and
`printer_extruder_id` gives each entry its 1-based extruder. These two size and address every
variant. Write all three keys and keep them in agreement. At load with
`single_extruder_multi_material` off, and in the app when the printer tab loads a printer with a
different number of extruders, the pair is rebuilt from `extruder_variant_list` (one
`Direct Drive Standard` per extruder when the list is absent) and the variant arrays are resized to
the rebuilt pair, padded with their first value or cut. The `machine_max_*` limits are padded the
same way, so extruder 2 and up of a list-less printer take extruder 1's normal limit as their
silent one too. The resize skips `hotend_heating_rate` / `hotend_cooling_rate`: they keep their
width, and an extruder beyond it reads their first value. With the three in agreement that changes
nothing; a pair written without the list is replaced. A listed variant the pair lacks is a menu
choice that reads variant index 0.
- The pair without `extruder_variant_list` slices, but the sidebar offers no variant switch and
the app cannot add variants to a list-less process: nothing is lost while every extruder
has exactly one variant, and every further variant is unreachable.
- A missing or one-value `printer_extruder_id` is extruder 1 at every index
([the id trap](#padding-truncation-and-composition)) unless the load-time rebuild above
replaces the pair (`single_extruder_multi_material` off).
-`check` reports against this rule ([variant arrays](validation.md#variant-arrays)). Errors: a
pair that is not the flattening, an id array that does not give each entry its extruder, a list
of more than one variant without the pair, and a pair without the list that puts several variants
on one extruder. A pair without the list and one variant per extruder is a warning where the
load-time rebuild would replace it (`single_extruder_multi_material` off, and a pair other than
one `Direct Drive Standard` per extruder), and passes otherwise.
3. Every variant string in an entry must start with that extruder's `extruder_type`.
4.`default_nozzle_volume_type` has one value per extruder and must name a nozzle volume type that
extruder's variants list; it seeds the sidebar. The live choice is `nozzle_volume_type` in the
project config, never a preset key.
5. Size every array by the set its key belongs to (the three sets are listed in full in
[The four key sets](#the-four-key-sets); the lengths are worked through in the
[sizing equation](#sizing-equation)): exactly the length below in each selectable preset that
writes it, or leave the key out and the preset takes what reaches it, the default or its base's
array. One value is no
shorthand for "the same for every variant"; write the value for every variant:
| Set | Length | Keys |
| --- | --- | --- |
| `printer_extruder_options` | extruders (`E`) | the 8 per-extruder keys outside the variant scheme (`extruder_type`, `nozzle_diameter`, `default_nozzle_volume_type`, …) |
| `printer_options_with_variant_1` | variant length (`S`) | the full list below; not guessable from names |
| `printer_options_with_variant_2` | 2 × variant length | the 16 `machine_max_*` limits, at [stride 2](#widths): a (normal, silent) pair per variant |
Put the variant layout on the shared base of all printers that share the hardware, and let the
nozzle-diameter siblings inherit it, restating only the variant arrays whose values change. `check`
does not judge a base on its own; its arrays count where they reach a selectable preset, under
`check --strict`. The loader stores a base at its **own**`printer_extruder_variant`, one variant
when it writes none whatever its extruder count, and cuts a wider array to its first values before
any child inherits it; so a multi-extruder base whose extruders need different values declares the
layout itself ([composition](#padding-truncation-and-composition)). When a variant array can move from
the presets to a base is in [shared-bases.md](shared-bases.md#variant-arrays-on-a-base).
## Process
1.`print_extruder_variant` + `print_extruder_id` list the (extruder id, variant string) pairs of the
process's variants. A variant is found by that pair, never by position: what is **required** is
that every pair a compatible printer (by list or condition) can select is present, in any order. A
missing pair reads the process's variant index 0 for that extruder, an extra pair is a variant
nothing selects, and neither is reported. The exception is a single-extruder printer whose
`extruder_variant_list` offers one variant string: no pair is matched there and variant index 0
is read, so a process shared with such a printer lists that printer's pair first. **Mirroring** the printer's `printer_extruder_variant` +
`printer_extruder_id` entry for entry is the convention; follow it, so the arrays compare by eye,
but a different order with every pair present is a nit, not a defect. A process shared by printers
whose pairs differ falls back to variant index 0 on the pairs it lacks; give each layout its own base.
2. Every key in `print_options_with_variant` ([the full list](#the-four-key-sets), which is not
"every speed") has exactly one value per variant, or is left out. `print_extruder_id` needs its
value per variant too: one value pads to extruder 1 everywhere. `check` holds a
`print_extruder_id` that reaches the preset, written or inherited, to one entry per variant on any
printer, without `--strict`, and warns when it is absent and a variant repeats.
3. Only add process variants if speeds or accelerations really differ per nozzle volume type or
extruder. Otherwise omit the variant keys: a one-variant process needs no list, because match and
no-match both read variant index 0, and that variant is copied to every extruder at slice time.
4. Put the lists on the process base for that printer layout, so leaves stay small.
## Filament
```json
"filament_extruder_variant":["Direct Drive Standard","Direct Drive High Flow"],
"filament_max_volumetric_speed":["21","29"],
"filament_flow_ratio":["0.98","0.98"],
"filament_retraction_length":["nil","0.4"]
```
1.`filament_extruder_variant` lists variants **without extruder ids**: a filament's High Flow variant
is used on whichever extruder is in High Flow. Its entries must be distinct, since the lookup
returns the first equal string and a repeated entry is a variant nothing selects. There is no
filament id key: `filament_extruder_id` exists only as a G-code placeholder, not as a filament
option (its option and its set entry are commented out in `PrintConfig.cpp`), so a filament file that writes it loses the key as unknown.
`filament_extruder_compatibility` is unrelated to variants (it says which extruders the filament
may be loaded into).
2. Every key in `filament_options_with_variant` ([the full list](#the-four-key-sets)) that reaches the
preset, whether written, included or inherited, is resized to its variant count: write it at exactly
that width, or leave it out. Keys outside the set
(`filament_type`, plate temperatures, `fan_max_speed`, `slow_down_min_speed`, …) are never addressed by variant index; the preset contributes their first value however wide a
file writes them.
3. Cover every variant the material is meant to print on across its `compatible_printers`. Leave out
a variant deliberately when the material should not be tuned for it (e.g. a TPU High Flow variant
only on TPU filaments); an extruder reporting that variant string then reads the first variant.
4. Order the variants like the printer's, Standard first, so the first-variant fallback is the
conservative one.
5. Tune what really differs: `filament_max_volumetric_speed` is the usual difference between nozzle
volume types, then flow ratio, temperature and retraction. Use measured values; never copy the
Standard value into the High Flow variant and call it tuned.
6.`nil` is legal per variant in the nullable override keys, occupies one entry like any value, and
keeps the printer's value for that variant only.
7. Declare a multi-variant list with the arrays it sizes: on the filament, in its own file or in a
template it pulls in with [`include`](vendor-bundle.md#inherits-and-include). The filament may be
compatible with printers that list fewer variants, or none: each printer takes the filament's variant
for its own variant string, else the filament's first variant. A one-variant list on a shared base
is harmless, because one variant is the width a list-less preset has anyway; and a list-less base is
stored at that width, so a wider array on it reaches its children as its first value only.
A key such a base writes reaches a multi-variant child as one value, which the loader spreads over
the child's variants; `check --strict` reports it at the child, which restates it at its own width.
## Widths
Every variant key is a flat array addressed by **variant index**: index `n` belongs to entry `n` of
the preset's own variant list (`printer_extruder_variant`, `print_extruder_variant` or
`filament_extruder_variant`). The index is found by exact compare of the string
`"<extruder_type> <nozzle volume type>"` (plus the 1-based extruder id for printer and process lists),
| `machine` | `len(printer_extruder_variant)`; without it, the variants `extruder_variant_list` offers; without both, **one per extruder** (`len(nozzle_diameter)`), since the list's default is one `Direct Drive Standard` per extruder. `extruders_count` is a printer-tab field, not a preset key. The loader honours that default only halfway for a system preset: it sizes the composed preset by the one-entry default `printer_extruder_variant`, cutting every variant array to its first value, and only then (with `single_extruder_multi_material` off) rebuilds the pair to one variant per extruder ([Printer rule 2](#printer-machine)) and pads the arrays with that value. The rule's width is still one per extruder; to give extruders different values, declare the layout. |
| `process` | `len(print_extruder_variant)`; without one, 1 |
| `filament` | `len(filament_extruder_variant)`; without one, 1 |
The per-extruder keys in `printer_extruder_options` ([listed with the sets](#the-four-key-sets)) are
outside this scheme and stay one value per extruder. Keys outside [the four sets](#the-four-key-sets)
are never variant-resized or addressed by variant index, however wide a shipped file writes them; a
per-extruder vector holds `len(nozzle_diameter)` values, and a shorter one acts as padded with its
first value ([machine-profiles.md](machine-profiles.md#multi-extruder-idex-and-tool-changers)).
### Sizing equation
For a key in one of the four variant sets:
```
E = extruders = len(nozzle_diameter) = len(extruder_type)
| `printer_extruder_variant` | `[Standard, Standard, High Flow]` | `[Standard, High Flow, Standard, High Flow]` (full strings in the file) |
| `retraction_length` (stride 1) | `["0.8", "1.0", "1.2"]` | `["0.8", "?", "1.0", "1.2"]`: one value at index 1 |
| `machine_max_speed_x` (stride 2) | `["500", "200", "600", "250", "700", "300"]` | `["500", "200", "?", "?", "600", "250", "700", "300"]`: a (normal, silent) pair at position 2 |
| `print_extruder_id` / `print_extruder_variant` / `outer_wall_speed` | mirror the printer | the same insertion at index 1 |
| `filament_extruder_variant``[Standard, High Flow]` | — | unchanged: filament variants carry no extruder id, so the existing High Flow variant now serves both extruders |
Every `?` is a measured value for that nozzle, on the machine limits as much as on retraction.
Removing or renaming a variant shifts the later values the same way in reverse; a variant string that
no longer matches the enum is simply a variant nothing selects.
### Padding, truncation and composition
**Every key in the set widens, whether or not a file restates it.** At load each variant key of the
composed config is resized to the length of the preset's own `*_extruder_variant` (the one-entry
default when it writes none) × stride: a short array is **padded by repeating its first value**, a
long one is truncated to its first values, without a word from the loader or the validator. The rule
does not follow from this padding: a file writes a variant key at **exactly `variant length × stride`**
or not at all. Any other length, one value included (which the loader spreads over every variant, at
stride 2 over normal *and* silent alike), is a mistake the loader hides and
`orca_profile_tool.py check` reports as an error in the selectable preset that writes it, even when
every value is the same; `check --strict` also reports an array that reaches a selectable preset at
another width
([variant arrays](validation.md#variant-arrays)).
The loader sizes a list-less preset of any type to one variant at this step, a list-less machine
included: a two-extruder machine without a layout that writes `retraction_length``["0.8", "0.9"]`, the
width the equation asks for, stores `["0.8"]`, which the pair rebuild and the slice-time collapse hand
to both extruders. A 3-value array on a preset of variant length 2 keeps its first two; at length 4 it
becomes `[a, b, c, a]`.
Composition hands down widths in two ways:
- **`inherits` hands down the parent's resized arrays.** A base is stored after its own resize, at
the length of its own `*_extruder_variant` (one variant for a base of any type that writes none,
whatever its extruder count), so an array wider than the base's list is cut to its first values
before any child sees it, and a child that adds variants gets those first values padded. Widen an
array only on a preset whose own resolved list already has the entries. `check` judges selectable
presets only, at the width each file wrote: an array a base's own resize cuts is not seen (a review
item), and an inherited array of another width than a child's list is left to the loader's resize
unless `check --strict`, which asks the child to restate it at its own width.
- **`include` hands down the template's diff at its pre-resize width.** The template contributes every
key where its composed config differs from the built-in defaults, taken before its own resize, so the
arrays it writes arrive at the width its file wrote, and the includer's own list sizes them. A key the
template sets to the built-in default is not passed on
`OrcaFilamentLibrary` is the filament-only bundle the loader reads **first**, so any vendor's filament may
inherit a library preset by name. It is the only cross-bundle parent: vendor-to-vendor inheritance
always fails.
## Where a filament goes
| Contribution | Location |
| --- | --- |
| Generic material for all printers | `OrcaFilamentLibrary/filament/Generic <mat> @System.json` |
| A brand's product, all printers | `OrcaFilamentLibrary/filament/<Brand>/` |
| A brand's tune for one printer vendor | `OrcaFilamentLibrary/filament/<Brand>/<PrinterVendor>/` (recommended); `<PrinterVendor>/filament/<Brand>/` also works |
| A printer vendor's tune of a generic, or its own product | `<PrinterVendor>/filament/` |
Both locations in the third row are supported: `OrcaFilamentLibrary/filament/<Brand>/<PrinterVendor>/<Name>.json`
(the shape the wiki shows) and `<PrinterVendor>/filament/<Brand>/`. The library path is the one a
filament brand should contribute to: `OrcaFilamentLibrary/filament/<Brand>/` is the brand's own folder,
while a printer vendor's folder belongs to that printer vendor.
Library layout: `filament/base/fdm_filament_*.json` material roots, root-level
`Generic <mat> @System.json` generics, and one subfolder per brand, which may nest printer-specific
tunes one level deeper. Adding a brand means adding a folder here; the folder name is a directory label
only, and `filament_vendor` inside the JSON is the real vendor string.
## The three-part shape
```jsonc
// OrcaFilamentLibrary/filament/Polymaker/Fiberon PA6-CF @base.json — the product root: identity + material values
`docs/HLSD/filament_id.md` is the authoritative design document for `filament_id`: the id landscape,
the checks CI runs, and the Bambu catalog map. This page is the tooling half.
| | `setting_id` | `filament_id` |
| --- | --- | --- |
| Identifies | one selectable preset | one filament **product** |
| Key hashed | `<vendor folder>/<type>/<name>` | `filament_product/<filament_vendor>/<filament_type>/<alias>`, using the resolved (inherited) first values and the name up to the first `@`, right-trimmed |
| `generate-id` | writes `setting_id` and `filament_id` |
| `normalize` | rewrites profile files into their canonical shape |
| `trim` | deletes profile files no `<Vendor>.json` list references |
| `update-index` | rebuilds the `*_list` sections from the files on disk |
The order after adding, renaming or deleting files is `normalize` → `update-index` → `generate-id` →
`check`. Each step feeds the next, so it is not interchangeable. The
[authoring workflow](../SKILL.md#creating-or-modifying-a-profile) has the commands.
> **`trim` deletes.** It removes every profile file the index does not list, including the one you just
> added and have not registered yet. Register first, or skip `trim` entirely: it is a cleanup sweep, not
> part of landing a profile. Preview with `--dry-run`.
**Register, then mint.** The `filament_id` pass reads `<Vendor>.json`'s `filament_list`, not the
filesystem (the `setting_id` pass walks the filesystem, so a bundle whose index has not landed yet is
still assignable). A new filament file is therefore invisible to `generate-id`'s `filament_id` pass
until it is registered; its `setting_id` is written regardless.
-`--dry-run` works on every writing command (`generate-id`, `normalize`, `trim`, `update-index`) and
writes nothing.
-`--filament-id` / `--setting-id` narrow `generate-id` to one pass; they exclude each other, and
passing neither writes both.
-`--vendor` is repeatable and narrows **only what is written**: an id is a function of its own key
alone, so a narrowed run writes exactly what a full run would. An unknown vendor exits 1 before any
write. `--vendor` on `check` narrows the per-vendor checks only; the `setting_id` and `filament_id`
passes stay tree-wide.
-`--profiles DIR` points any command at another tree; see
[Checking a copy of the tree](validation.md#checking-a-copy-of-the-tree).
-`--profile-type` narrows `normalize`, `trim` and `update-index` to `machine_model`, `process`,
`filament` or `machine`.
- Exit codes: 0 clean, 1 errors found (`generate-id` still writes what it could), 2 argparse misuse.
- Output is ANSI-coloured; searching for the literal `[ERROR]` still works.
`generate-id` is **idempotent and byte-preserving**: BOM and CRLF are kept, and each pass touches only
its one key line. A legitimate `generate-id` diff is one or two changed lines per file: a new instantiated
filament gets both a `filament_id` and a `setting_id`. `normalize` is the opposite by design (it
rewrites whole files into canonical shape), which is why `check` demands it already be a no-op. A file
committed with CRLF line endings changes on every line under `normalize`; read the diff before
committing it.
Exit 1 from `generate-id` does not mean nothing was written: it writes every id it can and reports the
rest, so read the diff before rerunning. On a clean tree `check` and `generate-id --dry-run` both exit 0
with zero findings; that is the baseline to restore before opening a PR.
## What `generate-id` does and does not fix
Writes:
- a `setting_id` into any instantiated preset that lacks one, or whose value does not match the formula;
- strips a `setting_id` from a base;
- deletes the misspelled key `settings_id`, moving its value to `setting_id` only on an instantiated BBL
preset that lacks one (everywhere else the old value is discarded and a fresh id minted);
- a `filament_id` into the id-less **root(s)** of an instantiated filament that resolves none;
- rewrites a **declared**`filament_id` that is not the mint of its own triple.
Refuses to write (reports only): a base62 collision between two products, an empty `filament_vendor` or
`filament_type`, a broken `inherits` chain, and roots of one filament resolving different
`(filament_vendor, filament_type)` pairs.
**Does not fix: a preset that *inherits* a wrong `filament_id`.** This is check 2b (the label `docs/HLSD/filament_id.md` and the tool use), and it is the trap most likely to bite.
It happens when a branded filament inherits a generic for its settings:
**Everything else is silently dropped**, a printer config key such as `default_bed_type` or a
misspelling included, so a neighbour carrying a key is no evidence it does anything. Printer config options belong on the `machine` preset, never here. The loader drops a model
silently if its index entry has no name or its `nozzle_diameter` yields no sizes; `check` also requires
| identity | **the `name` of the `machine_model_list` entry**, which is what a variant's `printer_model` must equal. `check` forces it to equal the file's `name`, so they coincide. |
| `model_id` | a *separate* cloud/device printer type. Optional, and not required to be unique. Not the model's identity; changing it changes device matching. |
| `machine_tech` | only a value starting with `SL` means SLA; everything else is FFF. Write `FFF`; `FGF` behaves as FFF. |
| `nozzle_diameter` | `;`-separated string, one token per available size. Order is free (`0.4;0.2;0.6;0.8` puts the default first). This list is the authoritative set of legal `printer_variant` values. |
| `default_materials` | `;`-separated filament **preset names**, not `,`; case-sensitive (`@System`); order is ignored. Used to preselect filaments in the setup wizard *and* to install a printer's filaments on first run, so a dangling entry costs a real user a filament. Every name must exist (`check` fails on a name matching no filament file, here or in `default_filament_profile`), and every variant of the model needs at least one entry compatible with it (`validate_system`). |
| `family` | a wizard grouping label only; give every model one. |
### Assets
`bed_model`, `bed_texture` and `hotend_model` are paths relative to the **vendor folder** (named by the
vendor id). Convention: `<Model>_buildplate_model.stl` and `<Model>_buildplate_texture.svg`. An
empty string is the legal "none", and is the norm for `hotend_model`.
**Nothing checks that the files exist.** A missing `hotend_model` falls back to
`resources/profiles/hotend.stl`; a missing `bed_model` makes the bed render as a generic custom bed, and a missing
`bed_texture` renders no texture.
Verify by hand, in exact case.
Every model also has a `<Model>_cover.png` in the vendor folder; treat it as required, not optional.
240×240 is the cap `scripts/optimize_cover_images.py` enforces. A missing cover degrades to a placeholder in both the wizard and the sidebar.
## `machine`: the variant
```json
{
"type":"machine",
"name":"Phrozen Arco 0.4 nozzle",
"inherits":"fdm_machine_common",
"from":"system",
"setting_id":"lvaYKTUZr5C9jSwk",
"instantiation":"true",
"printer_model":"Phrozen Arco",
"printer_variant":"0.4",
"nozzle_diameter":["0.4"],
"default_print_profile":"0.20mm Standard @Phrozen Arco 0.4 nozzle",
`max_layer_height`, …, given in full with the [key sets](extruder-variants.md#the-four-key-sets), plus
`extruder_offset` and `extruder_colour`) hold one value per extruder. The variant sets
(`retraction_length`, `z_hop`, `wipe`, `nozzle_type`, the `machine_max_*` limits at stride 2;
[the full lists](extruder-variants.md#the-four-key-sets)) are sized to the variant length:
`len(printer_extruder_variant)`, or one variant per extruder when the resolved preset writes no layout,
since `extruder_variant_list` defaults to one `Direct Drive Standard` per extruder
([widths](extruder-variants.md#widths)). `extruders_count` is a printer-tab field, not a preset key; the
loader drops it.
- Give **one entry per extruder** for ordinary per-extruder vectors such as `extruder_offset`,
`extruder_colour`, `min_layer_height` and `max_layer_height`. Size the variant sets to the variant
length × stride, one value per variant (per extruder when there is no layout; a (normal, silent)
pair for the `machine_max_*` limits), even where the values are the same; the loader keeps only the first value of a list-less printer's variant arrays, so declare
the layout when the extruders differ. A single `["0x0"]``extruder_offset` on a dual or
multi-extruder machine pads every extruder to the same offset, so the offset never applies.
- Overriding `nozzle_diameter` to a different count without restating every per-extruder vector is the
other half of the trap: a 4-extruder preset on a 5-extruder base inherits 5-entry vectors against 4
extruders.
The reverse is silent too: a base is stored resized to its own `printer_extruder_variant` (one variant
when it writes none, whatever its extruder count), so a wider variant array on it reaches the children
as its first value padded ([composition](extruder-variants.md#padding-truncation-and-composition)).
`check` does not judge the base on its own; where its extruders need different values, declare the
layout on the base.
Structure to copy: `Custom/machine/fdm_toolchanger_common.json` + `Custom/machine/MyToolChanger 0.4
nozzle.json` (a minimal variant on a base that gives the per-extruder vectors five entries), and
`Ratrig/machine/RatRig V-Core 4 IDEX 300 0.4 nozzle.json` for IDEX. Take the structure from them and
the widths from the [sizing equation](extruder-variants.md#sizing-equation). Both are list-less, so
every variant array holds one value per extruder and the loader keeps only the first: fine while every
extruder shares the same retraction and limits. Add the extruder-variant layout
| `machine` | `fdm_machine_common`, `fdm_<vendor>_common`, `fdm_klipper_common`, or an established machine-family base |
| `process` | `fdm_process_*`: shared roots and per-layer-height or per-nozzle bases such as `fdm_process_single_0.20` or `fdm_process_<vendor>_<lh>_nozzle_<n>` |
Run `./scripts/check_profile.sh` on the applied diff first ([validation.md](validation.md) says what CI
runs), then work through the items below: delivery, identity and backward compatibility first, then the
affected preset types. The table lists the gaps CI cannot see, so only a reviewer catches them.
| Not checked by CI | Consequence |
| --- | --- |
| The `version` bump | The change never reaches an upgrading user; an absent `version` hides the vendor from the setup wizard |
| A misspelled setting key | The setting silently has no effect |
| A filename Windows cannot check out, or one that differs from its `sub_path` only in case | Works on the author's machine, breaks the bundle on another platform |
| `bed_model` / `bed_texture` / `hotend_model` / cover pointing at a missing asset | A missing bed model renders a generic custom bed and a missing texture renders none, the hotend falls back to the generic model, the cover shows a placeholder |
| A nozzle size in a model's list with no matching variant | The size is offered and resolves to nothing |
| A non-default process | `validate_slice` gives non-default quality tiers no dedicated coverage |
| Whether the intended default survived compatibility selection | The sweep can select a different compatible preset |
| A dangling `compatible_printers` inside a base whose children all override it (or that has no instantiated children) | The reference check walks resolved selectable presets, so it reports a base's list only through a child that inherits it unchanged (a bad `inherits` in a base *is* caught) |
| A base nothing inherits | Dead weight, usually the leftover of an unfinished nozzle addition |
| A `renamed_from` whose old name is still a live preset | The redirect is inert while a live preset carries that name |
| A preset differentiated only by colour, or an all-printer library preset without `@System` | Per-colour presets split one product across several ids and the selector fills with near-duplicates; CI stays green |
| Plate temperatures for plates the printer has | The user's plate reads an unset or inherited temperature |
| Per-extruder vector length on a multi-extruder printer | Silently padded (with the **first** value) or truncated |
| A new variant appended instead of inserted at its variant index, a variant array widened on a base with a shorter list, or a variant a filament/process lacks | Values shift onto the wrong extruder, are cut before any child inherits them, or resolve to the first variant: High Flow silently gets Standard values |
| A name that ignores its type's convention: a printer model in a `process` quality position, or an unrelated target label or base name left in a copied preset | The selector misrepresents the preset's quality or intended printer |
| Values: temperatures, speeds, widths, pressure advance | A wrong value prints wrong while CI stays green |
## 1. Was the vendor `version` bumped?
For **every** bundle whose folder the diff touches, `resources/profiles/<Vendor>.json` must have its
`version` incremented: last component, carrying `.99` into the third component. A library change means
bumping `OrcaFilamentLibrary.json`.
*Why:* nothing in CI checks it, and the app reinstalls a bundled profile set only when its version is
newer than the installed one: without a bump the change reaches neither an upgrading user nor the
author's own running app. Without any `version` the vendor vanishes from the setup wizard, and neither
`check` nor the validator reports it.
## 2. Was the index rebuilt, and does the diff contain only this change?
`check` fails on an unregistered file, on an index `update-index` would reorder, and on a file
`normalize` would rewrite, so a PR that skipped them arrives red and you do not have to spot the
omission yourself. Three things are still yours:
- **The index diff belongs to this change.** `update-index` rewrites whole `*_list` sections. If the
bundle had drifted, the author's PR now carries someone else's reordering; ask for it in a separate
commit rather than reviewing it inline.
- **A deleted selectable preset needs a successor** as in item 4. `update-index` removes its
registration; `validate_custom` detects the break only for names covered by released fixtures.
- **`normalize` edits content, not just layout.** It drops `version` and `is_custom_defined` from preset
files, removes obsolete keys, deletes six print-speed keys from filament profiles, and resolves
`extruder_clearance_radius` against `extruder_clearance_max_radius` by keeping the larger
([what normalize changes](validation.md#normalize-and-update-index-are-part-of-the-check)). Check that
the keys it removed were meant to go.
Index order is dependency order, not alphabetical: parents and include templates before the presets
that use them, then, in name order, the entries that neither depend on nor are depended on by another
entry of their own list (such as a leaf filament whose only parent is in the library), and any entry on
a dependency cycle. Judge a hand-placed entry only by
`update-index --dry-run`: if it reports nothing to rebuild, the position is not a finding.
*Why:* the index is the loader's only entry point. Out-of-order entries fail with `can not find inherits`
and take the whole vendor bundle down; an unindexed file gets reviewed, merged and never loads.
## 3. Are ids generated, not written?
No hand-typed or copied `setting_id` / `filament_id`. Instantiated presets have a `setting_id`; bases do
not. `check` enforces all of that; what it cannot tell you is whether the identity *should* have moved.
A rewritten or removed `filament_id` means a product's identity moved (a renamed alias, or an edited
`filament_vendor` / `filament_type`), and the old id is not forwarded anywhere. Confirm that was
intended, and that a new id is not a rename in disguise.
*Why:* a duplicate `filament_id` on one printer makes AMS spool matching a coin toss; a copied
`setting_id` breaks preset identity. See [ids.md](ids.md).
## 4. Does anything disappear for existing users?
A rename, a deletion, or a flip of `"instantiation": "true"` → `"false"` on a shipped selectable preset
removes the name from the preset collection. It needs `renamed_from` on a selectable successor
([renamed_from](vendor-bundle.md#renamed_from)), and only one preset may claim a given old name. The
claimed old name must **not** still be a live preset; the redirect is inert if it is.
*Why:* user presets inheriting it die with `can not find parent <name> for config <file>!`; 3MF-embedded
presets are dropped with no error at all. Commit `33923464ae` reverted exactly this for Cubicon;
`6943b6ddc3` redid it correctly with `renamed_from`. CI's `validate_custom` catches the shipped-name
case, but not an inert `renamed_from`.
## 5. Is `compatible_printers` right?
Exact printer **variant** names, non-empty on every instantiated filament outside OrcaFilamentLibrary,
and written in the preset's own file ([SKILL.md rule 9](../SKILL.md#rules); the resolved-vs-own-key trap
is in [filament-profiles.md](filament-profiles.md#compatible_printers)). A `machine_model` name instead
of a variant name is the usual mistake: `check` passes it, the validator reports
`references unknown compatible_printers`. Watch for a nozzle-specific tune that inherited or copied the
base's full printer list, and for two presets of one product with overlapping lists: duplicate combobox
entries and an ambiguous AMS match.
Two presets of one product (`filament_id`) must not share a variant. Resolve it by specificity: move the
variant to the most specific preset and remove it from the more general ones, which is preferred over
deleting a profile. Then repoint the machine's `default_filament_profile` and the model's
`default_materials` at the profile that now covers it. See
[one variant, one profile](filament-profiles.md#overlapping-coverage-one-variant-one-profile-per-product).
*Why:* real shipped bugs twice (`b7b3418baf` filaments "showing up everywhere", `ff83aa41ef` duplicate
Flashforge entries). The Python `check` passes on an overlap; only the full `check_profile.sh`
`check` reports a variant array of any type that is not exactly the `variant length × stride` width
of the selectable preset that writes it, one value included, as an error whatever the values;
`check --strict` also reports one that reaches a selectable preset at another width. A base is not
judged on its own. On a machine it
also reports layout keys that disagree: `printer_extruder_variant` / `printer_extruder_id` not the
flattening of `extruder_variant_list`, a variant without its extruder's `extruder_type` prefix, a
`default_nozzle_volume_type` the extruder does not list, a list of several variants without the pair, a
pair without the list that puts several variants on one extruder. On a machine or a process, an id
array, written or inherited, whose length differs from its variant list is an error; the id and
layout rules judge the composed preset without `--strict`. Two layout findings are warnings: a
pair without the list that `single_extruder_multi_material` off would replace at load, and a process
without `print_extruder_id` whose variants repeat. The full rules are in
[variant arrays](validation.md#variant-arrays). `check_variant_names` separately holds every variant
string, `extruder_type`, `nozzle_volume_type` and `default_nozzle_volume_type` to the engine's enums
in every bundle, BBL included, failing on a dead variant, a legacy spelling and a variant list that
names one variant twice ([variant names](validation.md#variant-names)).
On a multi-variant printer, still check by hand:
- that a new variant's values sit at its variant index (machine limits as a (normal, silent) pair at
`2 × index`) in every file of the chain that restates the key, `include` templates included;
- that each process lists every (extruder id, variant) pair its printers can select;
- that no array was widened on a base with a shorter variant list: the loader cuts it before the
children inherit it ([composition](extruder-variants.md#padding-truncation-and-composition)), and
`check`, even `--strict`, composes without that cut, so it passes;
- that only keys in [the four sets](extruder-variants.md#the-four-key-sets) carry per-variant values;
- that the High Flow variants carry measured values rather than copies
([checking and testing](extruder-variants.md#checking-and-testing)).
## 15. Non-default processes get no slice coverage
`validate_slice` starts from printer defaults; it does not enumerate every process. Slice a new or
changed non-default tier explicitly with its intended printer
([on a copy of the tree](validation.md#checking-a-copy-of-the-tree)).
## 16. Do the preset names follow the conventions?
Check **every newly added profile and intentional name change**, including models and bases, against
[the naming conventions](naming.md#checking-names): no printer model in a process quality position, no
target label or base name left over from a copied preset, the bundle's established style. Preserve
shipped names during ordinary tuning; renaming a shipped selectable preset requires the migration in
item 4.
*Why:* CI checks name uniqueness, but does not enforce the naming conventions. Catch naming mistakes
before the names ship and existing projects depend on them.
## 17. Cross-platform filenames and paths (not checked)
Check for Windows-invalid characters, reserved device names, trailing path-component spaces or dots,
and case mismatches between `sub_path` or asset references and the files on disk. See
[filenames and paths](naming.md#filenames-and-paths).
## 18. Housekeeping worth a nit, not a block
`"from"` other than `"system"` (the bundle loader ignores it, though loading the file as a CLI config
rejects anything but `system` / `user` / `User`), `printer_settings_id` copied from another vendor,
redundant overrides that restate the parent's value, siblings that each repeat a value their base could
hold ([shared bases](shared-bases.md)), and a filename that disagrees with the preset's
`name` (the loader keys off `name`).
---
## Reporting the review
A finding is **one defect**: its file (or quoted lines), what breaks at runtime or in CI, and the fix.
Split independent defects into separate findings even when they live in one file: five id problems in
one bullet get one fix and four survivors. Say which findings `check` or the validator reports and which
only a reader catches: a missing version bump, a misspelled key and wrong temperatures pass CI, so a
contributor who only reruns the tools fixes what CI flags and resubmits the rest.
Severity discriminates only if it is earned:
| Severity | Means |
| --- | --- |
| blocker | the bundle fails to load, or a preset is unreachable at runtime |
| major | CI fails, existing users lose a preset, or a value prints wrong while CI stays green (PLA temperatures under an ASA name; a misspelled key whose intended value differs from the inherited one) |
| minor | wrong but working: redundant or dead keys that change nothing, `from`, naming |
Compute every number and id (`orca_profile_tool.py`, a scripted count) or omit it: one invented count
makes a reader stop trusting the right ones. Report a command's result only if you ran it. End with a
`snapshot` records every selectable preset of every bundle as the loader stores it: the parent's stored
config, each `include` at the width its file wrote, the preset's own keys, then every variant array
resized to the preset's own variant list ([composition](vendor-bundle.md#inherits-and-include)). It
models the step `check` does not see: a base is stored after its own resize, so an array wider than a
base's list reaches its children cut. It also leaves out what the loader reads from each file and never
hands down (`name`, `type`, `from`, `instantiation`, `inherits`, `include`, `setting_id`,
`renamed_from`, `description`, `version`, `url`, `is_custom_defined`), so those keys never move to a
base. Nor does it record `print_settings_id`, `printer_settings_id` or `filament_settings_id`: the app
replaces them with the selected presets' names before slicing, so a file's value never counts. Delete
them from the presets you restructure rather than moving them to a base.
It does not know the built-in defaults, so `compare` lists a key written on one side only (no file of
the preset's chain writes it on the other) apart from the differences, and exits 1 for either. A
difference is a value the edit changed: undo it, or make it a separate, deliberate change. A one-sided
key is no change only when its written value is the option's default in `PrintConfig.cpp`: confirm
each, as when a preset that loaded the default by leaving a key out must now write it
([Balance 5](#balance)). A value equal to the default needs writing nowhere when no base above writes
another: delete it from the presets rather than moving it, and confirm the one-sided keys.
## Levels
A base stands for a level of the vendor's catalogue, and a setting lives at the level that determines
it. The test for a shared value: if it had to change for one preset of the group, should it change for
all of them? If yes, it is the group's and goes on the group's base. A value that is only equal today
(two unrelated printers with the same acceleration) stays in each preset: a base built on coincidence
is later edited for one preset and silently changes the others. For a default with exceptions
([Balance 5](#balance)), ask the question of the presets that inherit the default.
The tables give each setting's usual level; the test decides for a given bundle: where each toolhead
(Bowden or Direct Drive) has its own default filament, `default_filament_profile` follows the
toolhead, not the nozzle. Levels run
coarse to fine, and a chain need not visit every level: each preset or base inherits the next coarser
level that has a base.
| Machine level | Base | Settings it determines |
| --- | --- | --- |
| Vendor | `fdm_machine_common`, `fdm_<vendor>_common` | the vendor's defaults for every printer |
| Firmware | `fdm_klipper_common`, `fdm_marlin_common` | `gcode_flavor`, G-code that calls the firmware's macros (layer change, pause, filament change), `host_type`, `print_host`, the thumbnail format |
| Hardware family: models that share a frame, motion system, toolhead or extruder layout | `fdm_<vendor>_<family>_common` (`fdm_qidi_x3_common`, `fdm_machine_eryone_ER20_common`) | the `machine_max_*` limits, `extruder_clearance_*`, the toolhead's retraction where every nozzle shares it, `z_hop` and wipe, fitted hardware such as `auxiliary_fan`, the extruder count and per-extruder vectors, the variant layout ([Printer rule 2](extruder-variants.md#printer-machine)) |
| Model: one `machine_model`, which for an IDEX printer includes its mode | the default-nozzle preset where the bundle hangs its other nozzle presets off it; otherwise `fdm_<vendor>_<model>_common` | `printable_area`, `printable_height`, `bed_exclude_area`, the model's start G-code, a COPY or MIRROR mode's settings |
| Nozzle: the selectable preset | none | `printer_model` and `printer_variant`, which name the preset's model and nozzle and stay in every preset as the tree writes them; `nozzle_diameter`, `min_layer_height`, `max_layer_height`, `default_print_profile`, `default_filament_profile`, retraction the vendor tunes per nozzle |
A family may split once more, into toolhead or revision groups that exist only within it:
`fdm_<vendor>_<family>_common` → `fdm_<vendor>_<family>_mk1_common`. A split that crosses another axis is not a
level: when every controller comes with every toolhead, a toolhead base under each controller base
repeats the same values in each ([Balance 6](#balance)).
| Process level | Base | Settings it determines |
| --- | --- | --- |
| Vendor | `fdm_process_common`, `fdm_process_<vendor>_common` | strategy: seam, wall order, infill and support patterns |
| Printer family or variant layout | `fdm_process_<vendor>_<family>_common` (`fdm_process_arena_common`, BBL's `fdm_process_dual_common`) | speeds, accelerations and jerk of that motion system; the variant layout ([Process rule 4](extruder-variants.md#process)) |
| Layer height × nozzle | `fdm_process_<vendor>_<lh>_nozzle_<n>` (BBL's `fdm_process_single_0.20`) | `layer_height`, line widths, shell layers, speeds scaled to the layer |
| Quality × printer: the selectable preset | none | `compatible_printers`, and what is unique to that combination |
| Filament level | Base | Settings it determines |
| --- | --- | --- |
| Material | `fdm_filament_<material>` in OrcaFilamentLibrary, shared by every bundle | material defaults |
| Product | `<Product> @base` | `filament_id` (minted here, [ids](ids.md)), `filament_vendor`, `filament_type`, density, cost, the product's temperatures and cooling |
| Printer or nozzle tune: the selectable preset | none | `compatible_printers` (always in its own file, [Rule 9](../SKILL.md#rules)), volumetric speed, flow ratio, pressure advance and retraction measured on that printer |
**Equal where they must differ is a copy.** A key that follows a finer level, such as the layer-height
limits and line widths that follow the nozzle diameter or `printable_area` that follows the bed, never
moves above that level. When presets that differ in it carry the same value, the value was copied: leave
it in the presets and report it (the [worked example](#worked-example-an-idex-family) has one).
## Balance
1.**One group, one base; use the existing one first.** A base is the home of the presets below it,
whatever its name. In a bundle whose presets are all one family, the vendor base is the family base,
so the family's values go there. In a bundle that hangs the other nozzles off the default-nozzle preset, that preset is the model's home.
Never create a base whose presets are exactly its parent's. Where two existing bases already serve
the same presets (a copied `fdm_machine_common` above the vendor's own base), the finer one is the
home, and merging the pair is a change of its own. Moving a key into an existing home adds no file.
A base value that no preset below it loads is dead: replace it with the group's value when there is
one; otherwise leave it and report it, since a future preset would inherit it.
2.**A new base must stand for a level and pay for itself.** Its file costs five metadata keys and an
index entry, so create it only when it takes `k` keys off `n` selectable presets with
`(n − 1) × k > 6`, where a custom G-code value counts as one key per G-code line: what it removes
must outnumber what it adds. A model with two nozzles that share three short keys keeps them in both.
3.**No ad-hoc bases:** never a base for presets that merely agree (the ones with 0.8 mm retraction),
and never a base with one preset below it.
4.**Keep every selectable preset within four ancestors**, selectable parents included. When a level would push a preset past that, fold it into the level above or leave
its keys in the presets.
5.**A default with exceptions.** A value that only some presets below a shared base load moves to that
base, whichever axis it follows, when two conditions hold. More presets load it than any other value
(on a tie the key stays in the presets), and `w − a > 1`, where `w` presets drop their copy and `a`
presets that take the key from the base and load another value, the built-in default included, must
now write theirs. Presets that write another value keep it and are unaffected. The base then holds
the group's default, and the exceptions stay visible in their own files.
6.**One chain; the other axes stay in the presets.**`inherits` follows one axis. When presets vary
along several (bed size × controller × toolhead), first fill the existing homes by
rules 1 and 5. Then give new bases to the axis whose bases pay most: sum rule 2's count over its
bases, less the keys a re-parented preset must now write because it no longer inherits them from its
old parent; on a tie, follow the layering the bundle already has. Leave the other axes' keys in the
presets, and never repeat one axis's bases under each group of another. Outside BBL, whose synced
presets need theirs, add no `include` template for a second axis: the loader reads `include` only
since #15869 (2026-09-25), and an app that predates it ignores the key, so the template's settings
never reach the preset.
Name a new base after its level ([base names](naming.md#bases)). The name must be unique in its bundle
and must not equal a selectable preset's: two such presets load silently and the first in the index wins
([uniqueness](naming.md#uniqueness)).
## Restated values
`candidates` lists every key a file writes that it would load unchanged without writing it. Delete it
when the value is what the presets below the base that supplies it share: more of them load it,
written or inherited, than any other value. A family that restates machine limits, clearances and
G-code every other printer of the bundle loads from `fdm_klipper_common` drops its copies. When most
presets below that base load another value, the match is a coincidence: keep the key, and move it to
the level of the presets that share it. Keys of the nozzle level stay in the preset in
either case.
After a restructure the report still lists keys that are right where they are: the nozzle-level keys
each preset keeps, and arrays a list-less multi-extruder base writes at its presets' width for
`check --strict`.
## Variant arrays on a base
- The loader stores a base with its variant arrays resized to the base's own variant list, one variant
when it writes none, so an array on a narrower base reaches its presets as its first value, padded.
Move a variant array to a base only when the base declares the presets' variant list (and, for a
machine or process, their ids), moving the layout keys with it as
[Printer rule 2](extruder-variants.md#printer-machine) and [Process rule 4](extruder-variants.md#process)
ask; or when the presets load its first value anyway: every value is equal, or the presets are
list-less machines, which the loader cuts to one variant. `compare` catches a cut.
- Write the array on the base at the width of the presets it serves, their `N`
([sizing equation](extruder-variants.md#sizing-equation)), so `check --strict` judges the right width
where it reaches them. When the presets below a base need different widths (single- and
dual-extruder models on one base), leave the array in the presets, or on bases that each serve one
width.
- On a list-less multi-extruder family base, declare the extruder count: `nozzle_diameter` and the other
per-extruder vectors at one entry per extruder. The base then has its presets' width, and
`fix-variant --strict` writes an array that reaches the presets at another width into the base once,
instead of into every preset.
- Deleting a restated variant array leaves the preset on the inherited array. Plain `check` accepts
that; `check --strict` reports it when the inherited width differs, which is why a bundle held to
`--strict` restates the array at each preset's width.
## Procedure
1.**Snapshot** the tree before any edit, `fix-variant` included. After `fix-variant`, run `compare`: a
difference is a value its padding changed (it repeats the last value, the loader the first). Set that
value deliberately, then snapshot again as the baseline for the restructure.
| `validate_custom` | `validator -p <tree + fixture> -l 2` | a shipped preset name that a past release offered no longer resolving |
**`-f` is a no-op.** It defaults to on, so the duplicate-`filament_id` check runs whether or not you
pass it, and `validate_system` already fails on duplicates. The binary's own `--help` ("Off unless this
flag is present") does not reflect that default.
**A `--vendor` run reads differently from CI.** The validator's `-v` loads that vendor plus
OrcaFilamentLibrary and nothing else, so every library tune whose `compatible_printers` names another
vendor's printers fails `validate_system`, `validate_filament_subtypes` and each `validate_custom`
fixture with thousands of `references unknown compatible_printers "Bambu Lab …"` lines. Under
`--vendor`, `profile_tool` and `validate_slice` are the meaningful results; for the other three, filter
the log for your vendor's files and treat only those lines as findings. The unscoped run is the CI
result; run it before the PR.
### `validate_custom`: the backward-compatibility gate
It downloads one fixture archive per past release (v1.9.0 onwards) of *generated mock* user presets: a
`<vendor>_<preset>_orca_test` copy of every system preset that release shipped, cut with the
validator's own `-g 1` mode. It unpacks each over a copy of the current tree and loads it. Each entry
holds only `inherits` plus a canned diff, so the one failure it adds over `validate_system` is a shipped
preset name disappearing. This is what makes a rename, a deletion or an `instantiation` flip a CI
failure rather than just a user complaint, and the reason `renamed_from` is mandatory.
The whole current tree sits under each fixture, so every `validate_system` error fails
`validate_custom` too: fix `validate_system` first. Under `--vendor` it copies only the top-level index
files, `<Vendor>/` and `OrcaFilamentLibrary/`, and picks fixtures by the index's display `name`
(`Bambulab` for `BBL`), not the file stem; fixture presets without that prefix are covered only by an
unscoped run, and it warns `validate_custom checked nothing` when none match.
### `validate_slice`
It slices a two-colour cube on every instantiable printer in the tree, sequentially, forcing the prime
tower. It selects `default_print_profile` and the first `default_filament_profile`, then updates
compatibility; that update can select a different compatible preset. Confirm the intended defaults
yourself rather than treating a passing sweep as proof that those exact presets were sliced.
A printer fails if it cannot be selected, falls back to a Default preset, throws, produces no G-code,
or emits no `CP TOOLCHANGE START` (`change_filament_gcode` never expanded). Non-default processes and
filaments get no dedicated coverage; [slice them on a copy](#checking-a-copy-of-the-tree). In the
default set, a bundle without a `machine/` folder is recorded as SKIP; naming `validate_slice`
explicitly for it fails (`No instantiable printer presets found for vendor OrcaFilamentLibrary`). The
validator logs `[error]` lines that do not fail a check (such as `could not found extruder_type`); only
each check's PASS or FAIL counts.
## `orca_profile_tool.py check`
`check` is one subcommand of the tool that also owns `fix-variant`, `generate-id`, `normalize`,
`trim` and `update-index`; [ids.md](ids.md#the-tool) has the writing half.
| Catches | Scope | Function in the tool |
| --- | --- | --- |
| two files in one bundle claiming one type + name, indexed or not | per vendor | `check_preset_name_uniqueness` |
| a file on disk that no `*_list` references (**an error, not a warning**) | per vendor | `check_index_coverage` |
| an index entry whose `name` disagrees with the file, or whose `sub_path` is missing | per vendor | `check_name_consistency` |
| a file `normalize` would rewrite, an index `update-index` would rebuild | per vendor | `check_normalized` |
| duplicate JSON keys in a file | every file read | the JSON loader |
| an instantiated non-library filament with no non-empty `compatible_printers` of its own | per vendor | `check_filament_compatible_printers` |
| `extruder_clearance_radius` alongside `extruder_clearance_max_radius` | per vendor | `check_conflict_keys` |
| a scalar `filament_type` (`"filament_type": "PLA"`); the five other filament vectors `normalize` arrayifies surface as `normalize would convert <field> to an array` | per vendor | `check_vector_type_keys`, `check_normalized` |
| a variant string the two enums cannot build (a dead variant, a legacy spelling included), a variant list naming one variant twice, an `extruder_type`, `nozzle_volume_type` or `default_nozzle_volume_type` that is not an enum name ([variant names](#variant-names)) | per vendor | `check_variant_names` |
| a variant array not exactly `variant length × stride` wide in a selectable preset that writes it (with `--strict`, also in one it reaches); machine variant layout keys that disagree; a process id array that does not pair each variant ([variant arrays](#variant-arrays)) | per vendor | `check_variant_arrays` |
| a declared `filament_id` longer than 8 characters | per vendor | `check_filament_id_length` |
| a `default_materials` name, or a `default_filament_profile` name, matching no filament file ([below](#default-material-references)) | per vendor | `check_machine_default_materials` |
| per-key warnings for ignored options, **filament files only** | per vendor | `check_obsolete_keys` |
| `setting_id` uniqueness, every `filament_id` rule, and `machine_model` names duplicated across bundles | **tree-wide, ignoring `--vendor` entirely** | `check_setting_id_uniqueness`, `check_filament_ids`, `check_machine_model_name_uniqueness` |
Because the id and model-name checks stay tree-wide, a vendor-scoped run can and does fail on another
vendor's files, and it saves seconds, not minutes.
Unscoped, the per-vendor pass covers every bundle. The only exclusion is the stray `user/` directory
(below); OrcaFilamentLibrary is held to the same rules as any vendor, its sole exemption being that a
library filament may leave `compatible_printers` empty. `check_normalized` covers every bundle with an
index.
Notes that matter:
- Exit codes: **0** clean, **1** errors found, **2** argparse misuse. Warnings never change the exit
code.
- A nonexistent `--vendor` is a hard error: `[ERROR] unknown vendor "<V>" in <dir>`, exit 1.
-`--vendor ""` means all vendors; `check_profile.sh` relies on that. `--vendor` is repeatable
(`check --vendor A --vendor OrcaFilamentLibrary`); `check_profile.sh` takes one.
- A **stray directory** under `resources/profiles/` still gets counted as a vendor by the per-vendor
pass and warned about (`No profiles found for vendor: <dir> at …/<dir>.json`, and the "Checked
vendors" count goes up by one): usually an emptied folder, or a `user/` left by a direct validator
run. An unscoped `check` skips `user/` by name; `--vendor user` still checks and warns about it.
`normalize`, `trim` and `update-index` ignore strays too: they define a bundle as *a directory with a
matching index file*.
- Each remedy is printed once for the whole run, not once per file, as a `[WARNING]` under the errors
(`2 unreferenced file(s) above: delete them, or run … update-index`). Read those lines: they name the
command that fixes the batch.
- When there are errors or warnings, the trailing summary suggests `normalize`. That is right for the
shape errors and misleading for everything else: an id error needs `generate-id`, a dangling
`default_materials` needs a human.
- Other options: `--dry-run` on every writing command, `--profiles DIR` to point any command at another
tree, `--profile-type` to narrow `normalize`, `trim` and `update-index` to one type
([ids.md](ids.md#the-tool)).
-`resources/profiles/check_unused_setting_id.py` is a legacy BBL-only diagnostic, not part of
profile CI. Use `orca_profile_tool.py check` for current id validation.
### Obsolete keys
`check` always reports per-key warnings for obsolete options in filament profiles. Its normalization
check also rejects obsolete keys across all preset types (`normalize would remove <key>`); `normalize`
removes them.
### Default-material references
The materials check finds `default_materials` / `default_filament_profile` entries naming a preset that
does not exist. It reads each `machine/` file's own key (a model's `default_materials`, a variant's
`default_filament_profile`; a file that writes both is checked on `default_materials` only) and accepts
any `name` found in the vendor's or OrcaFilamentLibrary's `filament/` files, bases and unindexed files
included. The three authoring errors it surfaces are `,` instead of `;`, wrong case (`@system`), and a
whole `;`-joined string stuffed into one array element. Only the validator (`validate_system`) requires
an instantiated system filament that is compatible with each variant.
### Variant arrays
`check_variant_arrays` composes every selectable preset the loader's way (the parent, then each
`include` in order, then the file's own keys; a filament's `inherits` may fall through to
OrcaFilamentLibrary) and holds each key of [the four variant sets](extruder-variants.md#the-four-key-sets)
that the preset writes itself to exactly its `variant length × stride`
([widths](extruder-variants.md#widths)). The variant length is the length of the composed preset's
own `*_extruder_variant` list; without one, a machine's is the number of variants its
`extruder_variant_list` offers, else its extruder count (the list's default is one
`Direct Drive Standard` per extruder), and a process's or filament's is 1. Any other width is an
error, one value included and even when every value is the same. A key the preset does not write
takes what reaches it, the default or an array it inherits or includes, which the loader resizes; it
is not checked. A base is not judged on its own: what it writes counts only where it reaches a preset
that does not override it.
`check --strict` also holds every selectable preset to its own width for each key that reaches it,
so a preset whose variants differ from those of the file its array comes from restates the array
(the error names that file). That is BBL's practice and the target for new printer-specific presets;
CI runs `check` without `--strict`.
An id array that reaches a selectable preset, `printer_extruder_id` or `print_extruder_id`, written or
inherited, must have one entry per entry of its variant list on any printer. This rule and the
machine layout rules below judge the composed preset without `--strict`, whichever file writes the
keys. Beside a written variant list this rule reports it instead of the width
rule, so it is reported once. A process that
lists variants without `print_extruder_id` gets a **warning** when a variant repeats (every entry then
reads as extruder 1, so the repeated variant is unreachable), nothing otherwise.
On a machine the layout keys are held to [Printer rules 1–4](extruder-variants.md#printer-machine),
every failure an error unless marked:
- With `extruder_variant_list` written: the list has one entry per extruder, as many as
`nozzle_diameter`; every variant starts with its extruder's `extruder_type`;
`default_nozzle_volume_type` names a nozzle volume type that extruder lists;
`printer_extruder_variant` is the list flattened extruder-major and `printer_extruder_id` gives each
entry its 1-based extruder (an id array left out reads as extruder 1 everywhere, which passes when
those are the flattening's ids). Without the pair, the list may offer one variant in total.
- With the pair written and no `extruder_variant_list`: one variant per extruder at most. A pair that
`single_extruder_multi_material` off would replace with the default at load is a **warning**; a pair
the rebuild would leave as it is passes.
It does **not** see a base's own resize: it composes at the width each file wrote, so an array wider
than a base's list, which the loader cuts before any child inherits it, passes even with `--strict`
([composition](extruder-variants.md#padding-truncation-and-composition)). Nor does it see per-extruder
vectors outside the sets (`extruder_offset`, `printer_extruder_options`, …), which no variant list
sizes. What a variant holds (a High Flow variant copied from Standard, a variant inserted at the wrong
| Linux | `$XDG_CONFIG_HOME/OrcaSlicer`, or `~/.config/OrcaSlicer` when unset |
| Windows | `%APPDATA%\OrcaSlicer` |
A portable `data_dir` next to the executable takes precedence. Use a separate test configuration for a
clean-install check; preserve the normal configuration and user presets.
## Error → remedy
| Message | Fix |
| --- | --- |
| `can not find inherits <parent> for <preset>` | parent missing, unregistered, misspelled, or listed **after** the child |
| `can not find include` | the template is misspelled, registered after the includer, or selectable |
| `can not find filament_id for <name>` | nothing in the chain declares one: run `generate-id` |
| `can not find parent <name> for config <user preset>!` | a shipped name disappeared: add `renamed_from` |
| `Failed loading configuration file <file>` | that file could not be loaded and the whole bundle was discarded: a JSON error, or a value its option cannot take, such as `nil` in a non-nullable key (`Invalid value provided for parameter <key>: nil`, `Deserializing nil into a non-nullable object`); the lines above it name the cause |
| `Missing instantiation attribute for <name>` | key absent **or** not the string `"true"` / `"false"` |
| `contains incorrect keys: <keys>, which were removed` | a key valid for a different preset type |
| `defines invalid printer variant "<v>"` | not a token of the model's `nozzle_diameter` list |
| `has printer_variant "<v>" that does not match its nozzle_diameter` | [the set comparison](machine-profiles.md#printer_model-and-printer_variant) |
| `references unknown compatible_printers "<p>"` | the printer was renamed or deleted, or a `machine_model` name was used instead of a variant name: fix the reference. Under `--vendor`, usually another vendor's printer ([why](#the-five-checks)) |
| `references renamed compatible_printers "<old>" (now "<new>")` | in-tree references must name the current preset; `renamed_from` does not excuse them |
| `Filament preset "<f>" is missing compatible_printers setting` | non-library filaments need a non-empty list in their **own** file; the resolved-vs-own-key trap is in [filament-profiles.md](filament-profiles.md#compatible_printers) |
| `Ambiguous AMS filament match: N filament presets share filament_id "X" and are all compatible with printer "Y"` | make the lists disjoint by [specificity](filament-profiles.md#overlapping-coverage-one-variant-one-profile-per-product): the specialized profile keeps the variant, the general ones drop it; prefer this over deleting a profile. Or fix an `inherits` pointing at another product's `@base`. `orca_profile_tool.py check` does not catch this; only `validate_system` does |
| `Layer height cannot exceed nozzle diameter.` / `Line width too small` / `Line width too large` | the [slicing limits](process-profiles.md#values-to-review-per-nozzle) |
| `[ERROR] … no <V>.json list references it, so it never loads` | `update-index`, or delete the file |
| `[ERROR] … no <V>.json list references it and it declares no profile type` | set the correct `type` explicitly, then `normalize` and `update-index` |
| `[ERROR] … normalize would <change>` / `<V>.json: update-index would rebuild <lists>` | run that command and commit the result |
| `[ERROR] <V> has N <type> profiles named "<name>"` | identify the intended preset and remove or rename the duplicate; use `trim --dry-run` only for deliberate unindexed-file cleanup |
| `Duplicate key error in <file>: Duplicate key detected: <key>` | a key written twice in one file; keep the intended one |
| `… must not have a setting_id` / `… is missing a setting_id` / `setting_id "X" in <file> does not match the expected "Y" …` | `generate-id --setting-id` (by hand in `BBL/`, see [ids.md](ids.md#bbls-exception-precisely)) |
| `filament_id "X" declared by … does not match the mint of its triple …` | `generate-id --filament-id`; if the id should not be declared here at all, remove it so the preset inherits its root's |
| `inherits filament_id "X" but its own triple "V/T/N" mints "Y"` | `generate-id` will **not** fix this; see [ids.md](ids.md#what-generate-id-does-and-does-not-fix) |
| `"<key>" has N values for variant length S at stride k, which takes M` (with ` (no <list key>, so …)` after `S`, naming where `S` came from, when the preset writes no list, and ` (it comes from <file>)` at the end under `--strict` for an array the preset does not write) | exactly `S × k` values in variant order, or leave the key out ([widths](extruder-variants.md#widths)); on a list-less multi-extruder printer whose extruders differ, declare the layout; `fix-variant` cuts or pads to that width ([variant arrays](#variant-arrays)); an `S` you did not expect means the variant list did not resolve through `include` or `inherits` |
| `printer_extruder_variant […] is not extruder_variant_list flattened extruder-major […]` / `printer_extruder_id […] does not give each entry of printer_extruder_variant its 1-based extruder […]` / `printer_extruder_id has N entries for the M entries of printer_extruder_variant` | write the pair as the flattening of the list, ids in step ([Printer rule 2](extruder-variants.md#printer-machine)) |
| `extruder_variant_list has N entries for M extruder(s)` / `extruder_variant_list entry i "…" holds a variant that does not start with extruder i's extruder_type` / `default_nozzle_volume_type "…" is not a nozzle volume type extruder i lists` | [Printer rules 1, 3 and 4](extruder-variants.md#printer-machine) |
| `extruder_variant_list offers N variants but the preset writes no printer_extruder_variant/printer_extruder_id` / `printer_extruder_variant lists several variants for extruder N but the preset writes no extruder_variant_list` / `[WARNING] … with single_extruder_multi_material off the loader replaces printer_extruder_variant …` | write all three layout keys ([Printer rule 2](extruder-variants.md#printer-machine)) |
| `print_extruder_id has N entries for the M entries of print_extruder_variant` / `[WARNING] … print_extruder_variant repeats a variant but print_extruder_id is absent` | one id per variant entry, mirroring the printer's pairs ([Process rule 1](extruder-variants.md#process)) |
| `<key> <where> holds "…", which no extruder can select: "…" is not a nozzle volume type the enum has (…)` / `… it does not start with an extruder type the enum has (…)` / `… is empty` / `… Hybrid names the sub-nozzles of one hybrid extruder at runtime` | write a legal variant string, `<extruder type> <nozzle volume type>` from the two enums ([variant names](#variant-names), [variant strings](extruder-variants.md#variant-strings)); in every bundle, BBL included |
| `… holds the legacy variant "…"` / `extruder_type spells the legacy name "…"` / `nozzle_volume_type spells the legacy name "…"` / `default_nozzle_volume_type spells the legacy name "…"` | write the enum name the message gives: the loader still rewrites the legacy one, but `check` rejects it |
| `<key> lists "…" N times` / `print_extruder_variant lists the pair (extruder N, "…") N times` | drop the repeat and its value from every variant array: the lookup returns the first equal string |
| `extruder_type "…" is not one of (…)` / `nozzle_volume_type "…" is not one of (…)` / `default_nozzle_volume_type "…" is not one of (…)` / `… names "Hybrid", which the engine computes for a hybrid extruder at runtime` | they are enum options, so an unknown value fails the validator's load of the whole bundle and silently becomes the default in the app; use `s_keys_map_ExtruderType` / `s_keys_map_NozzleVolumeType`, `Hybrid` excepted ([variant strings](extruder-variants.md#variant-strings)) |
| `[WARNING] No profiles found for vendor: <dir>` | a directory with no matching index (an emptied folder, or a `user/` left by a direct validator run); remove it |
| `… has no compatible system filament in its model's "default_materials"` | add a system filament preset compatible with that variant to the model's `default_materials` |
| `… names the unknown system filament "<n>" in its "default_materials"` / `… "default_filament_profile"` | name an existing system (not user, not base) filament exactly; `;` separators, exact case |
| `Missing filament profile: '<n>' referenced in <file>` | the same, caught by `check`; usual causes are `,` instead of `;`, wrong case (`@system`), or a `;`-joined list packed into one array element |
| `machine_model name "<n>" is declared by N bundles` | model names are unique across the tree; rename the new model |
| `vendor <V>'s config version: <s> invalid` | the `version` string does not parse; write `MM.mm.pp.bb` |
| `[json.exception.type_error.302] type must be string` | locate the non-string value in the index or a model; see [failure scopes](vendor-bundle.md#failure-scopes) |
| `Printer "<p>" fell back to a default preset` | the final process or filament selection is a generic Default preset: check the named defaults, their visibility and that compatible presets exist. An incompatible default may instead be replaced without this error |
| `Printer "<p>" sliced but the filament change never fired (no CP TOOLCHANGE START)` | `change_filament_gcode` never expanded |
## CI
`.github/workflows/check_profiles.yml`, job **"Check profiles"**, runs on `pull_request` into `main` or
1.**Everything is registered, bases included.** Every preset file on disk has exactly one entry in the
matching list, and no unindexed preset file is left in the tree.
2.**Parents before children, includes before includers.** The lists load processes first, then
filaments, then printers, each in index order, and `inherits` and `include` resolve only against
presets of that type already loaded from it. A parent
listed after its child produces `can not find inherits <parent> for <child>` and the bundle is
discarded; an include listed after its includer is `can not find include`, a counted error that
leaves the includer without those keys.
3.**The index entry's `name` equals the `name` inside the `sub_path` file.**`renamed_from` does not
excuse a mismatch.
All three are `check` errors, and `update-index` writes an index that satisfies all three from the
files on disk, including the dependency ordering (parents and templates before the presets that use
them, then, in name order, entries that neither depend on nor are depended on by another entry of their own
list, and any entry on a dependency cycle). Hand-editing the index is
fine for a one-line addition, but the committed result must equal what `update-index` writes, because
`check` compares them.
The loader itself reports none of this: an unregistered file, or an entry with a misspelled key
(`"subpath"`), is silently dropped. (A misspelled `sub_path` value is a `check` error naming the entry.)
`BBL/cli_config.json` and, in `BBL/filament/`, `filaments_color_codes.json`, `filament_id_map.json`,
`filament_name_map.json` and `support_recommended_params.json` are auxiliary data files read by path,
not presets. The last three carry a `type` key and look like presets; the tool excludes all five from
preset maintenance.
## `version`
Four components, `MM.mm.pp.bb`, compared as a version number in which the fourth is folded into the
third (`patch × 100 + build`). Write all four components, zero-padded.
- **Bump the version for every bundle the change touches.** The app installs bundled profiles only
when their version is newer than the installed one, and the `.opc` preset cache is also keyed on
the version. Nothing in profile CI checks the bump.
- **Keep the last component ≤ 99.** `02.04.00.100` and `02.04.01.00` both read as
`2.4.100`. A bundle that
reaches `.99` carries into the third component (`02.03.02.99` → `02.03.03.00`).
- An **absent** version is worse than a stale one: it reads as `0.0.0`, which is not a valid version.
`check` and the validator still pass, but the vendor is dropped from the setup wizard entirely and gets no preset cache. Confirm the key exists. An
*unparseable* version is not silent: it discards the whole bundle (`vendor <V>'s config version: <s>
invalid`).
## Common preset keys
| Key | Value |
| --- | --- |
| `type` | `machine_model`, `machine`, `process` or `filament` |
| `name` | the preset name, the identity every reference uses; the filename is *not* authoritative |
| `inherits` | the parent's exact `name`: no path, no `.json` |
| `include` | a template's exact `name`, or an array of them, layered under this preset's own keys ([below](#inherits-and-include)) |
| `instantiation` | the **string**`"true"` (selectable) or `"false"` (base) |
| `from` | `"system"` for shipped presets |
| `setting_id` | generated; required on instantiated presets, forbidden on bases |
| `renamed_from` | `;`-separated old names this preset supersedes ([below](#renamed_from)) |
These are config-preset keys; `machine_model` records have their own
[key set](machine-profiles.md#machine_model-a-record-not-a-config-preset). Keep `from` as `"system"`:
the bundle loader ignores it, but loading the file as a CLI config accepts only `system`, `user` or
`User` and handles their inheritance differently.
`instantiation` is the one metadata key the validator gates: a missing key or any value other than the
strings `"true"` / `"false"` is a counted error (`Missing instantiation attribute for <name>`) that fails
the validator, though the preset still loads and is treated as selectable. A file with no
`instantiation` whose name contains `gcode`, or that has no `name`, silently becomes an include-only
template. A JSON boolean `true` fails harder: it takes the **whole vendor bundle** down.
## `inherits` and `include`
`inherits` resolves by exact name **within the same bundle**, plus one exception: filaments may inherit
from OrcaFilamentLibrary, which is loaded first. Vendor-to-vendor inheritance always fails, and an
unresolved `inherits` discards the bundle. You can inherit from an instantiated preset as well as from a
base.
`"include": ["<name>", …]` (or one bare name) pulls in `instantiation: "false"` presets of the same type
from the same bundle (never the library), registered before the includer. It shares a block of keys
between presets that do not share a parent: a variant layout, a G-code template. Only `"false"` presets
can be included, so a name that resolves to nothing (misspelled, registered after the includer, or a
selectable preset) is a counted error (`can not find include`) and the preset loads without it.
**How a preset's config is composed:** start from the parent's stored config (a root starts from the
built-in defaults), apply each preset named in `include` in the order listed, then the preset's own
keys. Later layers win, so precedence is own keys > later includes > earlier includes > the `inherits`
chain. Only then is every variant key of the composed config resized to its variant length
([widths](extruder-variants.md#widths)), and keys of another preset type removed. The two routes hand
down different widths:
- **`inherits` hands down the resized config.** A base is stored *after* its own resize, at the length
of its own `*_extruder_variant` (one variant for a base of any type that writes none, whatever its
extruder count). A child therefore inherits the base's arrays at the base's width: an array wider than
that is cut to its first values before any child sees it, and a child that adds variants gets those
first values padded. So widen an array only on a preset whose own variant list already has the
entries; a wide array on a narrow base is silently lost at load, and `check`, which composes at the
width each file wrote and judges selectable presets only, misses that cut.
- **`include` hands down the template's diff, at its pre-resize width.** An included preset contributes
every key where its own composed config (its parent, its own includes and its own keys) differs from
the built-in defaults, taken *before* its resize. So keys the template inherits are passed on too,
arrays it writes arrive at the width its file wrote, and a key it sets to the built-in default value
is not passed on at all, so it cannot override what the includer inherited. Resizing happens on the
includer, not on the template.
## `renamed_from`
One JSON string, `;`-separated for several old names.
- Write `"A;B"`, never `"A ; B"`: a space after a `;` is skipped, but a space before it stays part of
the name (`"A "`), which can never match.
- When `renamed_from` is **absent** and the name contains `@`, the loader auto-adds the `@`-removed form
(`X @Y` → `X Y`) as a rename alias. Declaring an explicit `renamed_from`**suppresses** that, so a
preset that needs both the `@`-removed form and a real old name must list both; a preset that gains
a `renamed_from` without it quietly loses its `X Y` alias.
- It rescues names stored **outside** the tree: user presets and 3MF projects. It does **not** rescue
in-tree `inherits` (exact lookup), it does **not** satisfy the index-name rule, the validator reports
an in-tree reference that only resolves through it (`references renamed compatible_printers "OLD"
(now "NEW")`), and `machine_model` records never read it at all.
- Only one preset may claim a given old name; two that do is a counted error
(`… was marked as renamed from "Y" … as well`). But the redirect is **inert while a live preset still
carries that name**, and nothing checks *that*, so a neighbour's `renamed_from` is no model.
## Failure scopes
| Scope | Cause |
| --- | --- |
| **Every vendor except OrcaFilamentLibrary, and all user presets** | a non-string where the index or a `machine_model` expects a string (`"version": 2` at the top level of an index, a numeric `name` or `url`, a non-string `nozzle_diameter` or other model key): `[json.exception.type_error.302] type must be string`, and the validator reports `Validation failed` |
| **The whole vendor bundle** | index JSON parse error; unparseable `version`; a listed file missing or unparseable; a value its option cannot take, such as `nil` in a non-nullable key (`Failed loading configuration file`); unresolved `inherits`; two selectable presets with one name; empty or unknown `printer_model` / `printer_variant`; a filament resolving no `filament_id`; a JSON boolean `instantiation` |
| **A counted error; the preset still loads** | `instantiation` missing or not `"true"` / `"false"`; keys belonging to another preset type (`contains incorrect keys: …, which were removed`); a non-string inside a `*_list` entry (`invalid value type for <key>`); an `include` naming nothing usable (`can not find include`, loads without it) |
| **The rest of the file, logged only** | an array with a non-string element (`[0.4]`, `invalid json array`): that key and every key after it in the file are dropped, and no error is counted |
| **One value, logged only** | a raw JSON number in a preset (`invalid json type for <key>`): the value is dropped and the exit code stays 0 |
| **Nothing reported by the loader** | unregistered file; two bases with one name, or a base and a selectable preset with one name (the first in the index wins); misspelled setting key; missing bed, hotend or cover asset. `check` catches the first two; the others reach users |
Deleting a file the index still lists surfaces as a *parse error* on line 1 (`unexpected end of input`), not "file
not found".
Selectable preset names are a **single namespace across every vendor**: a duplicate within one vendor
is a hard bundle failure, and a duplicate across vendors is reported as `Found duplicated preset: <name>
in vendor: <vendor>` and still counts as an error. `check` catches the within-bundle case earlier and
more precisely, bases included, and including an *unindexed* twin, which is one `sub_path` edit away
from silently becoming the parent every child resolves to (the first registered preset of a name wins,
so index order decides). Base names, by contrast, repeat across bundles by design:
every bundle may have its own `fdm_process_common` ([uniqueness](naming.md#uniqueness)).
## Starting a whole new vendor bundle
Nothing generates one; copy the smallest bundle that resembles the hardware. **`Voxelab`** is the
minimal shape: a shared machine base, the model, one variant, a shared process base, two processes, and
an empty `filament_list`, so the printer takes the library generics. Do *not* start from a bundle that
carries local `fdm_filament_*` copies, which drift from the library, or filament presets that restate
most of their parent, the style this skill advises against.
Write the machine files **last**, so you only visit them once:
1.**Choose the names first**: model, variant(s), process(es). Everything else references them
description:Use when updating, filling, reviewing or fixing OrcaSlicer translations in localization/i18n — running scripts/run_gettext --full, translating empty or new msgids in the .po catalogs, AI translation passes and their "# AI Translated" / "# Needs human review" comments, plural forms, msgctxt homonyms, and keeping catalogs consistent with the localization glossary (OrcaSlicer_WIKI localization_glossary.tsv), including adding new glossary terms. Also use for msgfmt or run_gettext failures, placeholder or line-break mismatches, and reviewing a translation PR against the glossary.
---
# OrcaSlicer translations
The UI strings live in gettext catalogs: `localization/i18n/OrcaSlicer.pot` is the template, and
`localization/i18n/<lang>/OrcaSlicer_<lang>.po` holds each language. The `en` catalog is empty on
purpose, so skip it. Recurring terms are fixed by the
- **Batch files go in a scratch directory, not the repository.**
- **Each entry's `id` identifies it,** so the translation file need not repeat the English. `apply`
also accepts `msgid` with `msgctxt`, or the pending file itself with its `msgstr` fields filled.
- **Translate every entry in the batch.** A blank `msgstr` is skipped, not cleared.
- **Next batch:** applied entries leave the list, so run `pending` again at offset 0. Use `--offset`
only to step past entries you deliberately leave blank.
- **What each entry carries:**
- `glossary`: the terms it contains, with the language's rendering.
- `note`: the developer's `TRN` comment.
- `flags`: `c-format` or `boost-format`.
- `description`: present when a sense has to be chosen.
3. Read every apply message:
- **REJECTED:** a hard rule broke. Fix the translation and apply the corrected items again.
- **WARNING:** something worth a second look, such as a glossary rendering not found, ending
punctuation, or snake_case. A glossary warning on an inflected form you checked is fine.
Otherwise correct the translation (`apply --overwrite` replaces your own entry), or add a
`review` reason.
`apply` adds `# AI Translated` to every entry it writes, and `# Needs human review: <reason>` where an
item has `review`. On an `--overwrite`, `"review": ""` removes an earlier note. Never add either to a translation a person wrote. `--human` is for applying their
text.
**The work is large:** every new string, times every language. When subagents are available, give each
one a language or a few. Each agent only touches its own catalogs, so they can run
in parallel. Settle step 3 first. Give each agent the language list, this skill's path, the rules file,
and these instructions:
- Fill every empty entry.
- Don't add glossary rows.
- Report what it could not translate, the review notes it added, and any term it thinks the glossary
still lacks.
Add the terms they report yourself, then have the affected languages revisit the entries.
### 5. Verify
```bash
python3 $T check --lang <L> --warnings # every AI-translated entry, plus msgfmt --check-format
```
- **Run the full gettext again.** It must exit 0 and leave the catalogs as the pass left them.
- **Run `python generate_glossary.py --check` in the wiki clone.**
- **Errors on entries you didn't write predate the pass.** Report them rather than fixing them in the
same change.
### 6. Report
Report:
- the entries filled, per language;
- the review notes added;
- anything left empty, and why;
- the glossary rows added or filled.
The catalogs (and `OrcaSlicer.pot`) and the glossary (`localization_glossary.tsv` and `.md`) are
separate changes in separate repositories. The wiki change should land with, or before, the catalog
change. Commit or open PRs only when asked.
## Rules
1. **Run the full gettext first.** Translate against the regenerated template, never a stale one.
2. **The glossary is binding.** Use the language's rendering for each term, in the right sense. When a
rendering looks wrong, use it and flag the entry. Then raise the change with the user, rather than
deviating in one string.
3. **New recurring terms go into the glossary before the translations that use them.** Without a wiki
clone, stop and tell the user (step 3).
4. **Fill, don't rewrite.** A fill pass writes only empty and fuzzy entries.
- Existing translations, human or AI, change only on request (`--overwrite`).
- A glossary violation in a human translation is reported, not changed.
5. **Machine output is labelled.** Every AI translation gets `# AI Translated`; uncertain ones also get
`# Needs human review: <reason>`. Never use `fuzzy`.
6. **Never edit `msgid`, `msgctxt`, the header, or obsolete `#~` entries,** and never fix the English in
a translation. Report source problems (see the rules file).
7. **Write catalogs only through `po_tool.py` or gettext.** PO libraries such as polib reflow unrelated
entries, and hand edits skip the validation.
## Changing an established term
A glossary cell is usually copied from the catalog's own label. When the label is wrong, the glossary
inherits the mistake, so fix both in the same change:
1. **Check the catalog's usage first.** `lookup --lang <L> "<term>"` searches the English, and with
`--msgstr` it searches the translations.
- When most entries already use the correct form, move the cell to it.
- When the wrong form is used consistently throughout, leave both as they are and ask a
translator.
2. **Update the cell:** `glossary-set --term "<Term>" --replace --set <L>=<value>`.
3. **Rewrite the entries that still use the old form,** only where the term has that meaning.
- Write `[{"id": "<id from lookup>", "msgstr": "<full corrected text>"}]` and run
`apply --overwrite`.
- The rewritten entries are marked `# AI Translated`.
## Reviewing a translation change
Review it for these:
- glossary compliance (`check --scope all --warnings` on the language, then read the hits);
- placeholders and line breaks (`check` errors);
- `# AI Translated` kept on text a person rewrote, or added to text a person wrote;
- purged obsolete entries;
- reflowed unrelated entries;
- plural forms short of `nplurals`.
When a reviewer disputes a term, compare it with the glossary first. A machine translation that follows
the glossary is correct even if the term itself is debatable. Then the fix is a glossary change, made
with the translators, followed by the catalogs.
## Source of truth
When this skill and the checkouts disagree, the checkouts win:
- `scripts/run_gettext.sh` / `.bat` for extraction and merging;
- `localization/i18n/list.txt` for the scanned sources;
- the wiki's `localization_guide.md` and `localization_glossary.md` / `.tsv` for the rules and terms;
- `AGENTS.md` (Localization & translations) for the repository's conventions.
## Editing this skill
State rules and procedures, not the state of the catalogs: no counts, lists of untranslated strings or
per-language defects. Language-specific terminology belongs in the glossary's *Language specifics*
description:Use when writing, modifying, reviewing or debugging any OrcaSlicer GUI code under src/slic3r/GUI, or when deciding how a wxWidgets API behaves in OrcaSlicer — dialogs, frames, panels, the sidebar, Preferences, device/monitor pages, custom widgets, popups and menus, sizers and layout, painting, DPI scaling, dark mode, colours, icons, fonts, translated strings, settings fields, keyboard shortcuts, mouse capture and focus, events, CallAfter, threads and timers, WebView, the OpenGL canvas, AUI docking, clipboard, drag and drop, file dialogs — and for platform-specific UI bugs on Windows, macOS or Linux GTK/X11/Wayland such as popups that close at once, a frozen or unclickable UI, collapsed or clipped dialogs, invisible dark-mode icons, or crashes on close, even when the task never mentions wxWidgets.
---
# OrcaSlicer wxWidgets GUI
This skill is how OrcaSlicer's wxWidgets GUI is written, changed, fixed and reviewed. Its core is
**wxWidgets 3.3.2 API usage** — the documented contracts, the per-platform behaviour and the known
limitations of the wx version Orca pins — layered with the OrcaSlicer conventions, wrappers, custom
widget library and stable component designs a contributor must follow. Every reference file opens
with a numbered **Rules** checklist (what a diff is checked against), followed by sections that give
the wx contract, the correct code shape, platform differences, the Orca layer, and pitfalls as
wrong → right pairs with the commit that fixed each one.
## Ground truth: the wx tree in deps/
Orca pins **wxWidgets 3.3.2** from the fork `github.com/SoftFever/Orca-deps-wxWidgets` (tag `v3.3.2`,
`deps/wxWidgets/wxWidgets.cmake`), built static (Flatpak: shared). Facts about this build that change
how you read the wx docs:
- **Linux builds against GTK3** (`DEP_WX_GTK3` defaults ON in `deps/CMakeLists.txt`). Target GTK3 on
both X11 and Wayland; GTK-guarded code must still compile on GTK2, an opt-out build Orca does not ship.
- **wx asserts never fire.** wx is built with `wxBUILD_DEBUG_LEVEL=0` and `libslic3r_gui` with
`wxDEBUG_LEVEL=0`. Wherever the docs say a call "asserts", Orca silently ignores it, returns early or
corrupts state. Check preconditions yourself; do not expect a debug build to catch misuse.
- **No SVG in wx** (`wxUSE_NANOSVG=OFF`): `wxBitmapBundle::FromSVG*` does not exist. Orca rasterises
its SVG icons itself (`BitmapCache`, `create_scaled_bitmap`, `ScalableBitmap`).
Look things up in the source the app is built from — it beats memory, and 3.3 changed real behaviour:
```bash
WX=$(find -L deps -maxdepth 5 -type d -path '*dep_wxWidgets-prefix/src/dep_wxWidgets' 2>/dev/null | head -1)
// sizes (no msw_buttons_rescale(): it would override the DialogButtons' style height), then:
GetSizer()->SetSizeHints(this);
Refresh();
}
};
MyDialog dlg(this); // modal on the caller's stack
if (dlg.ShowModal() == wxID_OK) { /* read results */ }
```
## Orca widgets at a glance
| Orca widget (`src/slic3r/GUI/Widgets/`) | Instead of | Must know |
|---|---|---|
| `Button` | `wxButton` | ctor `(parent, text, icon_name, style, iconSize, id)` — id is last; style with `SetStyle(ButtonStyle::…, ButtonType::…)`; emits `wxEVT_BUTTON`; not a `wxButton` (no dialog default/escape emulation) |
| `::CheckBox` | `wxCheckBox` | no label parameter — pair with a `wxStaticText`/`Label`; emits `wxEVT_TOGGLEBUTTON` (never `wxEVT_CHECKBOX`), and a handler bound on it must `Skip()` or the bitmap is not refreshed |
| `::ComboBox` (+ `DropDown`) | `wxComboBox`/`wxChoice` | `wxCB_READONLY` for choices; `SelectAndNotify(n)` fires the event; ignores the ctor id; `void*` client data only |
| `::TextInput` | single-line `wxTextCtrl` | text via `GetTextCtrl()`; bind `wxEVT_TEXT_ENTER`/`wxEVT_KILL_FOCUS` on the widget, never on a parent |
| `SpinInput` | `wxSpinCtrl` | non-negative integers only (`-` cannot be typed); commits on Enter, kill-focus and arrow steps with `wxEVT_SPINCTRL`, a plain `wxCommandEvent` — bind a `wxCommandEvent&` handler and read `GetValue()` |
| `DialogButtons` | `wxStdDialogButtonSizer` | untranslated labels (custom ones written `L("…")`); OK/Cancel close by id, Yes/No/Apply/Confirm do not |
| `Label` | `wxStaticText` | `LB_AUTO_WRAP`, `LB_HYPERLINK` (only through `SetWindowStyleFlag`); also hosts the font table `Label::Head_*`/`Label::Body_*` |
| `SwitchButton`, `RadioGroup`, `TabCtrl`, `LabeledStaticBox`, `StaticLine`, `ProgressBar`, `PopupWindow` | toggle, `wxRadioBox`, notebook bar, `wxStaticBox`, `wxStaticLine`, `wxGauge`, `wxPopupTransientWindow` | quirks per widget in `orca-widgets.md` |
Disabling a parent does not repaint Orca widgets as disabled — enable/disable them individually (`::CheckBox`, a native
button, is the exception: it greys with its parent).
Containers stay raw (`wxPanel`, `wxBoxSizer`, `wxScrolledWindow`, `wxSimplebook`).
## Where to read next
| You are … / the symptom is … | Read |
|---|---|
| creating or closing a dialog or frame, `Destroy`/`delete`, modal results, liveness of a window pointer, message boxes | `references/windows-dialogs.md` |
| a sizer, a dialog that is collapsed, too big or clipped, a scrolled list, label wrapping, relayout after DPI change | `references/sizers-layout.md` |
| a custom-drawn control or a new widget in `Widgets/`, flicker, paint/erase, `wxGCDC`, best size | `references/painting-custom-widgets.md` |
| `FromDIP`/`em_unit`, rescale on DPI change, icons and bitmaps, image lists, fonts | `references/dpi-bitmaps-fonts.md` |
| colours, dark mode, theme toggle, `StateColor`, icons invisible in dark mode | `references/colours-dark-mode.md` |
| drag gestures and mouse capture, a frozen/unclickable UI on macOS, hover, wheel, keyboard shortcuts, focus, tooltips, cursors | `references/mouse-keyboard-focus.md` |
| popups and dropdowns (closing at once, not closing), context menus, the menu bar, `MenuFactory` | `references/popups-menus.md` |
| text/combo/check/radio/spin controls, book controls, `wxGrid`, `wxDataViewCtrl`, the object list | `references/controls-dataview.md` |
| WebView pages and dialogs, JS ↔ C++ messages, the GL canvas, ImGui overlays, docking panes, the top bar, camera view | `references/webview-gl-aui-media.md` |
| translations, string conversion and formatting, file/dir dialogs, clipboard, drag and drop, logging, `AppConfig` | `references/strings-i18n-files.md` |
| where new code goes, `GUI_App`/`MainFrame`/`Plater` structure, lazy pages, Preferences, notifications | `references/orca-architecture.md` |
| which Orca widget to use and its quirks | `references/orca-widgets.md` |
| adding or changing a print/filament/printer setting, a settings field, per-object overrides | `references/orca-settings-ui.md` |
| platform `#ifdef`s, wx build options, Wayland gaps, title bars, a bug on one platform only | `references/platforms.md` |
| old code, a wx call that behaves differently than you remember, wx-version migration | `references/wx-33-changes.md` |
**Reviewing a GUI diff:** for each area the diff touches, check it against that file's `## Rules`
list. **Debugging a UI bug:** find the symptom in the table above; most recurring Orca UI bugs are a
known class with a pitfall entry and a fixing commit.
## Platform gotchas worth memorising
- **macOS:** capture-lost is never sent (a leaked capture freezes all clicks); transient popups hover-
dismiss across a gap — anchor flush and re-verify the cursor; native modals (file/dir dialogs, native
message boxes) and generic progress dialogs re-activate the main window, so re-raise a secondary window
afterwards with a deferred, liveness-guarded `Raise()` — but never `Raise()` a `wxPopupWindow`, which makes
it the key window; a live menu accelerator consumes the key before
any wx key event; Control+click arrives as a right-click.
- **Windows:** `IsDark()` and `wxSYS_COLOUR_*` follow the system app mode, not Orca's theme — use
`dark_mode()`; menu bitmaps follow `check_dark_mode()`; windows are not double-buffered by default in
3.3.2; `ProcessLeftDown` is never called for popups; a popup that must close with the frame calls
`BindUnfocusEvent()`.
- **Linux GTK3:** dialogs without size hints collapse (only sizer-fitting calls, not `Fit()`, are
replayed at the first `Show()`); command events from a popup's children are not stopped at the popup
(MSW/macOS stop them); chained popups need `transient_for` set to the mapped parent right before
showing; native borders leak through custom widgets (`RemoveButtonBorder`/`RemoveInputBorder`).
- **Wayland:** no global pointer position, no window positioning, `wxClientDC`, `Update()` and `SetIcon`
do nothing, no floating AUI panes, GL is EGL only.
## Editing this skill
The skill describes how the pinned wxWidgets behaves and how Orca GUI code must be written, not what
the GUI tree currently contains. State rules, contracts, mechanisms, stable component designs and code
shapes; never usage counts, census lists, lists of today's offenders or dated measurements. Test each
sentence: if a change to Orca code the sentence does not name, with the wx pin unchanged, could make it
false, state the rule behind it or give a generic example instead, and cut it if there is no rule
behind it. Example files named as models to copy, and commit hashes cited as the reason for a rule,
are fine.
Keep the skill in step with what it describes. A change that alters a contract the skill states (an
Orca widget's API or quirk, a shared helper such as `DPIAware` or `UpdateDlgDarkUI`, the Shortcuts
registry) updates the skill in the same change. Moving the wx pin (`GIT_TAG` in
`deps/wxWidgets/wxWidgets.cmake`) means re-verifying every wx citation and **[source]** fact against
the new tree, and adding that version's behaviour changes the way `wx-33-changes.md` records 3.3's.
Cite wx by path relative to the wx tree root with line (`interface/wx/window.h:3805`), which stays
valid because the version is pinned. Cite Orca by file and symbol, never by line, and commits as the
rationale for a rule. Mark behaviour the wx docs don't state, or contradict, as **[source]**. Each
concept lives in one reference file and the others point to it. Verify every new claim in the wx tree
or the Orca code before adding it, and never drop a fact, qualifier or example to save words.
`evals/evals.json` holds the regression tasks (planted-defect review patches in `evals/files/`); rerun
them with and without the skill after substantial edits.
"prompt":"In the OrcaSlicer repo, add a new dialog that opens from the Help menu called 'Report Issue'. It needs a dropdown to pick the issue type (Bug / Feature request / Question), a multiline text box for the description, a checkbox 'Include system info', and OK/Cancel buttons at the bottom. It has to look correct in dark mode and on high-DPI screens.",
"files":[],
"assertions":[
"The dialog class derives from DPIDialog and overrides on_dpi_changed(const wxRect&) with a body that rescales its custom widgets (Rescale()/msw_rescale) and re-lays out (Layout/Fit/Refresh or equivalent)",
"The dropdown is Orca's custom ComboBox (::ComboBox from Widgets/ComboBox.hpp), not wxComboBox/wxChoice",
"The checkbox is Orca's custom ::CheckBox (constructed without a label) paired with a separate wxStaticText/Label for the text",
"The bottom row is DialogButtons constructed with untranslated labels such as {\"OK\", \"Cancel\"} (not _L(\"OK\")), and handlers end the dialog with EndModal(wxID_OK/wxID_CANCEL), not Destroy()",
"The multiline description is a raw wxTextCtrl created with wxTE_MULTILINE, not TextInput (TextInput::DoSetSize keeps the inner control at its single-line height), explicitly made dark-safe (UpdateDarkUI/UpdateDlgDarkUI or StateColor::darkModeColorFor colours), and its text is read via GetValue()",
"Every hard-coded pixel size/padding goes through FromDIP(n) (or em_unit multiples); no raw pixel literals in sizes/borders",
"All user-visible strings are wrapped in _L(...)",
"The top-level sizer is installed with SetSizerAndFit (or SetSizer followed by SetSizeHints)",
"wxGetApp().UpdateDlgDarkUI(this) is called as the last step of the constructor, after all children exist",
"The new .cpp/.hpp are added to SLIC3R_GUI_SOURCES in src/slic3r/CMakeLists.txt",
"A Help-menu item is added in MainFrame's menu construction (append_menu_item or equivalent) that opens the dialog with ShowModal()"
]
},
{
"id":2,
"name":"new-print-setting",
"prompt":"Add a new print setting to OrcaSlicer: 'seam_transition_gap', a float in mm, default 0.1, range 0-2, shown on the Quality page in the Seam group with a proper tooltip. It should be saved with process presets and searchable.",
"files":[],
"assertions":[
"A ConfigOptionDef is added in PrintConfig.cpp via this->add(\"seam_transition_gap\", coFloat) with label, tooltip, sidetext \"mm\", min 0, max 2, a mode, and default ConfigOptionFloat(0.1)",
"label/tooltip/sidetext in PrintConfig.cpp use the L(\"...\") extraction marker, not _L/_u8L",
"A ((ConfigOptionFloat, seam_transition_gap)) entry is added to the matching PRINT_CONFIG_CLASS_DEFINE block in PrintConfig.hpp",
"The key is appended to s_Preset_print_options in Preset.cpp",
"The answer states that search indexing comes automatically from get_option/append_single_option_line (no manual search registration)",
"Slicing invalidation is handled: the key is added to PrintObject::invalidate_state_by_config_options (or Print::invalidate_state_by_config_options) under an appropriate step"
]
},
{
"id":3,
"name":"macos-hover-popup-dismiss",
"prompt":"OrcaSlicer bug in a feature branch: the sidebar's filament-sync button now opens a small hover menu (code below). On macOS the menu appears and then disappears before the user can click either item — moving the mouse from the button down into the menu closes it, and sometimes it closes while the cursor is already over it. It works on Windows. Find the causes and fix the code.\n\n```cpp\n// Widgets/FilamentSyncMenu.hpp/.cpp (new, opened when the cursor enters the sidebar's sync button)\nclass FilamentSyncMenu : public PopupWindow\n{\npublic:\n explicit FilamentSyncMenu(wxWindow* parent) : PopupWindow(parent, wxBORDER_NONE)\n {\n auto* sizer = new wxBoxSizer(wxVERTICAL);\n auto* colours = new Button(this, _L(\"Sync colours only\"));\n auto* all = new Button(this, _L(\"Sync colours and types\"));\n colours->Bind(wxEVT_BUTTON, [this](wxCommandEvent&) { Dismiss(); wxGetApp().sidebar().sync_ams_list(); });\n all->Bind(wxEVT_BUTTON, [this](wxCommandEvent&) { Dismiss(); wxGetApp().sidebar().sync_ams_list(true); });\n sizer->Add(colours, 0, wxEXPAND | wxALL, FromDIP(4));\n sizer->Add(all, 0, wxEXPAND | wxALL, FromDIP(4));\n SetSizerAndFit(sizer);\n\n m_timer.SetOwner(this);\n Bind(wxEVT_TIMER, [this](wxTimerEvent&) { Dismiss(); });\n Bind(wxEVT_LEAVE_WINDOW, [this](wxMouseEvent&) { m_timer.StartOnce(300); });\n Bind(wxEVT_ENTER_WINDOW, [this](wxMouseEvent&) { m_timer.Stop(); });\n }\n\n void popup_under(wxWindow* btn)\n {\n wxPoint pos = btn->ClientToScreen(wxPoint(0, 0));\n Position(pos, {0, btn->GetSize().y + FromDIP(6)});\n Popup();\n }\n\nprivate:\n wxTimer m_timer;\n};\n\n// Sidebar constructor:\nm_sync_menu = new FilamentSyncMenu(ams_btn);\nams_btn->Bind(wxEVT_ENTER_WINDOW, [this](wxMouseEvent& e) { m_sync_menu->popup_under(ams_btn); e.Skip(); });\n```",
"files":[],
"assertions":[
"Identifies the opener-to-popup gap as a cause: the FromDIP(6) offset leaves a dead zone, and crossing it from the button into the menu ends the hover (LEAVE / hover-timer path) and dismisses the menu on macOS",
"Identifies that on macOS ENTER/LEAVE events around a wxPopupTransientWindow arrive spuriously or out of order (capture handling), so the hover timer must not trust them: re-verify the real cursor position with GetClientRect().Contains(ScreenToClient(wxGetMousePosition())) (or equivalent geometry) before starting/stopping the timer or dismissing",
"Fix anchors the menu flush with (or slightly overlapping) the button instead of leaving a gap, at least on macOS",
"Adds wxPU_CONTAINS_CONTROLS to the PopupWindow style because the menu hosts interactive Button children",
"Guards against calling Popup() again while the menu is already shown (re-entering the button) and/or debounces reopening right after a dismissal",
"The fix is gated to macOS where needed (#ifdef __WXOSX__/__APPLE__) or argued harmless elsewhere, and does not rely on unexplained SetFocus/CallAfter/Raise hacks"
]
},
{
"id":4,
"name":"dark-mode-icons",
"prompt":"Users report that a couple of the AMS status icons in OrcaSlicer are nearly invisible when dark mode is on, but fine in light mode. What's going on and how do we fix it properly?",
"files":[],
"assertions":[
"Explains that BitmapCache::load_svg recolors dark-mode icons by literal substitution of a fixed palette (e.g. #262E30 -> #EFEFF0) and off-palette colors pass through unchanged",
"Offers the asset fix: re-author the SVG fills in palette colors (e.g. near-black line art as #262E30)",
"Offers the variant fix: a *_dark asset selected by name from wxGetApp().dark_mode()",
"States that name-selected variants must be re-picked/reloaded on a runtime theme switch (on_sys_color_changed / sys_color_changed / msw_rescale path)",
"Investigates the actual AMS icon code/assets (e.g. checks both branches of a dark_mode() ternary for a copy-paste _light/_light bug, or inspects the SVG fills)"
]
},
{
"id":5,
"name":"custom-painted-drag-widget",
"prompt":"In OrcaSlicer, add a small owner-drawn widget for the filament area of the sidebar: a horizontal strip of colour swatches (one per filament) where the user can click-and-drag across swatches to select a contiguous range. Hovered swatches should highlight, it must emit an event with the selected range when the drag ends, and it must look right with high-DPI and dark mode. Write the widget.",
"files":[],
"assertions":[
"Paints only inside a wxEVT_PAINT handler using wxAutoBufferedPaintDC/wxBufferedPaintDC/wxPaintDC with SetBackgroundStyle(wxBG_STYLE_PAINT); never draws through wxClientDC; triggers repaint with Refresh()/RefreshRect()",
"CaptureMouse() on press is guarded (if (!HasCapture())) and ReleaseMouse() is guarded (if (HasCapture()))",
"Handles wxEVT_MOUSE_CAPTURE_LOST by ending the gesture: drag state reset and repaint, no recapture (cancelling rather than committing the selection is what the wx contract asks for)",
"Ensures capture cannot leak: release on every end-of-drag path regardless of drag flags (and/or in the destructor), mentioning the macOS consequence (UI unclickable) or the wx contract",
"Sizes use FromDIP (or em_unit) and the widget reports a size via DoGetBestSize/DoGetBestClientSize or SetMinSize, rescaled on DPI change (Rescale/msw_rescale/wxEVT_DPI_CHANGED)",
"Colours go through StateColor / StateColor::darkModeColorFor or are re-derived from wxGetApp().dark_mode() and re-applied on theme change (sys_color_changed/on_sys_color_changed)",
"Defines a custom event with wxDECLARE_EVENT/wxDEFINE_EVENT (e.g. a wxCommandEvent carrying the range) and uses Bind(), not a new static event table",
"Hover handling uses event-relative coordinates (evt.GetPosition()), not wxGetMousePosition(), and only refreshes when the hovered index actually changes"
]
},
{
"id":6,
"name":"worker-thread-progress",
"prompt":"In OrcaSlicer, we're adding a non-modal 'Firmware download' dialog. A network library calls our progress callback (int percent, std::string status) on its own worker thread, and a completion callback (bool ok, std::string error) at the end. The dialog shows a progress bar and status text and has a Cancel button; the user may also close the dialog at any time while the download continues. Implement the dialog and the callback wiring.",
"files":[],
"assertions":[
"No wx/GUI calls are made on the worker thread; callbacks marshal to the main thread via CallAfter (wxGetApp().CallAfter / window CallAfter) or wxQueueEvent",
"Deferred lambdas capture data by value (copies of percent/status strings), not references to worker-owned data",
"Deferred work re-checks the dialog's liveness inside the lambda (shared_ptr<atomic<bool>> alive flag, wxWeakRef, or registry lookup) because the dialog can be closed before the callback runs",
"If events are used across threads, they are heap-allocated/owned (wxQueueEvent with new/Clone, or wxThreadEvent) — not wxPostEvent/AddPendingEvent with wxString payload from the worker",
"Closing the dialog cancels or detaches the download callbacks and the non-modal dialog is destroyed with Destroy() (not delete), with any wxTimer stopped first",
"Uses Orca UI conventions: DPIDialog base, Orca ProgressBar or custom widgets, DialogButtons/Button for Cancel, _L strings, FromDIP sizes, UpdateDlgDarkUI",
"Strings crossing to the GUI are converted with from_u8() (UTF-8 std::string -> wxString), not implicit/ToStdString conversions"
]
},
{
"id":7,
"name":"scrolled-dynamic-layout",
"prompt":"OrcaSlicer bug on Linux: in a settings dialog, a wxScrolledWindow holds a list of rows that the user can add with a '+' button. After adding rows, the scrollbar doesn't appear/update and new rows are cut off; when the dialog first opens on GTK it's sometimes collapsed to a tiny size; and some German labels in the rows are clipped. Explain the causes and give the correct wx code patterns to fix all three.",
"files":[],
"assertions":[
"For the scroll issue: after adding rows, call FitInside() on the scrolled window (or SetVirtualSize/Layout so the virtual size updates), with SetScrollRate set so scrollbars are enabled",
"Explains that changing children doesn't change the scrolled window's size, so automatic layout doesn't run — Layout() (and FitInside) must be called explicitly after adding content",
"For the collapsed dialog: uses SetSizerAndFit on the top-level dialog or SetSizer + sizer->SetSizeHints(dialog) so the min size is set (GTK needs the min-size hint), and avoids an unconditional Fit() in on_dpi_changed/refresh paths",
"For clipped labels: avoids fixed widths/heights on wxStaticText (use -1 / best size) and re-Wrap()s after SetLabel or uses Label with LB_AUTO_WRAP; mentions translations being longer",
"Mentions InvalidateBestSize()/Layout() of the containing hierarchy (or parent->Layout()) after dynamic content changes, and/or Freeze()/Thaw() around bulk row creation",
"Notes GTK specifics: Linux builds wx against GTK3 by default (deps/CMakeLists.txt DEP_WX_GTK3 ON) and/or the first GTK size pass can run with a bogus tiny client width, so wrap/height calculations need a guard"
]
},
{
"id":8,
"name":"keyboard-shortcut",
"prompt":"Add a global keyboard shortcut Ctrl+Shift+E (Cmd+Shift+E on macOS) to OrcaSlicer that exports the current plate's G-code (same as the existing export action). It must work on all three platforms and show up wherever OrcaSlicer documents shortcuts.",
"files":[],
"assertions":[
"Explains that wxACCEL_CTRL / 'Ctrl' in menu accelerator strings maps to Cmd on macOS (and WXK_RAW_CONTROL / wxACCEL_RAW_CTRL is the real Control key)",
"Registers the shortcut through Orca's shortcut registry (Shortcut enum + shortcut_table in src/slic3r/GUI/Shortcuts.cpp, Global context dispatched from MainFrame's wxEVT_CHAR_HOOK) and/or the menu item whose accelerator text is derived from it, dispatching to the existing export handler",
"Checks for conflicts with existing shortcuts (registry defaults, GLCanvas3D/ObjectList contexts, menu items) before choosing the binding",
"The shortcut appears in KBShortcutsDialog by virtue of the registry (or is explicitly added there), and the answer checks/handles a conflict with an existing default chord",
"Accounts for platform differences in Orca's menus: native wxMenuBar on macOS vs BBLTopbar/custom menus on Windows/Linux, so the shortcut works where no native menubar exists"
]
},
{
"id":9,
"name":"webview-dialog",
"prompt":"Add a dialog to OrcaSlicer that shows a local HTML page from resources/web/ in a WebView. The page has a 'Done' button that posts a JS message; when it arrives the dialog must close and return a result to the caller. It must work on Windows (Edge), macOS (WKWebView) and Linux (WebKitGTK).",
"files":[],
"assertions":[
"Creates the browser through Orca's WebView wrapper (WebView::CreateWebView in Widgets/WebView.hpp), not wxWebView::New directly",
"Receives messages via wxEVT_WEBVIEW_SCRIPT_MESSAGE_RECEIVED using the 'wx' handler the wrapper registers (window.wx.postMessage / postMessage), without adding a duplicate script message handler (which throws an uncatchable NSException on WKWebView)",
"Closing/ending the dialog from the script-message handler is deferred with CallAfter (not done synchronously inside the WebView callback stack)",
"Builds the local page URL from Orca's resources dir (resources_dir()/from_u8 + file:// URL or wxFileName::FileNameToURL), handling paths portably",
"Accounts for asynchronous backend creation (Edge): no script runs/calls before the page is loaded (wxEVT_WEBVIEW_LOADED) or the wrapper's deferral",
"Dialog follows Orca conventions: DPIDialog base, on_dpi_changed, UpdateDlgDarkUI/dark-mode handling, EndModal with a result id"
]
},
{
"id":10,
"name":"macos-frozen-ui",
"prompt":"On macOS, sometimes after dragging the handle of the new ratio bar widget inside a dialog in OrcaSlicer, the whole app stops responding to mouse clicks — even the window's close button — but Cmd+S still saves, timers keep running and the 3D view still repaints. Windows is fine. What is going on, how do we confirm it, and how should widget code like this be written?",
"files":[],
"assertions":[
"Identifies a leaked wx mouse capture (CaptureMouse without matching ReleaseMouse) as the cause, explicitly not a deadlock/hang",
"Explains why only the mouse is dead: on macOS wx routes all mouse events to the capturing window while key events are unaffected",
"Explains why macOS differs: wxEVT_MOUSE_CAPTURE_LOST is not delivered on macOS (wxOSX never generates it; the docs' @onlyfor{wxmsw} is stale, wxGTK sends it too) so nothing ever unwinds the leaked capture",
"Fix: guard CaptureMouse with !HasCapture() and ReleaseMouse with HasCapture(), release on every exit path (mouse-up independent of drag flags, capture-lost handler, before the dialog closes/destructor)",
"Handles wxEVT_MOUSE_CAPTURE_LOST (required by the wx contract wherever it can fire) without recapturing",
"Gives a way to confirm/locate it (e.g. sampling the process / checking GetCapture(), or auditing every CaptureMouse call site)"
]
},
{
"id":11,
"name":"review-planted-wx-defects",
"prompt":"Review this proposed OrcaSlicer change before it is merged: the patch at .claude/skills/orca-wxwidgets/evals/files/filament-notes.patch adds a non-modal 'Filament notes' dialog (src/slic3r/GUI/FilamentNotesDialog.hpp/.cpp). It is not applied to the tree; read it from that path and use the repository for context. Report every correctness, cross-platform (Windows/macOS/Linux GTK, X11 and Wayland), threading, lifetime, DPI, dark-mode and wxWidgets API-misuse problem you find, each with its consequence and the fix. Do not report style nits.",
"files":[
"files/filament-notes.patch"
],
"assertions":[
"D1: Flags FromDIP() called on the SwatchPreview object inside its own base-class initializer (before wxPanel is constructed) as invalid/UB, and fixes it (parent->FromDIP / static wxWindow::FromDIP(sz, parent) / SetMinSize after construction)",
"D2: Flags the mouse-capture handling: CaptureMouse unguarded, ReleaseMouse gated on m_dragging rather than HasCapture(), and no wxEVT_MOUSE_CAPTURE_LOST handler — with the consequence (leaked capture freezes mouse input on macOS; wx's capture asserts are compiled out in Orca, so elsewhere the misuse fails silently) and the guarded pattern",
"D3: Flags hit-testing with ScreenToClient(wxGetMousePosition()) in the motion handler (unreliable on Wayland, stale vs the event) and fixes with e.GetPosition()",
"D4: Flags drawing the hover outline through wxClientDC outside the paint handler (no effect on macOS or GTK3 Wayland; on MSW and X11 it races and is overwritten by the next paint) and fixes by storing m_hover and Refresh()/RefreshRect() + drawing in the paint handler",
"D5: Flags the hard-coded wxColour(\"#009688\") pen as not dark-mode aware (should go through StateColor::darkModeColorFor or a dark-mode branch, re-derived on theme change)",
"D6: Flags SetValue() inside the wxEVT_TEXT handler (SetValue emits wxEVT_TEXT again: recursion/re-entrancy and caret reset) and fixes with ChangeValue() (plus an insertion-point restore or guard)",
"D7: Flags the wxEVT_KILL_FOCUS handler that does not call e.Skip() (breaks native focus handling of the text control)",
"D8: Flags that the wxStaticBoxSizer's child wxTextCtrl is created with the dialog as parent instead of box->GetStaticBox()",
"D9: Flags the raw pixel size wxSize(400, 160) (needs FromDIP)",
"D10: Flags m_notes->GetValue().ToStdString() as lossy for non-ASCII text and fixes with into_u8()/ToUTF8()",
"D11: Flags `delete this` in the wxEVT_CLOSE_WINDOW handler and fixes with Destroy() (deferred deletion after pending events)",
"D12: Flags SetSizer()+Layout() on the top-level dialog without SetSizerAndFit/SetSizeHints/Fit (no initial size / min size; GTK collapse) per the project rule",
"D13: Flags m_status->SetLabel() called on the worker thread (GUI call off the main thread)",
"D14: Flags the worker's wxPostEvent(this, evt) with a wxString payload: posting a wxString-carrying event from a worker thread is unsafe (use wxQueueEvent with a heap event / CallAfter), and `this` may be destroyed because the destructor detaches the thread — needs an alive flag/join/cancellation",
"D15: Flags that wxEVT_MENU is bound on m_more_btn although PopupMenu() is called on the dialog (menu events go to the invoking window and propagate up, never to a child), and that Bind runs on every popup (accumulating handlers); fix: bind on the dialog once or use GetPopupMenuSelectionFromUser",
"D16: Flags the heap wxTimer that is never stopped or deleted (leak; a pending tick can fire into the destroyed dialog) and fixes with Stop()+delete in the destructor or a by-value member",
"D17: Flags that start_sync() move-assigns m_sync_thread while the previous std::thread is still joinable (never joined or detached; a finished thread stays joinable), so the second sync ('Reload from cloud', live once the D15 routing is fixed) calls std::terminate; fix: no reassigned thread member (a detached worker per request behind the D14 alive flag) or detach/join the previous thread before reassigning, and/or refuse a reload while a sync is in flight",
"D18: Flags that SwatchPreview never releases the capture if it is destroyed while holding it (no guarded ReleaseMouse() in its destructor or on the dialog's close path): e.g. Esc, which DPIDialog maps to Close(), pressed with the button held on the strip destroys the panel while it is still on wx's capture stack (the 'Destroying window before releasing mouse capture' assert, or in Orca's assert-free build a dangling capture that freezes or crashes mouse input, notably on macOS); fix: drop the capture, or if (HasCapture()) ReleaseMouse() in the destructor / before close",
"Does not report false defects about correct code (e.g. DialogButtons' untranslated labels, wxBG_STYLE_PAINT + wxAutoBufferedPaintDC, Bind with lambdas)"
]
},
{
"id":12,
"name":"review-subtle-wx-defects",
"prompt":"Review this proposed OrcaSlicer change before it is merged: the patch at .claude/skills/orca-wxwidgets/evals/files/preset-note.patch adds a sidebar 'note chip' widget and a modal 'Preset note' editor dialog (src/slic3r/GUI/PresetNoteDialog.hpp/.cpp). It is not applied to the tree; read it from that path and use the repository and the wxWidgets source for context. Report every correctness, cross-platform, i18n, DPI, dark-mode, lifetime and wxWidgets/Orca API-misuse problem you find, each with its consequence and the fix. Do not report style nits.",
"files":[
"files/preset-note.patch"
],
"assertions":[
"P1: Flags the StateColor built with the Normal entry first: colorForStates returns the first matching entry and Normal (mask 0) matches every state, so the Hovered/Pressed colours never show; fix: list Pressed, Hovered, then Normal last",
"P2: Flags the wxEVT_SIZE lambda taking wxSizeEvent by value: Skip() on the copy does not reach the original, so the event counts as handled and default/base size handling stops; fix: take the event by reference",
"P3: Flags that the capture-lost handler routes through on_left_up and therefore fires EVT_NOTE_CHIP_CLICKED (commits a click) when capture is lost; per the wx contract capture loss must cancel the operation: reset m_pressed/visual state without emitting the event and without recapturing",
"P4: Flags _(\"Preset note\") as never extracted for translation (Orca's xgettext keywords are L/_L/_u8L/..., not _), fix: _L",
"P5: Flags wxBitmapBundle::FromSVGFile as unavailable in Orca's wx build (NanoSVG off / no wxHAS_SVG — does not compile) and replaces it with Orca's ScalableBitmap/create_scaled_bitmap/ScalableButton icon loading by name",
"P6: Flags Bind(wxEVT_TEXT_ENTER, ..., m_input->GetId()) on the dialog as never firing: TextInput re-dispatches TEXT_ENTER only to the wrapper (ProcessEventLocally, no propagation to parents) and the inner control's handler does not Skip; fix: bind on m_input (the wrapper) or on GetTextCtrl()",
"P7: Flags RichMessageDialog::SetYesNoLabels as having no effect in Orca (labels are stored but never applied to the buttons), fix: MsgDialog::SetButtonLabel(wxID_YES/wxID_NO, ...) or a dialog with custom buttons",
"P8: Flags EndModal(wxCANCEL): wxCANCEL is a style flag, not the return code wxID_CANCEL, so ShowModal returns the wrong value; fix: EndModal(wxID_CANCEL)",
"P9: Flags that restore_original() runs only in the Cancel button handler: ESC (DPIDialog's char hook -> Close()) and the close box end the dialog with wxID_CANCEL without running the custom Button's handler, so the draft is not cleared; fix: do cleanup on every non-OK exit (after ShowModal returns, or in a close/EndModal path)",
"P10: Flags m_preview->SetLabel(...) followed by Wrap(FromDIP(360)) on every update: in wx 3.3.2 Wrap() is a no-op when the width equals the last wrap width, so updated labels stay unwrapped; fix: Wrap(-1) then Wrap(w), or wxST_WRAP / Orca Label with LB_AUTO_WRAP",
"P11 (not a defect): Does not claim that m_options->Enable(...) leaves the ::CheckBox or its wxStaticText label looking enabled: ::CheckBox is a native wxBitmapToggleButton, so the ancestor's disable greys it on every port without calling its Enable() override (MSW: NotifyWindowOnEnableChange -> DoEnable -> EnableWindow, and the owner-drawn button paints its SetBitmapDisabled art; GTK3: native insensitivity, and the button's wxGtkImage draws a greyed copy of its current bitmap while !IsEnabled(); macOS: setEnabled:NO, and AppKit dims the image), and wxStaticText greys natively. Suggesting pin->Enable() to get the designed disabled art on GTK is at most a nit",
"P12: Flags that the patch registers neither new file: they are not in SLIC3R_GUI_SOURCES in src/slic3r/CMakeLists.txt (never compiled), and PresetNoteDialog.cpp is not in localization/i18n/list.txt (run_gettext scans only the listed files, so its _L strings never reach OrcaSlicer.pot); fix: add both entries",
"Does not report false defects about correct code: the weak_ptr-guarded CallAfter is not a use-after-free (alive is checked before `this` is used), GetSizer()->SetSizeHints(this) in on_dpi_changed is fine, DialogButtons' untranslated labels are correct"
2. Theme with palette colours, not `wxSYS_COLOUR_*`: on Windows wx serves its dark palette whenever the
system app mode is dark, and macOS system colours are dynamic so the dark map cannot key them. → [System colours](#system-colours-and-appearance)
3. Write colours as `wxColour("#RRGGBB")` or `wxColour(r, g, b)`. A packed integer `wxColour(0xRRGGBB)` is
read as `0x00BBGGRR`; `StateColor` integers are the opposite (`0xRRGGBB`). → [wxColour](#wxcolour)
4. Give every `wxPanel`/`wxScrolledWindow` you create an explicit palette background (`*wxWHITE`, `#F8F8F8`, …),
and set it **before** creating Orca widgets inside it. → [Window colours](#window-colours-inheritance-and-native-control-limits), [Update*DarkUI](#the-updatedarkui-walk)
5. Background colour is never inherited; foreground only at creation, only from the immediate parent, only for
`wxControl`-like classes. Set colours on the window that shows them. → [Window colours](#window-colours-inheritance-and-native-control-limits)
6. In a `StateColor`, list specific states first and `Normal` last; negate with `Not*`/`Disabled`, never `~X`. → [StateColor](#statecolor)
7. A literal light colour given to `SetForegroundColour`/`SetBackgroundColour`/`wxPen`/`wxBrush` goes through
`StateColor::darkModeColorFor()` when it is a `gDarkColors` key, else branch on `dark_mode()`. → [Re-applying colours](#runtime-theme-switch-and-re-applying-colours)
8. End every dialog constructor (after all children exist, before `ShowModal()`) with `wxGetApp().UpdateDlgDarkUI(this)`;
frames use `UpdateFrameDarkUI`; a subtree built after the app-wide pass ends with `UpdateDarkUIWin(this)`.
`UpdateDarkUI(win)` themes one window only. → [Update*DarkUI](#the-updatedarkui-walk)
9. Apply deliberate non-palette colours **after** the walk, and again on theme change; the dark walk rewrites every
and mistakenly using wxColour ctor from bool … doesn't compile any longer").
| Fact | Cite |
|---|---|
| Packed integers are `0x00BBGGRR` (`0xAABBGGRR` for `SetRGBA`): "Notice the right-to-left order of components!" `wxColour(0x009688)` is R=0x88 G=0x96 B=0x00, not Orca teal. Only symmetric greys (`0xEEEEEE`) read the same both ways. wx writes its own MSW dark palette this way (`wxColour(0x9e5315)` is a blue). | `interface/wx/colour.h:179-183`; `include/wx/colour.h:77-83`; `src/msw/darkmode.cpp``wxDarkModeSettings::GetColour` |
| `Set(const wxString&)` accepts colour-database names, CSS `rgb(r,g,b)`/`rgba(r,g,b,a)` (case-insensitive) and `#` + 6 hex digits; returns `false` on failure. **[source]** the parser also takes `#rgb`, `#rgba`, `#rrggbbaa`, but the documented form (and XRC, "but not "#rgb"") is `#RRGGBB` — write that. | `interface/wx/colour.h:288-301`; `src/common/colourcmn.cpp``FromString`; `docs/doxygen/overviews/xrc_format.h:229` |
| The string ctor is `{ Set(colourName); }`: a typo silently yields `IsOk() == false`. An invalid colour passed to `SetBackgroundColour`/`SetForegroundColour` means "reset to the default colour". | `include/wx/colour.h:235`; `interface/wx/colour.h:239-243`; `interface/wx/window.h:2438-2439` |
| 3.3 changed `wxColourDatabase` to CSS values ("GREEN" is `#008000` in the CSS scheme, `#00ff00` traditionally; wxGTK already used CSS); `UseScheme()` reverts. **[source]** stock objects did not change: `*wxGREEN` is still (0,255,0), so `*wxGREEN != wxColour("green")`. | `docs/changes.txt:31-33`; `interface/wx/gdicmn.h:999-1013`; `src/common/gdicmn.cpp:516`, `:805-807` |
| `wxTransparentColour` = `wxColour(0,0,0,wxALPHA_TRANSPARENT)`: valid, black, alpha 0. `IsTransparent()/IsOpaque()/IsTranslucent()` are new in 3.3.1. `Alpha()` returns `wxALPHA_OPAQUE` "on platforms where alpha is not yet supported". | `include/wx/colour.h:31`; `interface/wx/colour.h:110-114`, `:260-282` |
| `ChangeLightness(ialpha)`: 0 = black, 100 = unchanged, 200 = white; returns a copy. **[source]** values are clamped to 0..200 and the result is built as `wxColour(r,g,b)` — alpha is dropped. | `interface/wx/colour.h:365-377`; `src/common/colourcmn.cpp``wxColourBase::ChangeLightness` |
| `MakeDisabled(brightness = 255)` "modifies the object in place and returns the object itself". **[source]** each channel becomes `brightness + 0.4·(c − brightness)` — it keeps 40% of the colour and moves 60% toward `brightness`, so on a dark background the default makes a "disabled" colour light. | `interface/wx/colour.h:336-342`; `src/common/colourcmn.cpp``MakeDisabled`, `AlphaBlend` |
Alpha in window colours: **[source]** MSW brushes are `CreateSolidBrush(COLORREF)` (alpha dropped) and pen/brush
transparency is style-based, so a solid brush of a transparent colour paints **black** through GDI. Use
`*wxTRANSPARENT_BRUSH` or `wxGCDC`/`wxGraphicsContext` (`src/msw/brush.cpp:187`; `include/wx/brush.h:60-68`).
Background styles and transparent windows: see `references/painting-custom-widgets.md`.
**OrcaSlicer.** Build colours from hex strings or RGB triples, never from names. Text over user/filament
swatches is chosen by luminance (`clr.GetLuminance() < 0.51 ? *wxWHITE : *wxBLACK`, `wxExtensions.cpp` swatch
helpers, `PresetComboBoxes.cpp`). A packed `wxColour(0x…)` literal is harmless only for a symmetric grey
(`wxColour(0xEEEEEE)`); integer literals in `StateColor` contexts are RGB.
**Pitfalls**
- **Rule:** Never write a packed integer `wxColour`; keep `StateColor` integers as they are.
**Why:**`wxColour(unsigned long)` is BGR; `StateColor(unsigned long)`/`append(unsigned long, …)` byte-swap the value
so it is RGB (`Widgets/StateColor.cpp``StateColor::append`). "Fixing" one to look like the other inverts R and B.
```cpp
// Wrong: R=0x88 G=0x96 B=0x00 (olive), not #009688
label->SetForegroundColour(wxColour(0x009688));
// Right:
label->SetForegroundColour(wxColour("#009688"));
// Also right (StateColor integers are 0xRRGGBB, alpha byte 0 = opaque):
| `IsDark()` | "checks the appearance of the current application and not the other applications on the system, so under MSW … will return false even if dark mode is used system-wide unless the application opted in using dark mode using wxApp::MSWEnableDarkMode()" (`:332-346`). An incompatible 3.3 change (`docs/changes.txt:71-73`). |
| `AreAppsDark()` (3.3.0) | system-wide app dark mode "even if it's not enabled for this particular application"; same as `IsDark()` off MSW (`:308-321`). |
| `IsSystemDark()` (3.3.0) | the "Windows mode", which can differ from the "app mode" (`:348-358`). |
| `IsUsingDarkBackground()` | luminance fallback, "generally not very useful to call directly" (`:360-370`). |
| `GetName()` | "only implemented for macOS", e.g. "NSAppearanceNameAqua"; empty elsewhere (`:323-330`). |
| `wxSystemSettings::SelectLightDark(light, dark)` (3.3.0) | "just a convenient helper using wxSystemAppearance::IsDark()" (`:460-473`); literally `GetAppearance().IsDark() ? dark : light` (`include/wx/settings.h:234-237`). |
**Platforms** (all **[source]**):
| Port | `IsDark()` | `GetColour()` |
|---|---|---|
| MSW | `wxMSWDarkMode::IsActive()` ‖ luminance fallback (`src/msw/settings.cpp:431-442`). Under `DarkMode_Auto`, `IsActive()` is uxtheme's `ShouldAppsUseDarkMode()` (`src/msw/darkmode.cpp` `ShouldUseDarkMode`) — the **system app mode**, not the app's own choice. `AreAppsDark()/IsSystemDark()` read `AppsUseLightTheme`/`SystemUsesLightTheme` from the registry. | `GetSysColor`, except GRIDLINES→BTNFACE, LISTBOXTEXT→WINDOWTEXT, LISTBOXHIGHLIGHT→HIGHLIGHT, LISTBOX→WINDOW, MENUBAR→MENU unless flat menus (`src/msw/settings.cpp:99-148`). LISTBOXHIGHLIGHTTEXT maps only to a raw index `GetSysColor` does not define — use HIGHLIGHTTEXT. **When wx dark mode is active every index is answered by `wxDarkModeSettings::GetColour()` first**; indices it leaves invalid (GRAYTEXT, 3DLIGHT, borders, …) fall back to the light `GetSysColor` value. |
| macOS | `[NSApp effectiveAppearance]` best match is DarkAqua (`src/osx/cocoa/settings.mm` `IsDark`). | WINDOW/LISTBOX = `controlBackgroundColor`, BTNFACE = `windowBackgroundColor` (≥10.14), caption/border/MENU/MENUBAR = `windowFrameColor`, BTNTEXT/WINDOWTEXT/MENUTEXT/CAPTIONTEXT/INACTIVECAPTIONTEXT/INFOTEXT/LISTBOXTEXT = `controlTextColor` (GRAYTEXT = `disabledControlTextColor`, HIGHLIGHTTEXT = `selectedTextColor`), GRIDLINES = `gridColor`, INFOBK/APPWORKSPACE = `windowBackgroundColor` (commented "bogus"), HOTLIGHT = `linkColor` (`src/osx/cocoa/settings.mm` `GetColour`). The result wraps a **dynamic NSColor**: components resolve at call time under the effective appearance (`src/osx/cocoa/colour.mm`), and as a view background it adapts by itself. |
| GTK3 (the default Linux build) | `IsUsingDarkBackground()`: luminance(WINDOWTEXT) − luminance(WINDOW) > 0.2 (`src/common/settcmn.cpp:96-112`); `AreAppsDark()/IsSystemDark()` = `IsDark()` (`:71-84`). wxGTK3 follows the freedesktop portal `org.freedesktop.appearance` `color-scheme` (GNOME's dark style) by setting `gtk-application-prefer-dark-theme` and stripping a `-dark`/`-Dark` theme-name suffix; no portal when `GTK_THEME` is set (`src/gtk/settings.cpp:251-334`, `:369-400`, `:1426-1450`). | From synthetic `GtkStyleContext`s (button, textview, treeview, headerbar, tooltip, menu), **cached** in `gs_systemColorCache` until "notify::gtk-theme-name" or a colour-scheme change (`src/gtk/settings.cpp:728-880`). |
| GTK2 (opt-out build, `-DDEP_WX_GTK3=OFF`) | as GTK3 (luminance of the GTK2 theme); no portal/colour-scheme support (`#ifdef__WXGTK3__`). | `GtkStyle` of helper widgets (`src/gtk/settings.cpp:913ff`). |
| Wayland | no Wayland-specific branch in colour/appearance code. | — |
MSW dark palette (`wxDarkModeSettings::GetColour`, `src/msw/darkmode.cpp:300-370`; "not documented and are subject
to change", `interface/wx/msw/darkmode.h:74-84`): WINDOW/LISTBOX/INFOBK/APPWORKSPACE/ACTIVECAPTION `0x202020`;
the *TEXT indices `0xe0e0e0` except INACTIVECAPTIONTEXT `0xaaaaaa` (GRAYTEXT is left invalid); BTNFACE/GRIDLINES `0x333333`; MENU/INACTIVECAPTION `0x2b2b2b`; MENUBAR/LISTBOXHIGHLIGHT
`0x626262`; HIGHLIGHT/MENUHILIGHT `0x9e5315` (a blue — packed BGR); HOTLIGHT `0xe48435`.
**OrcaSlicer.** Orca does not theme with system colours or with `wxVisualAttributes`/`GetClassDefaultAttributes()`
(the stock advice for matching native controls): it uses a fixed palette with a hand-written dark twin map
(`StateColor`), because the exact-match map needs deterministic RGB that dynamic macOS colours and the MSW dark
palette do not give, and because Windows needs an app-level live toggle wx cannot do (see
[wxMSW dark mode](#wxmsw-dark-mode-mswenabledarkmode-setappearance-wxdarkmodesettings)). `GUI_App::init_label_colours`
still reads `wxSYS_COLOUR_WINDOWTEXT`/`wxSYS_COLOUR_WINDOW` for two light-mode values — on Windows with a dark system
and Orca light those come from wx's dark palette. macOS: Orca's wx is patched to read NSColor components in sRGB
instead of `NSCalibratedRGBColorSpace` (`deps/wxWidgets/0001-macos-use-srgb-colour-components.patch`, commit
a7775296b0 "Fix macOS custom color accuracy"), so RGB read back from native colours matches the expected hex.
**Pitfalls**
- **Rule:** Never branch on wx's appearance in Orca GUI code.
**Why:** Orca calls `MSWEnableDarkMode(DarkMode_Auto)`, so on Windows `IsDark()`/`SelectLightDark()` follow the
system app mode — Windows dark + Orca light reports dark.
```cpp
// Wrong:
auto c = wxSystemSettings::SelectLightDark(wxColour("#FFFFFF"), wxColour("#2D2D31"));
// Right:
auto c = StateColor::darkModeColorFor(wxColour("#FFFFFF")); // #FFFFFF is a gDarkColors key
auto d = wxGetApp().dark_mode() ? wxColour("#EFEFF0") : wxColour("#333333"); // unmapped colour
- **Rule:** Theme with palette colours, not `wxSystemSettings::GetColour(wxSYS_COLOUR_*)`.
**Why:** MSW returns wx's dark palette whenever the system app mode is dark, regardless of Orca's setting
(`src/msw/settings.cpp:99-107`); on macOS the value is a dynamic colour no `gDarkColors` key matches; on GTK it
is whatever the theme says. Do not cache macOS system-colour RGB across a theme change.
- **Rule:** Raw `wxButton` code that branches on `__WXMAC__` to `wxSYS_COLOUR_BTNFACE`/`BTNTEXT` only keeps the text
in the system colour: NSButton ignores the background colour anyway (see
[native limits](#window-colours-inheritance-and-native-control-limits)). New code uses Orca `Button`.
## wxEVT_SYS_COLOUR_CHANGED
**Contract.** "generated when the user changes the colour settings or when the system theme changes (e.g. automatic
dark mode switching on macOS)". "The default event handler for this event propagates the event to child windows,
since the system events are only sent to top-level windows. If intercepting this event for a top-level window,
remember to either call wxEvent::Skip() on the event, call the base class handler, or pass the event on to the
window's children explicitly" (`interface/wx/event.h:1942-1966`). **[source]** the default handler
(`wxWindowBase::OnSysColourChanged`, `src/common/wincmn.cpp:3011-3027`) sends a fresh event to every
**non-top-level** child and calls `Refresh()`; a child whose own handler does not `Skip()` stops it for its subtree.
**Who sends it** (all **[source]**):
| Port | Trigger |
|---|---|
| MSW | `WM_SYSCOLORCHANGE`, and `WM_SETTINGCHANGE` with "ImmersiveColorSet" (light/dark/accent switch) → `HandleSysColorChange()` on each TLW (`src/msw/window.cpp:3542`, `:5092`, `:5197-5200`). The default MSW handler re-sends a real `WM_SYSCOLORCHANGE` to native children (`wxWindowMSW::OnSysColourChanged`, `:5269-5293`); `wxFrame` also resets its background to `wxSYS_COLOUR_APPWORKSPACE` if `!UseBgCol()` and re-themes the menubar (`src/msw/frame.cpp:476-503`). |
| macOS | per-NSWindow KVO on `effectiveAppearance`, **and** `windowDidChangeBackingProperties` when the window's colour space changes (dragging to a display with another profile) — no theme change involved (`src/osx/cocoa/nonownedwnd.mm:684-740`). Popups are `wxNonOwnedWindow`s and get it too. No defined order across windows. |
| GTK3 | every TLW connects "notify::gtk-theme-name" with `g_signal_connect_after` so the colour cache is cleared before user handlers run (`src/gtk/toplevel.cpp:555-563`, `:951-955`); the portal colour-scheme handler `DoUpdateColorScheme` also loops over `wxTopLevelWindows` (`src/gtk/settings.cpp:251-334`). A dark switch that also renames the theme delivers the event **twice** per TLW. |
- **Rule:** Do not call `wxTheApp->SetAppearance(...)` in Orca.
**Why:** MSW returns `CannotChange` (Orca already called `MSWEnableDarkMode`), GTK2 fails, and macOS pins
`NSApp.appearance` while `GUI_App::dark_mode()` keeps reading the system `AppleInterfaceStyle` — the two would
disagree.
- **Rule:** Do not show wx's TaskDialog-based or common dialogs where dark mode matters; use `MessageDialog` and
friends (`MsgDialog.hpp`, see `references/windows-dialogs.md`).
## Window colours, inheritance and native-control limits
**Contract** (`interface/wx/window.h`):
| Call | Effect |
|---|---|
| `SetBackgroundColour(c)` | "may not affect the entire control and could be not supported at all depending on the control and platform"; does not refresh ("you may wish to call wxWindow::ClearBackground or wxWindow::Refresh"); "will disable attempts to use themes for this window"; returns `false` if the colour was already set (`:2427-2461`). |
| `SetOwnBackgroundColour(c)` | same, "but prevents it from being inherited by the children" (`:2588-2593`). |
| `SetForegroundColour(c)` | "not all native controls support changing their foreground colour so this method may change their colour only partially or even not at all" (`:2561-2585`). |
| `UseBgCol()`/`UseBackgroundColour()`, `UseForegroundColour()` | whether a colour was set for this window (`:2600-2632`). |
| `InheritsBackgroundColour()`/`InheritsForegroundColour()` | the inheritable flag (`:2595-2639`). |
| `ShouldInheritColours()` | "base class version returns false, but … overridden in wxControl where it returns true" (`:2645-2653`). |
| `InheritAttributes()` | called during creation; a child takes an attribute only if the parent set it explicitly (not via `SetOwn*`) and the child did not (`:4052-4075`). |
| `GetClassDefaultAttributes(variant)` → `wxVisualAttributes{font, colFg, colBg}` | "colBg may be wxNullColour if the controls background colour is not solid"; "All of them may be invalid if it was not possible to determine the default control appearance" (`:135-150`, `:4245-4273`). |
**[source] facts the docs do not spell out** (`src/common/wincmn.cpp`):
- **Background is never inherited.** The background branch of `InheritAttributes()` is `#if 0` ("inheriting (solid)
background colour is wrong as it totally breaks any kind of themed backgrounds", `:1524-1553`). Only font and
foreground are copied, only at creation time (`InheritAttributes()` runs from the ports' creation code), only
from the **immediate** parent.
- `ShouldInheritColours()` is false for `wxWindow`/`wxPanel` (`include/wx/window.h:1682`), `wxAnyButton`
`wxStaticBitmap`, `wxHyperlinkCtrl`, book controls, MSW `wxRadioButton`/`wxSlider` always do.
- macOS: a bare window with the default `wxBG_STYLE_ERASE` is cleared with its **own** `GetBackgroundColour()`
(`wxWindowMac::MacDoRedraw`, `src/osx/window_osx.cpp:1950-1983`; the `wxWindowDC` background is
`GetBackgroundColour()`, `src/osx/carbon/dcclient.cpp:89`) — the class default `windowBackgroundColor`, a dynamic
system grey.
- GTK3: a bare `wxPanel` is theme-enabled (`src/common/panelcmn.cpp:100`) and renders its own GTK style-context
background (`gtk_render_background` in `wxWindowGTK::GTKSendPaintEvents`, `src/gtk/window.cpp`), which depends on
the theme.
This is why "dark-theme bugs that only show on macOS" exist: MSW hides an unset panel behind the ancestor's brush,
macOS paints the system grey.
**Native controls that ignore colours** (**[source]**):
| Port | Behaviour |
|---|---|
| MSW | `wxButton`/`wxToggleButton` `SetBackgroundColour`/`SetForegroundColour` switch the native button to `BS_OWNERDRAW` (`src/msw/anybutton.cpp` `wxAnyButton::MakeOwnerDrawn`, `:1324-1390`) — the colour works but native theming is gone. `wxCheckBox`/`wxRadioButton` foreground makes them owner-drawn when themes are active (`src/msw/control.cpp` `MSWMakeOwnerDrawnIfNecessary`); 3.3.2 fixed checkbox accessibility in that mode (`docs/changes.txt:299-300`). Static-text-like children paint the ancestor brush only if that ancestor's background is inheritable. |
| macOS | `SetBackgroundColour` reaches the NSView only if it `respondsToSelector:setBackgroundColor:` and the style is not `wxBG_STYLE_TRANSPARENT` (`src/osx/cocoa/window.mm:3514-3531`): **NSButton (`wxButton`) ignores the background colour**. Foreground uses `setTextColor:` when the view has it (`:3885-3887`); stock NSButton has none. `wxStaticText` is an NSTextField with `setDrawsBackground:NO` — its background is never painted (`src/osx/cocoa/stattext.mm:154`). `wxTextCtrl`'s `setEnabled:` resets the text colour: multi-line (`wxNSTextView`) always, to `controlTextColor`/`disabledControlTextColor`; single-line (`wxNSTextField`) when it does not draw its background, to `controlTextColor`/`secondarySelectedControlColor` (`src/osx/cocoa/textctrl.mm`). |
| GTK3 | `SetBackgroundColour/SetForegroundColour/SetFont` install a per-widget CSS provider `*{color:..;background:..;font:..}` at `GTK_STYLE_PROVIDER_PRIORITY_APPLICATION` (`src/gtk/window.cpp` `wxWindowGTK::GTKApplyStyle`); `wxButton`/`wxCheckBox` apply it to their inner label too (`src/gtk/button.cpp:325-335`, `src/gtk/checkbox.cpp:233-237`). User CSS (priority USER) can still override. |
| GTK2 | colours applied with `gtk_widget_modify_style` — pixmap-engine themes may ignore them. |
**OrcaSlicer — `Utils/MacDarkMode.mm`** installs process-wide Objective-C categories/swizzles that affect *every*
window, so stock-wx advice about NSTextField/NSButton colours differs in Orca:
- every `NSTextField` is created with `drawsBackground = false` (category `NSTextField (drawsBackground)`), so with
the `textctrl.mm` rule above a custom text colour is lost on every `Enable()` and the background is not painted;
- `NSButton (NSButton_Extended)` adds `textColor`/`setTextColor:` (via the attributed title), which is what makes
`wxButton`/`wxCheckBox` foreground colours work on macOS; `setBezelStyle:` forces bordered (except shadowless
square); focus rings are off for NSButton and NSTextField; `NSTableHeaderCell`'s font is forced to `Label::sysFont(13)`;
- the main-frame title text colour is forced (`set_title_colour_after_set_title`; `NSTextField (textColor)` swizzle
for that one field), and `set_miniaturizable` makes the titlebar transparent over a dark NSWindow background.
Orca's widgets work with these limits: `Label` sets keyed colours (`SetForegroundColour("#262E30")`, background =
`StaticBox::GetParentBackgroundColor(parent)`) so the walk can map them — commit d408db2fde replaced `wxStaticText`
with `Label` to fix Linux dark mode this way; `TextInput::Enable` re-applies the inner control's background and
foreground from its `StateColor`s after enabling; `StaticBox::Create` copies the parent background into its own wx
background (a `StaticBox` parent's `background_color.defaultColor()`, the midpoint of a gradient, else
`parent->GetBackgroundColour()`), used for the corners outside the rounded rectangle.
**Pitfalls**
- **Rule:** Give every `wxPanel`/`wxScrolledWindow` you create an explicit palette background (typically
`SetBackgroundColour(*wxWHITE)`), even when white "already looks right".
**Why:** the walk re-keys a background only if it is a `gDarkColors` key. A bare panel paints its class default
(`wxSYS_COLOUR_BTNFACE`; on macOS the appearance-dependent `windowBackgroundColor`), which is never a key, so it
shows as a system-grey patch against Orca's palette in either theme. MSW hides this by painting the ancestor's
explicit brush, so it surfaces on macOS. An explicit palette colour is the key that lets the walk re-map it in
both directions.
```cpp
wxPanel* p = new wxPanel(parent); // Wrong: unkeyed; macOS shows system grey, not #2D2D31 / #FFFFFF
wxPanel* p = new wxPanel(parent);
p->SetBackgroundColour(*wxWHITE); // Right: #FFFFFF is a key → #2D2D31 in dark
- **Rule:** Negate with the `Not*` enumerators, never `~`.
**Why:** `(int) Hovered | NotFocused` = `0x40008` means "hovered and not focused"; `(int) Hovered | ~Focused` =
`0xFFFFFFFB` sets every on-bit and never matches.
- **Rule:** Do not pass data colours (filament/extruder/user colours) through `darkModeColorFor` or leave them as
window backgrounds the walk visits; paint them in a paint handler or re-set them after the walk.
**Why:** a white filament `#FFFFFF` maps to `#2D2D31`.
- **Rule:** Never use a dark-twin value as a light-mode colour (`#FFFFFE`, `#D9D9D9`, `#909090`, `#808080`, …).
**Why:** the light walk rewrites it through the reverse map.
## Orca dark-mode state
| State | Set by | Read by |
|---|---|---|
| app_config `dark_color_mode` ("1"/"0") | Windows: the Preferences checkbox (`PreferencesDialog::create_item_darkmode`, `#ifdef _WIN32`); `AppConfig::set_defaults` sets `"0"` when empty (`#ifdef _WIN32`). macOS/Linux: overwritten from `GetAppearance().IsDark()` at startup (`GUI_App::on_init_inner`, `#ifndef__WINDOWS__`) and by `update_dark_config()` on every SYS_COLOUR event. | `GUI_App::dark_mode()` (non-macOS); `Plater::priv::on_change_color_mode` and other GL/web code that reads the key directly — new code calls `dark_mode()` instead |
| `GUI_App::m_is_dark_mode` | `GUI_App::Update_dark_mode_flag()` (= `dark_mode()`) | `GUI_App::UpdateDarkUI` (the walk) only |
| `StateColor` file-static `gDarkMode` | `StateColor::SetDarkMode`, called only from `GUI_App::init_label_colours()` | `darkModeColorFor`, `colorForStates` |
| NppDarkMode `g_darkModeEnabled` (Windows) | `NppDarkMode::InitDarkMode` / `SetDarkMode` | title bars, explorer theme, scrollbars, DVC header |
| wx's own MSW mode | `MSWEnableDarkMode(DarkMode_Auto)` → follows the **system app mode** | wx internals, `IsDark()`, `GetColour()` |
`GUI_App::dark_mode()` (static, recomputed every call) is the only query GUI code uses:
- macOS: `wxPlatformInfo::Get().CheckOSVersion(10, 14) && mac_dark_mode()` (10.12/10.13 gave false positives);
`mac_dark_mode()` (`Utils/MacDarkMode.mm`) reads the `AppleInterfaceStyle` user default == "Dark" — the **system**
preference. The config key is ignored. wx's `IsDark()` (which feeds `update_dark_config()`) reads
`[NSApp effectiveAppearance]`; the two normally agree.
const auto& val = app_config->get("dark_color_mode");
if (val == "1") return true;
if (val == "0") return false;
return check_dark_mode(); // unset key only
```
Cite: 7d7f26ed69 (`GUI_App.cpp`, `GUI_App::dark_mode`, the non-Apple branch).
- **Rule:** Any new code path that changes the theme first sets `dark_color_mode` (Windows: `app_config->set` +
`save()`; macOS/Linux: `update_dark_config()`), then refreshes the cached state: `wxGetApp().Update_dark_mode_flag()`
(`m_is_dark_mode`, read by the walk), `wxGetApp().init_label_colours()` (StateColor's `gDarkMode` and the label
colours) and, on Windows, `wxGetApp().force_colors_update()` (NppDarkMode's `g_darkModeEnabled` via
`NppDarkMode::SetDarkMode(dark_mode())`, the main frame's title bar, and the walk flag `update_ui_from_settings()`
consumes). `dark_mode()` itself is live; wx's own MSW mode follows the system and is not set by Orca.
**Why:** they are updated at different moments; `update_dark_config()` does not touch StateColor's flag — for the
main frame that happens in `MainFrame::on_sys_color_changed` — and after startup only `force_colors_update()` calls
`NppDarkMode::SetDarkMode`.
## The Update*DarkUI walk
`GUI_App` helpers (`GUI_App.hpp/.cpp`):
| Helper | Does |
|---|---|
| `UpdateDarkUI(win, highlited = false, just_font = false)` | One window. `just_font` is unused. Skips `wxBU_AUTODRAW` buttons (`wxButton` or Orca `Button`). Windows only: a `wxButton` with id `wxID_OK`/`wxID_CANCEL` gets `wxNO_BORDER`, palette background/foreground (→ owner-drawn) and four hover/focus handlers **bound again on every call**. Then, using `m_is_dark_mode` (not `dark_mode()`): **dark** — `bg = darkModeColorFor(GetBackgroundColour())`, set only if it changed (exact key); `fg = darkModeColorFor(GetForegroundColour())`, then if ΔE(bg, fg) < 10 → LAB L = 90, if L(fg) < 45 → L = 70, and **fg is always set**; **light** — `lightModeColorFor` on background and foreground, each set only if changed. |
| `update_dark_children_ui(win)` (file-static) | recursive over `GetChildren()` at call time: a `ScalableButton` gets `ScalableButton::UpdateDarkUI()` (= `msw_rescale()`: `UpdateDarkUI(this, m_has_border)` plus re-rasterized icons), anything else `UpdateDarkUI(child)`. `GetChildren()` includes owned TLWs, so dialogs parented to the window are walked too. |
| `UpdateDarkUIWin(win)` | the walk. |
| `UpdateDlgDarkUI(dlg)` / `UpdateFrameDarkUI(frame)` | Windows: `NppDarkMode::SetDarkExplorerTheme` + `SetDarkTitleBar` on the HWND (both follow Orca's mode in both directions); then the walk. |
| `UpdateDVCDarkUI(dvc, highlited)` | **Windows-only body** (no-op on macOS/Linux): `UpdateDarkUI`, dark list header via `NppDarkMode::SetDarkListViewHeader`, header attr text colour `NppDarkMode::GetTextColor()`, `SetAlternateRowColour(m_color_highlight_default)` for `wxDV_ROW_LINES`, forces `wxBORDER_SIMPLE`. |
| `UpdateAllStaticTextDarkUI(parent)` | **Windows-only body**: `UpdateDarkUI(parent)` and `m_color_label_default` on direct `wxStaticText` children. |
Consequences: an unmapped background stays as it is; an unmapped dark foreground becomes grey (L≈70) instead of
dark-on-dark; every visited window ends with an explicit foreground (on MSW that makes raw buttons/checkboxes
owner-drawn); dark → light is lossy (lifted greys are not restored; shared twins collapse, see
[StateColor](#statecolor)); the walk only calls `Set{Background,Foreground}Colour`, so `wxPaintDC` painting is
unreachable by it. The source comment on `UpdateDarkUIWin` ("Don't use this function for Dialog contains
ScalableButtons") is stale: all three entry points run the same `update_dark_children_ui`, which already handles
`ScalableButton`; the only difference is the Windows HWND theming. Use `UpdateDlgDarkUI` for dialogs because of
that theming.
App-wide passes: `MainFrame`'s constructor ends with `UpdateDarkUIWin(this)` (on all platforms); a theme switch runs
`update_dark_children_ui(mainframe)` from `GUI_App::update_ui_from_settings`; the lazily built main-window tabs re-run
`UpdateDarkUIWin(this)` after insertion. Everything created later themes itself.
Raw wx controls do not follow Orca's mode by themselves (on Windows wx themes native controls by the *system* app
mode, see [wxMSW dark mode](#wxmsw-dark-mode-mswenabledarkmode-setappearance-wxdarkmodesettings)). When one cannot
be replaced by an Orca widget, make it dark-safe through the walk (`UpdateDarkUI(ctrl)` for a single control created
after the pass, `UpdateDlgDarkUI(dlg)` for its dialog) or with explicit `darkModeColorFor()` colours.
**Pitfalls**
- **Rule:** Theme a dialog with `UpdateDlgDarkUI`, not `UpdateDarkUI`.
**Why:** `UpdateDarkUI` touches one window; children keep light defaults and the Windows title bar stays light.
- **Rule:** Apply deliberate non-palette colours after the walk (and again on theme change).
**Why:** in dark mode the walk rewrites every visited window's foreground — white text on an accent chip becomes
mid-grey.
```cpp
wxGetApp().UpdateDlgDarkUI(this);
m_badge->SetForegroundColour(*wxWHITE); // after the walk
```
- **Rule:** Run the dialog walk once per theme state; do not call it from paint or size handlers.
**Why:** the map chains (`#FFFEFE → #D9D9D9 → #27272A`) and the Windows OK/Cancel branch stacks handlers per call.
## NppDarkMode (Windows)
Vendored Notepad++ dark-mode code in `src/slic3r/GUI/dark_mode.cpp/.hpp` and `src/slic3r/GUI/dark_mode/*.hpp`,
compiled only on Windows. Namespace `NppDarkMode`:
| Function | Does |
|---|---|
| `InitDarkMode(bool dark, bool sys_menu)` | loads the uxtheme entry points, records the system-menu setting, `SetDarkMode(dark)`. |
| `SetDarkMode(bool)` | sets `g_darkModeEnabled`; `AllowDarkModeForApp(dark)` → `SetPreferredAppMode(ForceDark|ForceLight)` (Windows 1903+) or `AllowDarkModeForApp` (1809); flushes menu themes when the system-menu setting is on; scrollbar fix. |
| `SetDarkTitleBar(HWND)` | allows dark for the window per `IsEnabled()`, refreshes the title-bar colour, applies the explorer theme. |
| `SetDarkListViewHeader(HWND)` | dark `ItemsView` theme on a list header. |
| `GetTextColor()` | `0xF0F0F0` when enabled, else `wxSYS_COLOUR_WINDOWTEXT` — which is wx's dark-palette `0xe0e0e0` when the Windows app mode is dark and Orca is light **[source]**. |
Unlike wx's MSW dark mode it switches live: every function follows `g_darkModeEnabled` in both directions, so
re-running them on existing HWNDs re-themes them. `GUI_App::force_colors_update()` re-arms it
apply it per TLW. Its `#if wxVERSION_NUMBER < 3300` block that themed the tooltip window is dead:
`wxToolTip::GetToolTipCtrl()` is private in 3.3 (`include/wx/msw/tooltip.h:92`), and wx dark-enables the
tooltip window itself through `wxMSWDarkMode::AllowForWindow`, so tooltips follow wx's mode (the system app
mode), not Orca's **[source]** (`src/msw/tooltip.cpp:321`). `update_dark_ui(wxWindow*)` (`GUI_Utils.cpp`, `_WIN32`), which `DPIAware`'s constructor and
→ …), `MonitorPanel::when_built`, `CalibrationPanel::when_built`, every `Tab::sys_color_changed()` (tabs, model tabs,
plate tab), `MenuFactory::sys_color_changed(m_menubar)` (its body is compiled out with `#if 0`, so menu-bar item
icons are not re-rasterized; the cached context menus are, via `Plater::sys_color_changed` →
`MenuFactory::sys_color_changed()`), `WebView::RecreateAll()`, then `Refresh()`. A cached,
hidden or lazily built window must be added here or chained from something here: the walk reaches dialogs parented
to the main frame (they are in `GetChildren()`), but only their colours (and `ScalableButton` icons) — not their
title bar, `ScalableBitmap`-based images or `on_sys_color_changed()`. Commit bab3c72e4f fixed the compare dialog keeping old row colours by calling
`diff_dialog.on_sys_color_changed()` from this fan-out (the dialog is not destroyed on close).
Who gets what on a switch:
| Window | Windows (Preferences) | macOS/Linux (system) |
|---|---|---|
| MainFrame subtree | walk + fan-out | walk + fan-out |
| Open dialog parented to the main frame | walk only (Preferences re-themes itself) | its own `on_sys_color_changed()` via the OS event + the walk |
| Hidden cached dialog parented to the main frame | walk only, unless chained in the fan-out | walk + its own handler (the OS event reaches every TLW) |
| Modal dialogs (`dialogStack`, via `DPIAware::ShowModal`) | cannot be open across a Preferences change | live |
**What a dialog must do to survive a toggle:**
1. Derive from `DPIDialog` (`DPIAware<wxDialog>`, `GUI_Utils.hpp`; `on_sys_color_changed()` is a protected virtual
no-op by default) and use palette colours / Orca widgets.
2. End the constructor with `wxGetApp().UpdateDlgDarkUI(this)` — it also sets the Windows dark title bar and
explorer theme the children walk alone does not.
3. Override `on_sys_color_changed()` to re-create `ScalableBitmap`s (`msw_rescale()` + re-`SetBitmap`), re-pick
`*_dark` icon names, re-apply construction-time and owner-drawn colours, call widget `Rescale()` where styles
depend on the theme, then `Refresh()`. It runs on macOS/Linux from the OS event; on Windows only if the main-frame
fan-out calls it.
4. If it outlives a show (cached singleton, lazily built), register it in `MainFrame::on_sys_color_changed`.
**Pitfalls**
- **Rule:** Never hand a literal light-theme colour straight to `SetForegroundColour`/`SetBackgroundColour`/`wxPen`/`wxBrush`;
wrap it in `StateColor::darkModeColorFor(...)` when it is a `gDarkColors` key, or branch on
`wxGetApp().dark_mode()` with an explicit dark counterpart when it is not.
**Why:** `darkModeColorFor` is an exact RGBA lookup; an unmapped colour (`#F0F0F0`, `#333333`, `#5C5C5C`) passes
through unchanged and renders as a light patch, or — set directly with no walk after it — dark-on-dark text. This
also applies to owner-drawn `wxPaintDC` painting (popup borders/fills), which the walk cannot reach.
The general wx rule is that setters do not send events. The exceptions are the bugs:
| Call | Event emitted? | Cite |
|---|---|---|
| `wxTextEntry::SetValue(s)` | **yes**, one `wxEVT_TEXT`, even when `s` equals the current text | `interface/wx/textentry.h:539-542`; [source] `src/common/textentrycmn.cpp:236-254`, `src/msw/textctrl.cpp:1133-1145` |
| `wxTextEntry::ChangeValue(s)` | no | `interface/wx/textentry.h:177-178` |
| generic `wxDataViewCtrl::Expand/Collapse` | EXPANDING/EXPANDED, COLLAPSING/COLLAPSED; `Collapse` can also send SELECTION_CHANGED when it hides selected rows | [source] `src/generic/datavgen.cpp:4110-4150` |
Orca's widgets add their own rows; see [Orca widgets](#orca-replacement-widgets-event-contracts).
Pitfalls:
- **Rule:** A model→view refresh must not loop back through the view's change handler.
**Why:**`SetValue` and `SetSelection` on a book fire synchronously inside the refresh; the handler writes the
value back, marks the preset dirty, or rebuilds the page it is running in. On macOS a programmatic DVC
| Current item | may be unselected | may be unselected | always selected; `SetCurrentItem` selects in multi-selection | 1456-1458, 1655-1665 |
| `IsExpanded()` | correct | correct | may be true for leaves (documented bug) | 1597-1603 |
| `CreateEditorCtrl()` | called | called | **never called** | 2626-2629 |
| `wxDataViewRenderer::SetAlignment()` | yes | yes | ignored; aligns as the column header | 2079-2085 |
| `IsEditCancelled()` / veto `EDITING_DONE` | yes | documented as unavailable; [source] text commits (`src/gtk/dataview.cpp:2240-2246`) and custom-editor commits (`src/common/datavcmn.cpp:791-853`) do go through `DoHandleEditingDone`, so veto works | documented as unavailable; [source] native text commit reaches the model before `EDITING_DONE` | 3942-3958 |
| Icon+text markup (`EnableMarkup` on `wxDataViewIconTextRenderer`) | no | yes | no | 2215-2218 |
| `EnableDropTargets` | full | first format only | full | 1377-1385 |
Note: `wxObject`'s default copy and assignment share ref-data with correct refcounting
(`include/wx/object.h:324-338`); `DataViewBitmapText`'s field-copying operators are harmless but not required.
## Orca replacement widgets: event contracts
Orca's interactive controls replace raw wx ones (catalog, constructors and quirks: `references/orca-widgets.md`).
They emit the native event *types* through `GetEventHandler()->ProcessEvent`, so command events propagate to parents
like native ones and stop at dialogs (`wxWS_EX_BLOCK_EVENTS`) and, on MSW and macOS, at popups (not on wxGTK,
`references/events.md` §5) — but ids, event objects, event classes and setter side effects differ. "Ancestor" in the
Bind column means an ancestor bound by event type; binding an ancestor by the widget's id is discouraged
(`references/events.md` §10):
| Widget | Replaces | Emits (user action) | Id / object | Setter side effects | Bind |
|---|---|---|---|---|---|
| `::TextInput` | `wxTextCtrl` | inner `wxEVT_TEXT` (propagates); `wxEVT_TEXT_ENTER`, `wxEVT_KILL_FOCUS` re-sent to the wrapper only | TEXT: inner ctrl's id/object; ENTER/KILL_FOCUS: wrapper's id, inner object | `GetTextCtrl()->SetValue` → `wxEVT_TEXT` | TEXT: wrapper or `GetTextCtrl()`; ENTER/KILL_FOCUS: wrapper or `GetTextCtrl()`, never a parent |
| `::ComboBox` | `wxComboBox`/`wxChoice` | `wxEVT_COMBOBOX` (int = index, string = text); DROPDOWN/CLOSEUP | COMBOBOX: combo's auto-generated id (the ctor `id` is ignored) and object; DROPDOWN/CLOSEUP: id 0, no object | `SetSelection`/`SetValue` → `wxEVT_TEXT` when the text ctrl is shown or `CB_NO_TEXT`; `SelectAndNotify` → `wxEVT_COMBOBOX` | COMBOBOX on the combo (or an ancestor); DROPDOWN/CLOSEUP on the combo without id |
| `::SpinInput` | `wxSpinCtrl` | `wxEVT_SPINCTRL` (a `wxCommandEvent`); `EVT_SPINCTRL_TEXT` per parsable keystroke; inner `wxEVT_TEXT` | SPINCTRL / SPINCTRL_TEXT: spinner id/object | `SetValue(int)` → `EVT_SPINCTRL_TEXT` + `wxEVT_TEXT`, no `wxEVT_SPINCTRL` | SPINCTRL on the spinner or an ancestor, handler takes `wxCommandEvent&` |
| `::CheckBox`, `SwitchButton` | `wxCheckBox` | `wxEVT_TOGGLEBUTTON` | native | `SetValue` silent | on the widget (with `Skip()`) or an ancestor; never `wxEVT_CHECKBOX` |
| `::RadioGroup` | `wxRadioBox` | `wxEVT_COMMAND_RADIOBOX_SELECTED` | group id, no object | **every** `SetSelection` emits | on the group or an ancestor |
| `::TabCtrl` | tab bar | `wxEVT_TAB_SEL_CHANGING` then `wxEVT_TAB_SEL_CHANGED` (plain `wxCommandEvent`) | tab ctrl id/object; CHANGING carries the old index | `SelectItem(i)` emits both when `i` changes | on the ctrl or an ancestor; cannot veto |
| `Button` | `wxButton` | `wxEVT_BUTTON` | button id/object | — | on the button or an ancestor |
- **Rule:** Derive a scale with `GetDPIScaleFactor()` or `wxDPIChangedEvent::Scale*`, never by
dividing a DPI by 96.
**Why:** macOS's standard PPI is 72 (`interface/wx/display.h:203-205`), so `GetDPI().x / 96.0`
is 1.5 on a 2x Retina screen and 0.75 on a 1x one.
```cpp
double s = GetDPI().x / 96.0; // Wrong on macOS
double s = GetDPIScaleFactor(); // Right
```
- **Rule:** Choose icon/raster resolution from `GetDPIScaleFactor()`, not
`GetContentScaleFactor()`.
**Why:** content scale is always 1 on MSW (`interface/wx/window.h:1587-1592`), so icons never grow there.
## Choosing a size unit: FromDIP, em_unit, text metrics
Two scaling currencies coexist in Orca, plus the text metrics wx recommends:
| Unit | Use for | Value per platform |
|---|---|---|
| `FromDIP(n)` | fixed sizes in new code: icon sizes, borders, gaps, control heights, min sizes not driven by text | MSW n·DPI/96; macOS/GTK identity |
| `em_unit` (`em_unit(this)`, `wxGetApp().em_unit()`, `DPIAware::em_unit()`) | the settings code (`Tab`, `OptionsGroup`, `Field`, `ObjectList` columns) and any size that must follow text on Linux; sizes written as multiples, `wxSize(65 * em, 30 * em)` | MSW `max(10, 10 × scale_factor)`; macOS always 10; GTK width of "m" in the window font − 1 (min 10) |
| text metrics (`GetTextExtent`, best sizes, `ConvertDialogToPixels`) | widths that hold translated text | follow the font everywhere |
The overview prefers text metrics or dialog units over pixel values and calls `FromDIP` "the
**Contract** (`interface/wx/event.h:3560-3593`): sent "to each wxTopLevelWindow affected by the
change, and all its children recursively (post-order traversal)" — on a move to a monitor with a
different DPI or a system DPI change. "You should almost always call event.Skip() … as many
controls rely on processing this event in order to update their appearance". The TLW's default
handler "only sets the new window size, by scaling the current size by the DPI ratio … and also
ensuring that the window is still bigger than its best size"; to prevent it, handle the event on
the TLW, `SetSize()` there and do *not* Skip. Documented generators: wxMSW "if and only if" Windows
10 1703+ with a PerMonitorV2 manifest; wxGTK "when using GTK 3.10 or later and only since
wxWidgets version 3.3.0".
Helpers (`interface/wx/event.h:3610-3660`): `GetOldDPI()`, `GetNewDPI()`, `Scale(wxSize)`, and since 3.3.0
`Scale(wxPoint)`/`Scale(wxRect)`, `ScaleX/ScaleY` — old-DPI→new-DPI via `wxMulDivInt32`. Prefer
them to `GetNewDPI()/96` (72-based on macOS).
**Platforms** [source unless noted]:
| Port | Who generates it | What wx itself rescales |
|---|---|---|
| MSW | `WM_DPICHANGED` → `wxNonOwnedWindow::HandleDPIChange` (`src/msw/nonownedwnd.cpp`), only for PMv2-aware windows | `wxWindowMSW::MSWUpdateOnDPIChange` (`src/msw/window.cpp`) recurses from the TLW through every non-TLW child; for each window it first rescales `m_min/maxWidth/Height`, invalidates best size, re-creates the window font at the new PPI (`MSWUpdateFontOnDPIChange`; the TLW's override `wxTopLevelWindowMSW::MSWUpdateFontOnDPIChange` only re-selects its icons and leaves the TLW font alone), rescales sizer borders, spacer sizes and nested-sizer min sizes (`UpdateSizerOnDPIChange`; window items keep their min size because each window scales its own), then recurses into its children, then sends the event to that window. If the TLW's event was **not processed**, wx `SetSize()`s the TLW to the suggested rect inflated to the sizer's min size |
| macOS | `windowDidChangeBackingProperties` when the backing scale changes (`src/osx/cocoa/nonownedwnd.mm`), DPIs built from the 72-based std PPI (`wxWindowBase::WXNotifyDPIChange`). Undocumented. The same notification sends `wxSysColourChangedEvent` when the colour space changes | nothing (logical = DIP) |
| GTK3 | TLW configure event when `GetContentScaleFactor()` changed (`src/gtk/toplevel.cpp``wxTopLevelWindowGTK::GTKConfigureEvent`, `__WXGTK3__` only); the initial scale is captured at creation, so there is no event at startup; DPIs are 96 × integer scale | nothing (logical = DIP) |
| GTK2 | never | — |
The default TLW resize exists only on MSW; on macOS/GTK3 logical sizes do not change with DPI.
Handler order: children receive the event before their TLW (post-order, documented; [source]
`src/common/wincmn.cpp``NotifyAboutDPIChange`; MSW recursion in `MSWUpdateOnDPIChange`).
Dynamic handlers run most recently bound first, before static tables
(`docs/doxygen/overviews/eventhandling.h:475-482`), so a handler you `Bind` on
a control runs before the control's own (`wxBookCtrlBase`, `wxComboCtrlBase`, `wxTreeCtrlBase` bind
one; wxMSW `wxStaticBitmap` uses a static table entry, `src/msw/statbmp.cpp`, and MSW button
bitmaps bind one, `src/msw/anybutton.cpp`).
**What wx rescales vs what Orca must re-apply:**
| Item | MSW (done by wx before the event) | macOS / GTK3 | Orca's `on_dpi_changed` must |
|---|---|---|---|
| window min/max sizes | rescaled by the DPI ratio | unchanged (logical = DIP) | re-set only if it recomputes them anyway |
| window font set by `SetFont` | re-created at the new PPI for non-TLW windows; the TLW keeps its old-PPI font [source] | unchanged (points) | nothing: a `wxFont` stores points and every window or DC `SetFont` re-adjusts it to its target's PPI; `DPIAware::rescale` re-reads it and updates `em_unit` |
| TLW size | resized only if the TLW event is unprocessed — never for DPIAware windows | unchanged | re-establish min size, `Fit()`/`SetSize()` |
| values given to setters (`SetRowHeight`, column widths, `SetItemMinSize`, custom-widget sizes, cached pixel members) | never | never | re-apply from `FromDIP`/`em_unit` |
| bitmaps on native controls | reselected from the control's bundle; for Orca's single bitmaps that is the old raster, unscaled or integer-upscaled | GTK never; macOS from the bundle | `msw_rescale()` + `SetBitmap()` |
| Orca widgets (cached measures, named icons) | never | never | call `Rescale()` |
**Pitfalls**
- **Rule:** Call `Skip()` in every `wxEVT_DPI_CHANGED` handler you bind, on a child, a control or
the TLW (DPIAware's own handler is the one exception).
**Why:** your dynamically bound handler runs first; without `Skip()` the control's own handler
(book controls, combo controls, MSW static bitmaps and button images) never runs and keeps its
old-DPI appearance. Bound on a `DPIDialog`/`DPIFrame` (as `m_parent` is here), it also starves
DPIAware's own handler, so `on_dpi_changed` never runs.
| MSW | `FromDIP(px, win)` | `win->GetDPIScaleFactor()` (primary-display ratio when `win` is null) | `FromDIP(px)` px; the tag makes the implicit bundle's DIP size `px`, so wx uses it unscaled |
| 0 | `wxApp::FilterEvent()` and other `wxEventFilter`s (LIFO) | Anything but `Event_Skip` (-1) stops here. Called once per event, not again as it propagates ([source] `src/common/event.cpp:1537-1552`) |
| 1 | `TryBefore()` | Validators on windows |
| 2 | — | If `SetEvtHandlerEnabled(false)`, skip to step 5 (the overview's wording; the interface doc's "skips to step (7)" is inaccurate — [source] `TryHereOnly` returns `false` and `DoTryChain` still runs, `src/common/event.cpp:1582-1591, 1644-1648`) |
| 3 | Dynamic table (`Bind`) | **Most recently bound first**, before the static table (`docs/doxygen/overviews/eventhandling.h:474-483`) |
| 4 | Static event table | Macro order, derived class before base class |
| 4a | Implicit `CallAfter` entry | Runs a queued `wxAsyncMethodCallEvent` only when its event object is this handler ([source] `src/common/event.cpp:1644-1667``TryHereOnly`) |
| 5 | Next handlers in the chain | For windows: the pushed-handler stack (§14) |
| 6 | `TryAfter()` | Windows propagate to the parent (§5); finally `wxTheApp->ProcessEvent()` |
`ProcessEvent` returns `true` iff some handler ran and did not call `Skip()` (`interface/wx/event.h:619-622`).
Before each handler call wx resets the flag with `event.Skip(false)` ([source]
`src/common/event.cpp:1443-1475``ProcessEventIfMatchesId`), so a skip in one handler does not carry over to the
next: every handler that wants processing to continue must call `Skip()` itself.
A handler entry matches when the event type matches and the bound id is `wxID_ANY`, or equals the
event id, or the event id falls in `[id, lastId]` (`src/common/event.cpp:1443-1475`).
---
## 2. Bind and Unbind
**Contract.**
- Forms: `Bind(tag, functor, id = wxID_ANY, lastId = wxID_ANY, userData = nullptr)` and
`Bind(tag, &Class::method, handlerPtr, id, lastId, userData)` (`interface/wx/event.h:876-957`). The
method form accepts "an arbitrary method (doesn't need to be from a wxEvtHandler derived class)";
the handler pointer "must always be specified". `userData`: "wxWidgets will take ownership of
this pointer" — deleted when the handler is unbound or at program termination.
- Handlers can be bound at any time and removed with `Unbind`
(`docs/doxygen/overviews/eventhandling.h:237-252`). `Connect()` is the legacy form: "please use
[Bind] in any new code" (`interface/wx/event.h:705-706`).
- Lifetime (`docs/doxygen/overviews/eventhandling.h:332-335`), for a handler object not derived from `wxEvtHandler`: "the
lifetime of `myFrameHandler` must be greater than that of `MyFrame` object -- or at least it needs
to be unbound before being destroyed".
- `Unbind` "can only unbind functions, functors or methods which have been added using the Bind<>()
method. There is no way to unbind functions bound using the (static) event tables." Its note:
"functors are compared by their address which, unfortunately, doesn't work correctly if the same
address is reused for two different functor objects. Because of this, using Unbind() is not
recommended if there are multiple functors using the same eventType and id and lastId as a wrong
one could be unbound" (`interface/wx/event.h:958-997`).
**Mechanics** [source]:
- `Bind` takes `const Functor&`, stores a **copy** of the functor and records the **address of the
object you passed** (`include/wx/event.h:524-570``wxEventFunctorFunctor`, `:3951-3961`). `Unbind`
matches on that address plus the functor type. So a lambda is unbindable when the same lvalue
(a member `std::function`, a named lambda that stays at one address, heap storage) is passed to
both calls; an inline lambda literal never matches. Method + handler pairs match by value and
always work. Captures must be copyable.
- `DoUnbind` requires `entry->m_id == id`; only `lastId == wxID_ANY` and `eventType == wxEVT_NULL`
act as wildcards (`src/common/event.cpp:1806-1824`). A handler bound with `ctrl->GetId()` is not
removed by `Unbind(evt, fn)` (id defaults to `wxID_ANY`), and vice versa. `Unbind` on a different
emitter than the one you bound on returns `false` silently.
- `DoBind` always appends (`src/common/event.cpp:1769-1803`): binding the same handler twice runs it twice.
- No locking in `DoBind`/`DoUnbind`: main thread only.
**Lifetime by handler kind** [source]:
| Handler | Removed automatically when the handler object dies? |
|---|---|
| `src->Bind(evt, &C::m, sink)` where `sink` is a `wxEvtHandler` (any window) other than `src` | Yes. `DoBind` registers a `wxEventConnectionRef` on the sink (`src/common/event.cpp:1793-1802`); the sink's `~wxTrackable` calls `OnSinkDestroyed`, which deletes the entries (`include/wx/event.h:4184-4200`, `src/common/event.cpp:2022-2042`) |
| Lambda or functor (`[this]{…}`), free function | No. `GetEvtHandler()` is null for functors |
| Method of a class not derived from `wxEvtHandler` (`GLCanvas3D`, `Plater::priv`) | No |
| Handlers bound on `this` itself | Deleted with `this` (`~wxEvtHandler`, `src/common/event.cpp:1204-1245`) |
The automatic removal runs **late**: `wxEvtHandler` derives from `wxObject, wxTrackable`
(`include/wx/event.h:3705-3706`), so `~wxTrackable` runs after the derived destructor, after member
destruction, and after the port destructor has destroyed the native window and the children
(`src/osx/window_osx.cpp``~wxWindowMac`, `src/msw/window.cpp``~wxWindowMSW`). Events the source
emits during that teardown (activation, focus, size, show) still reach the half-destroyed sink.
Unbind explicitly in the sink's destructor whenever the source can fire while the sink dies.
**Usage.**
```cpp
// Method + wxEvtHandler sink on another window: removed on sink death (late) — unbind early anyway
- **Rule:** Move-only captures do not compile in `Bind` or `CallAfter`; use `std::shared_ptr`.
Cite: `include/wx/event.h:524-570` (functor stored by copy).
---
## 3. Static event tables
**Contract.** `wxDECLARE_EVENT_TABLE()` in the class, `wxBEGIN_EVENT_TABLE(Class, Base)` …
`wxEND_EVENT_TABLE()` in the `.cpp`; entries are searched in macro order, then the base class table
(`interface/wx/event.h:567-625` step 5). They run **after** every dynamic handler for the same event.
**Multiple inheritance.** "it is imperative that the wxEvtHandler(-derived) class is the first class
inherited such that the `this` pointer for the overall object will be identical to the `this`
pointer of the wxEvtHandler portion" (`interface/wx/event.h:376-380`). Put the wx base first in
`class X : public wxPanel, public Other`.
**OrcaSlicer.** New code binds with lambdas (`[this](auto& e)`) or method + `this`. The core widgets
keep their internal handlers in static tables (`DECLARE_EVENT_TABLE()` in `Widgets/StaticBox.hpp`,
`Widgets/Button.hpp`; tables in `Button.cpp`, `TextInput.cpp`, `SpinInput.cpp`, `ComboBox.cpp`,
`DropDown.cpp`, `TabCtrl.cpp`). That is why `Skip()` in user handlers matters (§4): a user `Bind` on
`Button` for `wxEVT_LEFT_DOWN` runs before `Button::mouseDown`. Do not add new static tables; when
changing such a widget, keep its internal handlers where they are.
---
## 4. Skip discipline
**Contract.** `wxEvent::Skip` (`interface/wx/event.h:238-252`): "Without Skip() (or equivalently if
Skip(false) is used), the event will not be processed any more. If Skip(true) is called, the event
processing system continues searching for a further handler function for this event, even though it
has been processed already in the current handler. In general, it is recommended to skip all
non-command events to allow the default handling to take place. The command events are, however,
normally not skipped as usually a single command such as a button click or menu item selection must
only be processed by one handler."
| Event | Rule | Cite |
|---|---|---|
| `wxEVT_SET_FOCUS` / `wxEVT_KILL_FOCUS` | "should almost invariably call wxEvent::Skip()"; a KILL_FOCUS handler "must not call wxWindow::SetFocus()" — defer it (`CallAfter`) | `interface/wx/event.h:3410-3416` |
| `wxEVT_SIZE` | "Sizers … rely on size events to function correctly … call Skip on all size events you catch" | `interface/wx/event.h:5058-5060` |
| `wxEVT_KEY_DOWN` | Not skipping suppresses `wxEVT_CHAR` for that key and "may also prevent accelerators … from working" | `interface/wx/event.h:1454-1461` |
| `wxEVT_CHAR_HOOK` | Propagates upward; handled (not skipped) → no `KEY_DOWN`/`CHAR`. Skip every key you do not handle | `interface/wx/event.h:1485-1508`; keyboard order: `references/mouse-keyboard-focus.md` |
| `wxEVT_PAINT` | The handler "must create a wxPaintDC"; skip only if default painting must also run | `interface/wx/event.h:2274-2285`; `references/painting-custom-widgets.md` |
| `wxEVT_DPI_CHANGED` | "should almost always call event.Skip() … as many controls rely on processing this event"; a TLW handler may deliberately not skip to suppress the default resize | `interface/wx/event.h:3571-3583` |
| `wxEVT_SYS_COLOUR_CHANGED` | The default handler propagates it to children; a TLW handler must Skip, call the base, or forward | `interface/wx/event.h:1950-1955` |
| `wxEVT_ACTIVATE` on a TLW | MSW: `wxTopLevelWindowMSW::OnActivate` (static table) saves and restores the last focused child; macOS: `wxFrame::OnActivate` (static table) installs the frame's menubar. A non-skipping dynamic handler disables both | [source] `src/msw/toplevel.cpp``wxTopLevelWindowMSW::OnActivate`, `src/osx/carbon/frame.cpp``wxFrame::OnActivate`; `references/popups-menus.md` §15 |
| Mouse/key events on custom widgets | Dynamic handlers run before the widget's static table: not skipping disables the widget's own press/release/capture logic | [source] `src/common/event.cpp:1644-1667` |
| Command events | Normally not skipped; `Skip()` lets the event continue to the static table, the next handler, then the parent | `interface/wx/event.h:247-251` |
wx's own controls bind some internals dynamically in their constructors — e.g.
`wxBookCtrlBase`, `wxComboCtrlBase` and `wxTreeCtrlBase` bind `wxEVT_DPI_CHANGED`
(`src/common/bookctrl.cpp:60`, `src/common/combocmn.cpp:839`, `src/common/treebase.cpp:172`) — so a
non-skipping handler you bind on
such a control later starves its rescale. DPI propagation order: `references/dpi-bitmaps-fonts.md`.
**OrcaSlicer — deliberate deviations.**
- `DPIAware<P>` (`src/slic3r/GUI/GUI_Utils.hpp`) binds `wxEVT_DPI_CHANGED` on the TLW (not on macOS)
**without** Skip: Orca rescales itself in `rescale()` and suppresses wx's default TLW resize, which
the doc allows. Its `wxEVT_SYS_COLOUR_CHANGED` handler Skips on macOS and Linux but not on Windows,
where the theme is app-forced.
- Because `DPIAware` binds in its own constructor, everything a derived dialog binds on itself later
runs first. `DialogButtons` binds the parent's `wxEVT_DPI_CHANGED` and calls `Skip()`, so it runs
before `DPIAware`'s handler and lets it run. Likewise `DPIAware` maps Esc to `Close()` in a
`wxEVT_CHAR_HOOK` handler on every `DPIDialog`: a derived dialog's own `wxEVT_CHAR_HOOK` handler must
`Skip()` every key it does not consume, or Esc stops closing the dialog (`CloneDialog`'s Enter→OK
hook is the model).
**Pitfalls.**
- **Rule:** Take the event parameter by reference.
**Why:** the functor form compiles with a by-value parameter; the lambda then receives a copy
(`include/wx/event.h:534-545` calls `m_handler(static_cast<EventArg&>(event))`), and `Skip()` on the
copy does nothing — the original counts as handled.
`src/common/popupcmn.cpp`) blocks at every dialog, but at popups only on MSW/macOS; on wxGTK an
unconsumed command event from inside a popup reaches the opener's ancestors and any unfiltered
handler there. Bind on the dialog/popup or on the control, consume the event there, or re-emit
explicitly.
- **Rule:** An ancestor that binds a command event without an id filter receives that event from
every descendant control of that type.
---
## 6. Emitting events synchronously
| Call | Use for | Note |
|---|---|---|
| `win->ProcessWindowEvent(e)` = `win->GetEventHandler()->ProcessEvent(e)` | Emitting a window's own event | "ProcessEvent() itself can't be called for wxWindow objects as it ignores the event handlers associated with the window; use this function instead" (`interface/wx/window.h:2736-2744`) |
| `win->HandleWindowEvent(e)` = `GetEventHandler()->SafelyProcessEvent(e)` | Same, from code that must not leak exceptions (native callbacks) | `interface/wx/window.h:2726-2734` |
| `win->ProcessWindowEventLocally(e)` | This window and its pushed handlers only, no propagation | `interface/wx/window.h:2746-2757` |
| `handler->ProcessEventLocally(e)` | Forwarding an event to another handler | "should, be called to forward an event to another handler instead of ProcessEvent() which would result in a duplicate call to TryAfter()" (`interface/wx/event.h:627-651`) |
| MSW | Keep running inside native modal loops (menu tracking, window move/size, `MessageBox`, common dialogs) through a `WH_GETMESSAGE` hook (`src/msw/window.cpp``wxIdleWakeUpModule::MsgHookProc` → `wxApp::MSWProcessPendingEventsIfNeeded`, `src/msw/app.cpp:717-730`) | Not processed inside those native loops — do not defer repaint/layout to idle if it must happen during a live resize |
| macOS | Processed by a run-loop observer at `kCFRunLoopBeforeTimers` in common modes (`src/osx/core/evtloop_cf.cpp:86-112`) | At `kCFRunLoopBeforeWaiting`. While a native `wxMessageDialog`, `wxFileDialog` or `wxDirDialog` is modal, `wxCFEventLoopPauseIdleEvents` stops **both** (`src/osx/cocoa/msgdlg.mm:62`, `src/osx/cocoa/filedlg.mm:601`, `src/osx/cocoa/dirdlg.mm:123`, `src/osx/core/evtloop_cf.cpp:362-379`): a `CallAfter` queued before or during such a dialog runs only after it closes |
| GTK3 (default build, X11 and Wayland) / opt-out GTK2 | Pending events and idle run from one GLib idle source at `G_PRIORITY_LOW` (`src/gtk/app.cpp:102-153, 631`, `wxApp::DoIdle`): they starve while higher-priority sources (timers, redraw, input floods) are busy; a masked `YieldFor` removes that source before each iteration, so nothing queued runs inside it (`references/threads-timers-app.md` §Event categories and yields) | same source; `RequestMore()` or remaining pending events keep it installed — a busy loop |
**OrcaSlicer.** `GLCanvas3D::post_event` sets the event object and `wxPostEvent(m_canvas, …)` —
asynchronous, dispatched on the `wxGLCanvas`, where `Plater::priv` binds the canvas events.
`BackgroundSlicingProcess` posts with `wxQueueEvent(wxGetApp().mainframe->m_plater, evt.Clone())` and
`new wxCommandEvent(...)`; `SlicingProcessCompletedEvent` carries the worker's exception as a
`ObjectList::is_live_model_item`), b779a7bfed (`src/slic3r/plugin/host/PluginHostUi.cpp`, the
`UiRegistry::is_open` re-check in `ui_create_window`).
**Liveness tools.**
| Tool | Use when | Limits |
|---|---|---|
| Queue on the target itself (`win->CallAfter`) | The lambda touches only `win` | Main thread; still runs for a `Destroy()`ed TLW until the idle delete |
| `std::shared_ptr<std::atomic<bool>>` alive flag (`PluginsDialog::m_alive`) | App-queued work, any thread | Flips at the start of the derived destructor |
| `wxWeakRef<T>` (`interface/wx/weakref.h`) | Main-thread checks of a `wxEvtHandler`/window | Reset in `~wxTrackable`, i.e. after the derived destructor and `DestroyChildren`; tracker list unsynchronised (`include/wx/tracker.h`) — create and test on the main thread only |
| Registry / model re-validation (`UiRegistry::is_open`, `ObjectList::is_live_model_item` in macOS-only code) | Ids or items that may be rebuilt | `is_live_model_item` compares node pointers, so it cannot detect a freed node whose address was reused |
| `wxTheApp->IsScheduledForDestruction(w)` | Detect a `Destroy()`ed TLW still in `wxPendingDelete` | — |
| `GUI_App::is_closing()` | Callbacks during shutdown | Check before posting and again inside the lambda |
Deletion and `wxWeakRef` details: `references/windows-dialogs.md`.
**OrcaSlicer.**
- `run_on_ui_blocking` (`src/slic3r/plugin/host/PluginHostUi.cpp`) captures `fn` and the promise **by
reference** in `wxGetApp().CallAfter` and runs `fn` inline when `wxIsMainThread()` (a `CallAfter` +
`future.get()` on the main thread would deadlock). It is safe only because the caller blocks until
the lambda has run; never copy by-reference captures into a fire-and-forget `CallAfter`.
- Webview script messages arrive synchronously inside the native callback on WebKitGTK and WKWebView;
defer window work from them with `CallAfter` plus an alive check, and in a `WebViewHostDialog`
subclass do it in your own `on_script_message`, because the base dispatches synchronously. Details
and cites (b779a7bfed, f2ccbfc8b5, 0a0d59b76b): `references/webview-gl-aui-media.md`.
**Pitfalls.**
- **Rule:** Do not use `IsBeingDeleted()` to detect a `Destroy()`ed TLW. Use an alive flag or
`IsScheduledForDestruction`.
- **Rule:** Do not disable a handler with `SetEvtHandlerEnabled(false)` while it still has
`CallAfter`s it needs (§14).
---
## 9. Custom events and payload ownership
**Declaring.** `wxDECLARE_EVENT(NAME, Class)` in the header, `wxDEFINE_EVENT(NAME, Class)` in exactly
one `.cpp` — "this is a definition so can't be in a header" (`docs/doxygen/overviews/eventhandling.h:634-640`).
[source] `wxDEFINE_EVENT` expands to `const wxEventTypeTag<T> NAME(wxNewEventType())`
(`include/wx/event.h:105-106`). A namespace-scope `const` has internal linkage unless an `extern`
declaration (`wxDECLARE_EVENT`) precedes it, so a header definition alone gives every translation unit
a **different** event type: binds and posts silently never meet, and nothing fails to link.
`wxDECLARE_EXPORTED_EVENT` exists only for DLL export (`include/wx/event.h:110-115`).
**Choosing the class** (`docs/doxygen/overviews/eventhandling.h:600-613`): `wxEvent` (no payload, no propagation) or
`wxCommandEvent` (int, long, string, client data; propagates). For richer payloads derive a class,
add members and implement `Clone() { return new MyEvent(*this); }` (`docs/doxygen/overviews/eventhandling.h:686-741`);
event-table macros need extra boilerplate, `Bind` does not. `wxEvent::Clone()` is pure virtual
(`include/wx/event.h:1009`), so a direct `wxEvent` subclass without `Clone` does not compile — but
a subclass of a concrete event (`wxCommandEvent::Clone` returns `new wxCommandEvent(*this)`,
`include/wx/event.h:1655`) silently inherits a slicing `Clone`.
**Payload ownership.**
| Payload | Ownership | Cite |
|---|---|---|
| `wxCommandEvent::SetString/SetInt/SetExtraLong` | Copied / by value | `interface/wx/event.h` wxCommandEvent |
| `wxCommandEvent::SetClientData(void*)` | Raw pointer, never owned | — |
| `wxCommandEvent::SetClientObject(wxClientData*)` | "not owned by the event … must be owned and deleted by another object (e.g. a control) that has longer life time than the event object" | `interface/wx/event.h:2210-2216` |
| `wxEvtHandler::SetClientObject(wxClientData*)` | Owned by the handler: "Any previous object will be deleted"; deleted in `~wxEvtHandler`; do not mix with `SetClientData` on the same handler | `interface/wx/event.h:1062-1074`; [source] `src/common/event.cpp:1240-1242` |
| `Bind(…, userData)` | Owned by the binding, deleted on unbind or at exit | `interface/wx/event.h:900-905` |
| `wxThreadEvent::SetPayload<T>` | Copied; T's copy must not share state | `interface/wx/event.h:3775-3790` |
[source] For a `wxEVT_TEXT` event whose event object is a text-entry window,
`wxCommandEvent::GetString()` always returns the control's **current** value and ignores the stored
string (`src/common/event.cpp:430-447`). The copy constructor materialises the stored string
(`include/wx/event.h:1622-1632`), but a queued or cloned copy still has the event object, so its
`GetString()` reads the control live at handling time. If the value at emit time matters, capture it
yourself.
**OrcaSlicer.**
- Custom event types are declared in headers (`GLCanvas3D.hpp` and `Plater.hpp` in
`Slic3r::GUI`; widget events such as `EVT_SPINCTRL_TEXT`, `wxEVT_TAB_SEL_CHANGED` at global scope)
and defined once in the matching `.cpp`.
- Payload classes in `src/slic3r/GUI/Event.hpp`: `SimpleEvent`, `IntEvent` (`get_data()`),
`Event<T>` (`.data`), `ArrayEvent<T, N>` (`.data`). All derive from `wxEvent`, propagate (§5), and
their `Clone()` copies type, data and event object (not the id). `LoadPrinterViewEvent` there shows
the `wxCommandEvent`-derived shape: a copy constructor that copies the extra member, and
`Clone() { return new LoadPrinterViewEvent(*this); }`.
13. Enter/leave handlers only record hover state, per child window. Apply `Show()`/`Hide()` +
`Layout()` from `wxEVT_IDLE` or `CallAfter`, only when the state differs. No `wxFindWindowAtPoint()`
there, and no `IsShownOnScreen()` as a guard. → [Hover](#hover-handlers-and-enterleave-feedback-loops)
14. To see a child's keys, bind `wxEVT_CHAR_HOOK` on the parent and `Skip()` everything not
handled; `wxEVT_KEY_DOWN`/`wxEVT_CHAR` do not propagate. → [wxEVT_CHAR_HOOK](#wxevt_char_hook)
15. A top-level `wxEVT_CHAR_HOOK` must not consume printable keys, Space or text-editing chords
(Ctrl+A/C/V/X/Z, Delete, Backspace, Home/End, arrows) while a text entry has focus. →
[wxEVT_CHAR_HOOK](#wxevt_char_hook)
16. Test modifiers with `GetModifiers() == wxMOD_…`, not `ControlDown()`; on macOS `wxMOD_CONTROL` is
Cmd and `wxMOD_RAW_CONTROL` is the Control key. → [Modifiers](#modifiers-and-the-cmd-mapping)
17. Match letters, digits and special keys on `wxEVT_KEY_DOWN` and punctuation on `wxEVT_CHAR`;
build chords with `KeyChord::from_event`. → [Key codes](#key-codes)
18. New user-facing shortcuts go through the registry (`Shortcut` enum + `shortcut_table` row + a case in
the context's dispatcher). No raw `wxAcceleratorTable`, hard-coded key test or literal key name in a
label. → [Registry](#orcas-shortcut-registry)
19. Global chords need Ctrl/Alt or a key that types nothing. On macOS a live menu accelerator is
consumed by the menu bar before any wx key event, so the menu handler and the dispatcher case
must do the same thing. → [Accelerators](#accelerators-and-menu-shortcuts)
20. `SetAcceleratorTable()` replaces the window's table; build one table per window. It fires only
while focus is inside that window and stops at the top-level window. → [Accelerators](#accelerators-and-menu-shortcuts)
21. `SetFocus()` only on a user action or when the top-level window `IsActive()`; never from hover or
timers unconditionally, never inside `wxEVT_KILL_FOCUS` (defer with `CallAfter`). Focus handlers
`Skip()`. → [Focus](#focus)
22. Clear a tooltip with `UnsetToolTip()`, not `SetToolTip("")`. A composite that forwards tooltips
overrides `DoSetToolTip` as well as `DoSetToolTipText`. → [Tooltips](#tooltips)
23. Tooltips on disabled controls only through the MSW-gated parent-motion hack
(`Button::EnableTooltipEvenDisabled`); never install parent-motion forwarding on macOS. →
[Tooltips](#tooltips-on-disabled-controls)
24. Set a window cursor once with `SetCursor()`, reset with `wxNullCursor`, never toggle it on
enter/leave. Busy cursors via `wxBusyCursor` RAII, on the main thread. → [Cursors](#cursors)
## Mouse capture
### Contract
`CaptureMouse()` "Directs all mouse input to this window". wx "maintains the stack of windows having
captured the mouse … you must release the mouse as many times as you capture it, unless the window
receives the wxMouseCaptureLostEvent event. Any application which captures the mouse in the
beginning of some operation must handle wxMouseCaptureLostEvent and cancel this operation when it
receives the event. The event handler must not recapture mouse." (`interface/wx/window.h:3805-3819`)
`wxMouseCaptureLostEvent` goes to **all windows on the capture stack** when capture is lost to an
"external" event (a dialog box shown, another application capturing the mouse). It is "not sent if
the capture changes because of a call to CaptureMouse or ReleaseMouse"
(`interface/wx/event.h:3496-3514`). The doc's "currently emitted under Windows only" /
`@onlyfor{wxmsw}` is stale: wxGTK sends it too, wxOSX never does (table below).
`wxMouseCaptureChangedEvent` is MSW-only and is sent to a window that loses capture "even if
wxWindow::ReleaseMouse was called by the application code"; the doc's purpose: "allows an application
to cater for unexpected capture releases" (`interface/wx/event.h:4659-4680`; only `src/msw/window.cpp`
`wxWindowMSW::HandleCaptureChanged` builds it).
### The capture stack
[source] `src/common/wincmn.cpp:3322-3456` (namespace `wxMouseCapture`; the stack is a
`wxVector`, i.e. `std::vector`):
| Call | Behaviour |
|---|---|
| `CaptureMouse()` (:3352) | Asserts (compiled out) if `this` is already anywhere on the stack. Natively releases the current holder, `DoCaptureMouse()`, pushes `this`. A second call on the same window pushes it **twice**. |
| `ReleaseMouse()` (:3371) | Calls `DoReleaseMouse()`**first**, dropping the native capture whoever owns it. Then `wxCHECK_RET(stack.back() == this)` returns silently if this window is not on top. Pops, and if the stack is non-empty **re-captures the new top** (:3412-3416). |
| `HasCapture()` | `AsWindow() == GetCapture()` (`include/wx/window.h:1125`): true only for the top of the stack. |
| `NotifyCaptureLost()` (:3438) | Does nothing while a wx `Capture/ReleaseMouse` is in progress. Otherwise sends the lost event to each stacked window, top first, popping each. The stack ends empty, so no `ReleaseMouse()` is owed afterwards. A handler that leaves the event unprocessed (`Skip()`) hits a compiled-out `wxFAIL` (:3423-3436). |
| `~wxWindowBase` (:452) | Asserts (compiled out) if the window is still on the stack, and does **not** remove it: the stack keeps a dangling pointer. |
Consequences:
- Double capture followed by one release leaves the **same window captured** (the restore step
re-captures it). That is a leaked capture.
- An unguarded `ReleaseMouse()` from a window that is not on top frees the real owner's native
capture and leaves the stack out of step; a later release can re-capture a stale window.
- An unguarded `ReleaseMouse()` inside the lost handler pops the stack while `NotifyCaptureLost`
is iterating it. The loop then pops an entry it never notified, or calls `pop_back()` on an
empty vector (undefined behaviour). A `HasCapture()`-guarded release is a harmless no-op there
(next table).
- The documented way to break capture before a modal, `dialog.CaptureMouse(); dialog.ReleaseMouse();`
(`docs/doxygen/overviews/eventhandling.h:878-901`), is undone by the restore step when the
current holder captured through wx: the release re-captures the previous holder. It only breaks
a capture taken natively, outside wx's stack.
### Per-port delivery
| | MSW | GTK3 / GTK2 | macOS |
|---|---|---|---|
| Native capture | `SetCapture` | `gdk_seat_grab` (GTK ≥ 3.20) or `gdk_pointer_grab`. On an unrealized window `DoCaptureMouse` fails a silent `wxCHECK_RET`, but `wincmn` still pushes it (`src/gtk/window.cpp:6732-6765`) | None. `DoCaptureMouse` sets `wxApp::s_captureWindow` and wx reroutes events (`src/osx/window_osx.cpp:596-612`) |
| Capture-lost sent | `WM_CAPTURECHANGED` → `HandleCaptureChanged` → `NotifyCaptureLost`, then `wxEVT_MOUSE_CAPTURE_CHANGED` (`src/msw/window.cpp:5176-5192`) | GTK `grab-broken-event` (`src/gtk/window.cpp:2636-2646`). `wxDialog::ShowModal` and `wxMessageDialog::ShowModal` release the grab and notify (`src/gtk/dialog.cpp:137`, `src/gtk/msgdlg.cpp:285` → `GTKReleaseMouseAndNotify`) | **Never.**`src/osx` has no `NotifyCaptureLost` call |
| `HasCapture()` inside the lost handler | false: `GetCapture()` returns null while `gs_insideCaptureChanged` (`src/msw/window.cpp:757-768`) | false: `g_captureWindow` is cleared before notifying (`src/gtk/window.cpp:6794-6815`) | n/a |
| Modal dialog shown while captured | Lost event only if something takes the native capture (`WM_CAPTURECHANGED`) | Capture released, lost event sent | Capture kept; the dialog cannot be clicked (`src/osx/dialog_osx.cpp``ShowModal` has no capture code) |
| Captor destroyed | wx stack dangles | `g_captureWindow` cleared (`src/gtk/window.cpp:3088-3089`); wx stack dangles | `s_captureWindow` dangles; the next mouse event dereferences it |
| `wxEVT_CHAR_HOOK` while captured | Not generated (any native `::GetCapture()`) | Not generated | Generated |
| Enter/leave while captured | Synthesized for the captor only (`src/msw/window.cpp:6019-6074`) | Synthesized for the captor only; grab crossings ignored (`src/gtk/window.cpp:2073-2117, 2384-2456`) | Other views' enter/exit events are rerouted to the captor |
### The cancel-on-lost pattern
```cpp
// EVT_MOUSE_CAPTURE_LOST(MyWidget::mouseCaptureLost) in the event table, or Bind(...)
void MyWidget::mouseDown(wxMouseEvent& e)
{
e.Skip(); // let focus move
m_pressed = true;
if (!HasCapture()) CaptureMouse(); // a second button mid-press must not push twice
Refresh();
}
void MyWidget::mouseReleased(wxMouseEvent& e)
{
e.Skip();
if (HasCapture()) ReleaseMouse(); // always, not only when m_pressed
if (!m_pressed) return;
m_pressed = false;
Refresh();
if (GetClientRect().Contains(e.GetPosition()))
sendButtonEvent(); // commit only on a real button-up inside
| `wxGetMousePosition()` | screen coordinates (`interface/wx/utils.h:360-365`) | `gdk_device_get_position` with no Wayland handling (`src/gtk/window.cpp:7026-7042`). Wayland exposes no global pointer position to clients [external]; Orca's comments record (0,0) |
| `wxGetMouseState()` | position, buttons and modifiers (`interface/wx/utils.h:367-375`) | position unreliable, same path (`src/gtk/window.cpp:2827-2867`) |
| `ClientToScreen()`, `ScreenToClient()`, `GetScreenRect()` | — | add the top-level window's `gdk_window_get_origin()` (`src/gtk/window.cpp:4630-4698`). Without a global origin the result is only meaningful relative to the same top-level window: fine for deltas and hit tests inside one window, wrong across windows or for absolute placement |
| `wxFindWindowAtPoint(pt)` | deepest window at a screen point; disabled children count, hidden ones are skipped (`interface/wx/utils.h:385-394`) | built on screen rectangles, so unreliable. On GTK it is `wxGenericFindWindowAtPoint`, a walk over every top-level window and child (`src/gtk/utilsgtk.cpp:98-101`, `src/common/utilscmn.cpp:1294-1345`) |
| `wxGetKeyState(key)` | "In wxGTK, this function can be only used with modifier keys … when not using X11 backend" (`interface/wx/utils.h:352-358`) | Ctrl/Alt/Shift and Caps/Num/Scroll Lock only; any other key returns false (`src/unix/utilsx11.cpp:2596-2662`) |
| `WarpPointer()` | Apple's HIG forbids it; on Wayland it works only with a compositor implementing the pointer-warp protocol, and mutter also needs a pressed button (`interface/wx/window.h:3895-3914`; `docs/changes.txt:294`) | — |
| `wxUIActionSimulator` | "doesn't work when using Wayland" (`interface/wx/uiaction.h:20`) | — |
| `PopupMenu(x, y)` | — | GTK ≥ 3.22 positions it relative to the window (`gtk_menu_popup_at_rect`, `src/gtk/window.cpp:6520-6565`), so it is safe |
On wxOSX `WarpPointer` synthesizes a `wxEVT_MOTION` to the window (`src/osx/window_osx.cpp:1370-1398`),
so warping from a motion handler recurses. On GTK2, pointer queries and warps go straight to X11.
**OrcaSlicer.** Detect the backend with `Slic3r::GUI::is_running_on_wayland()` / `is_running_on_x11()`
(`LinuxDisplayBackend.hpp`), never with environment variables (`references/platforms.md`). Models:
- `SearchDialog` (`Search.cpp`) dismisses by focus on Wayland (`focus_left_popup`) instead of
comparing `wxGetMousePosition()` with its screen rectangle.
- `BBLTopbar::FindToolByCurrentPosition` (`BBLTopbar.cpp`) uses the last event position and returns
null on Wayland rather than query the global pointer.
- `BBLTopbar::OnMouseMotion` and `Button::OnParentMotion` use `ClientToScreen(event.GetPosition())`
within one window ("wxGetMousePosition() … returns (0,0) on Wayland").
### Pitfalls
- **Rule:** On Linux, never rely on global screen coordinates.
**Why:** wx delegates straight to GDK and does not compensate on Wayland, so hit tests against
`wxGetMousePosition()` silently fail or misfire.
```cpp
// Wrong:
if (!GetScreenRect().Contains(wxGetMousePosition())) Dismiss();
if (wxFindWindowAtPoint(wxGetMousePosition()) != this) /* ... */;
// Right: event-relative, same window (convert through the event's GetEventObject() window)
if (!GetClientRect().Contains(evt.GetPosition())) Dismiss();
// or, for popups and dialogs, focus tracking (focus_left_popup is file-local to Search.cpp):
if (focus_left_popup(this, wxWindow::FindFocus(), related)) Dismiss();
```
- **Rule:** Track held keys with KEY_DOWN/KEY_UP, not `wxGetKeyState(letter)`.
**Why:** it is always false for letters on Wayland, so held-key and repeat detection built on it
(including pruning a held-key record with it) breaks there. Clear such a record from KEY_UP.
## Hover handlers and enter/leave feedback loops
**Rule.** In `wxEVT_ENTER_WINDOW`/`wxEVT_LEAVE_WINDOW` handlers, only record hover state, per child
window, because enter and leave fire per window. Apply `Show()`/`Hide()` + `Layout()` from
`wxEVT_IDLE` or `CallAfter`, and only when the desired state differs from the current one.
`Skip()` the events.
**Why.** A layout change inside the handler moves windows under the pointer, which re-fires
enter/leave. On Wayland compositors that keep hidden-workspace surfaces mapped (Hyprland), GTK
delivers a stream of synthetic leave events, and a handler that re-layouts pegs a CPU core
(69e16cd7ef). The fix that moved show/hide out of the handlers removed the freeze caused by the
dynamically hidden printer edit button (e87625e023). Two wx facts block the obvious guards:
- `wxFindWindowAtPoint()` is a full window-tree walk on GTK and unreliable on Wayland (table above).
- `IsShownOnScreen()` is not a visibility test on any platform. It only checks `IsShown()` up the
parent chain (`src/common/wincmn.cpp:1205-1212`, `interface/wx/window.h:3101-3102`), plus "surface
exists" for `wxGLCanvas` on Unix. It never reflects minimised, occluded or other-workspace state,
which is why `Plater::priv::set_current_panel` says "wxWidgets IsShownOnScreen() is buggy and
cannot be used reliably".
**OrcaSlicer.** The Sidebar printer, nozzle and bed panels (`Plater.cpp`, `Sidebar` ctor) follow
this design:
- each child's ENTER/LEAVE lambda inserts or erases the window in a per-group
`std::shared_ptr<std::unordered_set<wxWindow*>>` (e.g. `printer_preset_hovered`) and sets the border
colour;
- a `wxEVT_IDLE` handler compares `!hovered->empty()` with `btn_edit_printer->IsShown()` and calls
`Show()`/`Hide()` + `Layout()` only on a difference (keep such a handler a cheap comparison: it runs on every
| MSW | thread keyboard hook `wxKeyboardHook` → `wxEVT_CHAR_HOOK` to the focus window (or the active window) — skipped while any native `::GetCapture()` exists or a non-wx modal (IME) is open; with IME open a handled hook does not stop the key reaching the IME (`src/msw/window.cpp:7320-7395`) → `wxGUIEventLoop::PreProcessMessage`: accelerator walk from the focus up to the first `IsTopNavigationDomain(Navigation_Accel)` window, with the text-entry exemption (`src/msw/evtloop.cpp:62-119`) → `IsDialogMessage`, which is never given `VK_ESCAPE` (`src/msw/window.cpp:2769-2778`) → `wxEVT_KEY_DOWN` → `wxEVT_CHAR` |
| GTK | `wxEVT_CHAR_HOOK`, skipped while `g_captureWindow` is set → accelerator walk to the top-level window, no text-entry exemption → `wxEVT_KEY_DOWN` → input method → `wxEVT_CHAR` (`src/gtk/window.cpp:1266-1420`) |
| macOS | the main menu's `performKeyEquivalent:`: a menu-bar key equivalent consumes the key before wx sees it (`src/osx/cocoa/window.mm:1611-1644`) → `wxEVT_CHAR_HOOK`, also during capture → accelerator walk, only if no `wxEVT_CHAR_HOOK` handler processed the event (`src/osx/window_osx.cpp:2547-2590`) → Tab navigation by the first `wxTAB_TRAVERSAL` ancestor unless the window has `wxWANTS_CHARS` (`src/osx/cocoa/window.mm:4009-4046`) → `interpretKeyEvents` (IME) → `wxEVT_KEY_DOWN` → `wxEVT_CHAR` (`:4048-4101`). A disabled window gets no key events (`:1616-1617`) |
### wxEVT_CHAR_HOOK
Contract (`interface/wx/event.h:1487-1508`): "Unlike all the other key events, this event is
propagated upwards the window hierarchy … generated before any other key events". If a handler
processes it without `Skip()`, "neither wxEVT_KEY_DOWN nor wxEVT_CHAR events will be generated
(although wxEVT_KEY_UP still will be)". `DoAllowNextEvent()` handles it and still lets normal events
through (`:1683-1702`). It "is not generated when the mouse is captured". [source] Exceptions and
details:
- **macOS** has no capture check, so the hook fires during drags.
- **wxGTK** reads the allow flag from the original event rather than the hook event, so
`DoAllowNextEvent()` has no effect there (`src/gtk/window.cpp:1266-1282`).
- Propagation stops only at windows with `wxWS_EX_BLOCK_EVENTS` (`src/common/wincmn.cpp:3498-3522`).
`wxDialog` sets it (`src/common/dlgcmn.cpp:127`), and so does `wxPopupWindow` on MSW and macOS
| `RawControlDown()`, `wxMOD_RAW_CONTROL`, `WXK_RAW_CONTROL`, `wxACCEL_RAW_CTRL`, `"RawCtrl+"` | Ctrl (same value as the above) | the physical Control key (a distinct bit) |
| `CmdDown()`, `wxMOD_CMD`, `wxACCEL_CMD` | deprecated aliases of `ControlDown()`/`wxMOD_CONTROL`/`wxACCEL_CTRL` | same |
- **Rule:** Don't expect `MainFrame`'s keys inside a dialog: each top-level window needs its own
handling (dialogs block `wxEVT_CHAR_HOOK` propagation and the accelerator walk stops at them).
## Orca's shortcut registry
The registry is the one table every key dispatcher, menu label, tooltip and the shortcuts dialog
reads; the design is in `docs/HLSD/keyboard-shortcuts.md`.
**Data model.**
- `KeyChord` (`KeyChord.hpp`) is `{key, modifiers}`: the key as `wxEVT_KEY_DOWN` reports it, plus
`wxMOD_*` limited to CONTROL|SHIFT|ALT|RAW_CONTROL. Its canonical text (`to_string()`, e.g.
`Ctrl+Shift+S`) is platform-neutral and doubles as the wx accelerator string and the config format.
`display()` uses translated modifier names and the macOS glyphs. Predicates:
- `needs_char_event()`: a printable non-alphanumeric key with at most Shift, resolvable only from
the CHAR that follows.
- `is_punctuation()`: such a key with no modifier.
- `is_menu_accelerator()`: Ctrl, Alt or RawCtrl held, or a non-printable key other than Space.
- `is_system_shortcut()`: Alt+F4 and Alt+Space on Windows.
- `to_accelerator_entry()` maps RawCtrl to `wxACCEL_RAW_CTRL` on Apple.
- `Shortcut` (enum) and `shortcut_table` (`Shortcuts.cpp`): each row, written with
`SHORTCUT`/`REPEATING`/`STEPPING`, holds a config key, description, context mask, default chord, and
the `repeatable`/`modifier_variants` flags. `static_assert`s keep the table in enum order and
`section_table` ascending.
- `ShortcutRegistry` (`wxGetApp().shortcuts()`) overlays user overrides from the AppConfig section
`shortcuts`; only overrides are stored, and `none` records an unbound shortcut.
`ShortcutRegistry::load()` drops a Global override that fails `is_menu_accelerator()` ("A Global
chord is seen before any text field").
- Lookups: `lookup(context, chord)` matches exactly. `match()` falls back to `modifier_variants`
shortcuts with Shift/Ctrl added, which are reserved steps (`step_owner()`). `conflicts()` treats
Global as sharing every context.
**Contexts and dispatch.**
| Context | Dispatcher | Event |
|---|---|---|
| `Global` | `MainFrame::handle_global_shortcut`, from `MainFrame`'s `wxEVT_CHAR_HOOK`, before the focused child gets KEY_DOWN/CHAR (a child's own CHAR_HOOK handler still runs first); `Skip()` otherwise | CHAR_HOOK |
| `Plater`, `Preview` | `GLCanvas3D::handle_shortcut` (→ `ShortcutRegistry::match`) from `GLCanvas3D::on_key`; punctuation retried from `GLCanvas3D::on_char` | KEY_DOWN, CHAR |
| `ObjectList` | `ObjectList::dispatch_shortcut`, from `ObjectList::key_event` on CHAR (non-macOS); on macOS from a `wxAcceleratorTable` built by `ObjectList::update_shortcut_accelerators()` (ids from `wxWindow::NewControlId`), because the native data view delivers no keys there | CHAR / accelerator |
`sys_color_changed`, `Sidebar::msw_rescale` / `sys_color_changed`); nothing calls it otherwise. → [DPI and colour fan-out](#dpi-and-colour-fan-out)
9. Process and model-scope settings are in `ParamsPanel` inside the sidebar; filament and printer
settings are in the modeless `ParamsDialog`. `get_tab()` returns null until the tab is complete. → [Settings placement](#settings-placement-paramspanel-paramsdialog-tabs)
10. `Plater` and `Sidebar` are pimpl'd: new state goes into `Plater::priv` / `Sidebar::priv`. New
events are declared next to their emitter with the `Event.hpp` types; a short-lived listener on
the plater binds through `EventGuard`. → [Plater and Sidebar](#plater-and-sidebar)
11. Docked panes go through `Plater::add_dock_pane` with a stable, untranslated, delimiter-free name;
the window must be a child of the plater. → [Docking](#docking)
12. UI drawn over the 3D view is ImGui inside `GLCanvas3D`, never a wx child window over the GL
canvas; ask for a redraw with `set_as_dirty()` / `request_extra_frame()`. → [3D canvas](#3d-canvas-imgui-layer-and-notificationmanager)
13. `NotificationManager` is called on the UI thread only, and its notifications are visible only
while the plater is shown. → [NotificationManager](#notificationmanager)
14. UI-initiated background work is a `Job` on a `Worker`; slicing is `BackgroundSlicingProcess`;
network agents call back through `wxGetApp().CallAfter`. UI is touched only on the main thread. → [Background work](#background-work)
15. Device UI pulls state from `DeviceManager` on its own timer and mutates `MachineObject` only on
the UI thread. → [Device pages](#device-and-monitor-pages)
16. Web UI goes through Orca's hosts (`WebView::CreateWebView`, `WebViewHostDialog`, `WebPanel`,
`DockPanel`); window operations requested from a script message are deferred and liveness-checked. → [Web UI](#web-based-ui)
17. A preference is a `create_item_*` row in `PreferencesDialog::create_items` that writes
`app_config` and saves at once; its default goes in `AppConfig::set_defaults`; effects needed
after the dialog closes go in `GUI_App::open_preferences`. → [Preferences](#preferences)
18. `AppConfig` values are strings: match the key's own convention (`"true"/"false"` or `"1"/"0"`);
`save()` and every write run on the main thread. → [AppConfig](#appconfig)
19. Every new source file is registered in `src/slic3r/CMakeLists.txt`: `SLIC3R_GUI_SOURCES`, or the
`if (WIN32)` / `if (APPLE)` / `if (SLIC3R_CAD)` blocks, or the `GUI/DeviceCore` / `GUI/DeviceTab`
| `MainFrame` | `DPIFrame`; `GUI/MainFrame.hpp/.cpp` | borderless main window, top bar / menu bar, tab book, preset tabs, idle prebuild, DPI/colour fan-out | `wxGetApp().mainframe` |
| `Plater` | `wxPanel`, pimpl `Plater::priv`; `GUI/Plater.hpp/.cpp` | the Prepare and Preview page: model, three canvases, AUI docking, slicing, job worker, notifications, context menus | `wxGetApp().plater()` |
| `ParamsPanel` | `wxPanel`; `GUI/ParamsPanel.hpp` | process and model-scope `Tab`s, reparented into the sidebar | `wxGetApp().params_panel()` (null-safe) |
| `ParamsDialog` | `DPIDialog`; `GUI/ParamsDialog.hpp` | its own `ParamsPanel` with the filament and printer `Tab`s; modeless | `wxGetApp().params_dialog()` (null-safe) |
| Web hosts | `Widgets/WebView`, `Widgets/WebViewHostDialog`, `WebViewDialog.hpp` (`WebViewPanel`), `PrinterWebView`, `WebPanel`, `DockPanel`, `WebDialog` | HTML UI | per class |
`load_current_presets()`, `mainframe->Show(true)`; the splash is destroyed; `update_mode()`.
11. The app-level `wxEVT_IDLE` handler is bound.
### post_init and the app idle handler
The idle handler bound at the end of `on_init_inner` runs `post_init()` exactly once (guarded by
`m_post_initialized`, and postponed while a WebView script handler is being added), then on every
idle **saves `app_config` if it is `dirty()`**. `post_init` initialises the WebView2 runtime on
Windows (`init_webview_runtime`, before the first WebView), opens command-line files, loads the GL
resources on the Prepare canvas when the app starts on Prepare (when it starts on Home they load
later as an idle task, so Home paints first), starts the idle prebuild
(`MainFrame::prebuild_pages_when_idle`), and `CallAfter`s the config wizard and update checks — the
code comment: on Mac this is "the only way to popup a modal dialog on start without screwing combo
boxes". If the GL context cannot be made current yet, Linux resets `m_post_initialized` so the next
idle retries (a Wayland surface commits late).
### Accessors
| Accessor | Null? |
|---|---|
| `app_config`, `preset_bundle`, `imgui()`, `shortcuts()`, `action_registry()` | created in the constructor or before `MainFrame`; non-null for the GUI lifetime (`preset_updater` is null in the G-code viewer) |
| `mainframe`, `plater()` | null before `MainFrame` exists |
| `sidebar()`, `obj_list()`, `model()` | dereference `plater_`**without a check** |
| `params_panel()`, `params_dialog()`, `notification_manager()` | null-safe (return null without a main frame / plater) |
| `get_tab(Preset::Type)` | null for a tab not found **or not yet `completed()`**; `tabs_list` / `model_tabs_list` are cleared by `MainFrame::shutdown` |
| `get_model_tab(part)`, `get_layer_tab()` | index `model_tabs_list`**without a bounds check** — undefined once `MainFrame::shutdown` has cleared it |
| `getDeviceManager()`, `getAgent()` | may be null; check before use |
| `em_unit()` | app-wide; per-window value via the free `em_unit(wxWindow*)` — `references/dpi-bitmaps-fonts.md` |
| `dark_mode()` | static, recomputed per call — `references/colours-dark-mode.md` |
| `is_closing()`, `is_recreating_gui()`, `input_idle_ms()` | state flags; `input_idle_ms` is fed by `GUI_App::FilterEvent` |
### recreate_GUI (language switch)
`GUI_App::recreate_GUI` sets `m_is_recreating_gui`, destroys the cached Speed Dial dialog (its
translated strings are injected once), calls `mainframe->shutdown()`, swaps the `Field` control pools
(`switch_window_pools()`; the old pools are released only when the old frame is destroyed), creates a
**new `MainFrame`**, `Destroy()`s the old one, reloads presets, shows the new frame and calls
`prebuild_pages_when_idle()` again. `GUI_App::shutdown` returns early while recreating, so
`is_closing()` never becomes true during a language switch.
Consequences: every `MainFrame` child, `Tab`, lazy panel and cached dialog is a new object afterwards.
The `LazyInstance` statics follow automatically (the new frame's holders replace the old ones); raw
pointers do not. Settings fields are recycled through pools: `references/orca-settings-ui.md`.
### Pitfalls
- **Rule:** On startup and shutdown paths, test `plater()` before `sidebar()` / `obj_list()` / `model()`.
**Why:** those accessors dereference `plater_` unchecked; before `MainFrame` exists they crash.
## Deferred construction (Lazy, LazyPage, StagedBuild, IdleScheduler)
Design doc: `docs/HLSD/deferred-page-construction.md`. Startup pays only for what the first frame
shows (the start page and the Prepare plater); every other tab, and heavy dialogs and GL resources,
build on first show or in small units while the user is idle. A click during the idle build waits for
one unit at most. The parts are independent and wx-free where possible (`Lazy`, `StagedBuild`,
`PrebuildQueue` are unit-tested in `tests/slic3rutils`: `test_lazy.cpp`, `test_staged_build.cpp`,
`test_prebuild_queue.cpp`).
### The holder: `Lazy<T>` and `LazyInstance<T>`
`Lazy<T>` (`GUI/Lazy.hpp`) holds a factory and the object it makes: `Lazy(name, order, factory)`.
| Member | Contract |
|---|---|
| `get()` | the object, **null until completely built** (a staged object mid-build is null) |
| `ensure()` | builds whatever is left now (busy cursor + log line) and returns the object; null if the factory returned null or a nested call finds it mid-build |
| `when_built(fn)` | runs `fn` now if built, otherwise once the build completes |
| `build_step()` | one unit: the factory first, then one `StagedBuild` step per call; a nested call (a unit that pumps the loop) does nothing |
| `prebuild_order()` | position in the idle queue; lower first; **negative = never prebuilt** |
The holder does not own the object — its wx parent does. A factory that returns null or a unit that
throws leaves the holder and scheduler able to carry on. `LazyInstance<Self>` is a mixin that gives a
type with one instance app-wide the statics `Self::if_built()`, `Self::ensure()`,
`Self::when_built(fn)`; the `Lazy<Self>` constructor registers itself, and a recreated `MainFrame`'s
holder replaces the old one. All statics are harmless (null / no-op) while no holder exists —
including `when_built`, which then drops `fn`.
### The placeholder page: `LazyPage<Panel>`
`LazyPage<Panel> : wxPanel, Lazy<Panel>` (`GUI/LazyPage.hpp`) is the notebook page (the book needs a
page object to insert and remove by pointer). `LazyPage(parent, name, order, factory)`; the default
factory is `new Panel(parent)`. Its `Show(true)` builds the panel the first time (only once the
top-level frame is shown — `MainFrame::Show` completes the start page on the frame's first show) and
forwards later shows/hides to the panel, so the panel's own `Show()` override stays its activation
hook. The build runs before the placeholder's own `wxPanel::Show(true)`, so an on-demand build creates
its controls in a hidden window as a prebuild does: on MSW each control created or moved inside a
shown window re-clips and repaints its shown siblings, which made a large panel's first show take
seconds. A lazy panel's constructor therefore runs off screen (except the start page's) and must not
rely on `IsShownOnScreen()`. A panel built while its page is hidden stays hidden, and `when_built`
gives it the dark-UI pass the frame ran before it existed (`apply_dark_ui_to_lazy_panel`).
`pending()` is true only while the page is in the book.
### Staged construction: StagedBuild
`StagedBuild` (`GUI/StagedBuild.hpp`) splits a constructor too big for one unit: the constructor builds
a skeleton and queues the rest with `add_build_step(fn)`; `add_build_steps_of(child)` forwards a child
panel's steps, and the parent is `built()` only once every child is. Constraints, all from the design:
- members created in steps start null, so a partly built panel can be destroyed;
- timers, event handlers and the destructor that touch step content check `built()` first;
- nothing takes focus while off screen (a unit may run while the user types elsewhere);
- a widget added by a step keeps its place through an empty sizer slot the skeleton creates.
### The idle scheduler: IdleScheduler, PrebuildQueue
`PrebuildQueue` (`GUI/PrebuildQueue.hpp`) orders `LazyBase` tasks by `prebuild_order()` (equal order:
insertion order) and runs one slice of units of the first pending task. `IdleScheduler`
(`GUI/IdleScheduler.hpp/.cpp`, `MainFrame::m_idle`) drives it from a self-owned `wxTimer`:
- it ticks every 250 ms and runs a slice only after 500 ms without user input (`GUI_App::input_idle_ms`,
stamped by `GUI_App::FilterEvent` for non-command user-input events and main-frame resizes);
- a slice spends at most 40 ms, then the next slice is `StartOnce(5)` — a separate timer message, so
paint, timers and input queued meanwhile run first. Posting slices as pending events would not do
that, because wx drains every pending event, including ones posted meanwhile, before the next
native message (`src/common/appbase.cpp``wxAppConsoleBase::ProcessPendingEvents` loops until the
list is empty **[source]**);
- it skips while `wxEventLoopBase::GetActive()->IsYielding()` (a slice inside a `wxYield()` would build
pages in the middle of the code that yielded) and guards re-entry with `m_in_slice`;
- it stops its timer when nothing is pending, so it costs nothing afterwards.
`MainFrame::prebuild_pages_when_idle` (called from `post_init` and `recreate_GUI`) clears the queue
and registers the GL resources (`GLResourcesPrebuild`), the Prepare settings page one option group at
a time (`ParamsPanel::settings_page_prebuild`), the Prepare layout at the book's page size
(`m_prepare_layout_prebuild`), every lazy page with a non-negative order, and the lazily built
dialogs (`m_diff_dialog`); the queue then runs them by `prebuild_order()`. `MainFrame::shutdown`
stops it. Units should fit in one slice on a fast machine; a constructor over that is staged.
**Platforms.** GTK: a timer that is always due (`g_timeout_add`, default priority,
`src/gtk/timer.cpp`) runs ahead of the lower-priority GLib sources that repaint and that deliver
posted events and idle (wx's single `G_PRIORITY_LOW` idle source, `src/gtk/app.cpp`
`wxApp::WakeUpIdle`**[source]**) — hence the 5 ms gap rather than 0. macOS: wxOSX rejects a 0 ms
timer (`src/osx/core/timer.cpp:74``wxCHECK_MSG(m_milli > 0, …)`**[source]**; with asserts compiled
out, `StartOnce(0)` silently never fires). Windows: a slice also waits while the native queue holds input
(`GetQueueStatus`), not counting mouse moves, which Windows synthesises when a window appears under the
cursor. GTK GL resources: the prebuild task `gtk_widget_realize`s the hidden canvas before making the
context current, since GTK creates the surface only on realize.
### Reaching a lazy object
| Need | Use |
|---|---|
| work the object can live without (refresh, status update) | `if (X* x = X::if_built()) x->…;` |
| navigating to it or showing it | `X::ensure()->…` (as `MainFrame::jump_to_monitor`) |
| state it would not fetch for itself when constructed; rescale/recolour of a staged panel (null from `if_built()` while mid-build) | `X::when_built([](X& x) { … });` |
A panel that pulls its own state in its constructor only ever needs `if_built()`.
### Usage
The shape to copy for a new tab (`MainFrame::init_tabpanel`):
```cpp
// Panel: one instance app-wide; heavy constructors also derive StagedBuild
class CalibrationPanel : public wxPanel, public StagedBuild, public LazyInstance<CalibrationPanel> { … };
// MainFrame::init_tabpanel: id, order (gaps leave room between neighbours; <0 = never prebuilt)
m_calibration_page = new LazyPage<CalibrationPanel>(m_tabpanel, TAB_ID_CALIBRATION, 30);
`itInfo`). It lives in the sidebar, is initialised by `obj_list()->init()` after the main frame is
created, gets keys through the shortcut registry (`ObjectList::dispatch_shortcut`; on macOS a
`wxAcceleratorTable` regenerated by `update_shortcut_accelerators`, because the native control
delivers no key events), and shows context menus through `MenuFactory` + `Plater::PopupMenu`. Per-object
overrides appear as `itSettings` children that open the model-scope tabs. Model ownership, renderers,
drag and drop, native-vs-generic data view: `references/controls-dataview.md`.
## 3D canvas, ImGui layer and NotificationManager
### GLCanvas3D
`GLCanvas3D` is **not a window**: it wraps a `wxGLCanvas* m_canvas` (`get_wxglcanvas()`) created by
`OpenGLManager`, binds its size/idle/key/mouse/paint/focus/timer handlers in `bind_event_handlers`, and
must be unbound before teardown (`Plater::unbind_canvas_event_handlers`, from `MainFrame::shutdown`).
Rendering is idle-driven: handlers mark `set_as_dirty()`, `request_extra_frame()` or
`schedule_extra_frame(ms)`, and `on_idle` renders. Outgoing events go through `GLCanvas3D::post_event`.
The view, preview and assemble canvases and the UV editor share one `wxGLContext`. Paint/idle/swap
details, the shared-context attribute rule and EGL/GLX: `references/webview-gl-aui-media.md`.
### What is ImGui and what is wx
| Drawn with ImGui inside the canvas | wx windows |
|---|---|
| gizmo panels (`GLGizmoBase::on_render_input_window`), `NotificationManager` and its hint / slicing-progress notifications, the preview layer slider (`IMSlider`), `IMToolbar`, the G-code legend (`GCodeViewer`), plate labels (`PartPlate`), and the overlays in `GLCanvas3D::_render_overlays` (plate-select toolbar, variable-layer-height dialog, 3D navigator, toolbar item windows) | everything outside the canvas: sidebar, tabs, dialogs, top bar, Home/Device/Project pages |
`GLToolbar` is not ImGui: its icons are OpenGL-textured quads; only its item option windows are
ImGui callbacks. ImGui input arrives only through the canvas's own handlers (`ImGuiWrapper::update_mouse_data` /
`update_key_data`), so text entry needs canvas focus; ImGui sizes are physical pixels (scale by
`GLCanvas3D::get_scale()`); ImGui strings are UTF-8 (`_u8L`). wx theming, `DPIDialog`, sizers and
`Widgets/` do not apply there.
- **Rule:** Never place a wx child window over the GL canvas; draw the overlay in ImGui or put a wx
window beside the canvas.
**Why:** on GTK the GL canvas is a native child window or, on Wayland, a subsurface drawn outside
GTK, and a wx child over it does not reliably stack above the GL content.
```cpp
// Wrong
auto* banner = new wxPanel(canvas->get_wxglcanvas());
// Right: ImGui from the gizmo / overlay pass, or a sibling of the canvas in the plater layout
void on_render_input_window(float x, float y, float bottom_limit) override; // GLGizmoBase
```
Cite: `docs/HLSD/design-tab.md` (sketch banner "a sibling of the canvas, not a child over it").
### NotificationManager
`NotificationManager` (`GUI/NotificationManager.hpp`) is owned by `Plater::priv` and initialised after
the canvas exists (`Plater::init_notification_manager`; notifications pushed before `init()` are
| UI-initiated task (arrange, orient, fill bed, send) | `Job` subclass in `GUI/Jobs/`; `replace_job(plater->get_ui_job_worker(), std::make_unique<OrientJob>())`, or a dialog-owned `PlaterWorker<BoostThreadWorker>` | `Job::finalize` and `Ctl::call_on_main_thread`, delivered from the owner window's idle/paint |
| slicing and export | `BackgroundSlicingProcess` (`Plater::priv::background_process`) | `wxQueueEvent(plater, evt.Clone())`; `execute_ui_task` for a synchronous UI call |
The contracts (which side runs what, cancellation, the `eptr` rethrow, deadlock rules, platform stalls)
are in `references/threads-timers-app.md`.
## Device and Monitor pages
Design doc: `docs/HLSD/printer-agent.md`; implementing a printer agent: the `orca-printer-communication`
skill. Data flow:
1. `NetworkAgent` (façade over the active `IPrinterAgent` and the cloud agents) calls the callbacks
installed by `GUI_App::init_networking_callbacks` (`set_on_message_fn`, `set_on_local_message_fn`,
`set_on_printer_connected_fn`, `set_queue_on_main_fn`, …) and by `GUI_App::post_init` /
`restart_networking` (`set_on_ssdp_msg_fn`) on its own threads.
2. Each callback returns if `is_closing()`, then `CallAfter`s a by-value lambda that re-checks
`is_closing()` and, on the UI thread, updates the `MachineObject` (`parse_json`), refreshes
`Sidebar::load_ams_list` and `Plater::update_machine_sync_status`. `MachineObject` and
`DeviceManager` state is main-thread-only.
3. The Device UI is **pull-based**: `MonitorPanel : wxPanel, StagedBuild, LazyInstance<MonitorPanel>`
(`GUI/Monitor.hpp`) starts its refresh `wxTimer` in its `Show(true)` override, stops it on hide, and
`on_timer` → `update_all()` reads `DeviceManager::get_selected_machine()` and pushes it into the
`StatusPanel` (`StatusBasePanel : wxScrolledWindow, StagedBuild`), HMS and media pages inside a
`Tabbook`. `DeviceManager::start_refresher` / `stop_refresher` follow the main frame's `wxEVT_SHOW`.
4. Camera playback: `MediaPlayCtrl` selects and tears down the stream backend; the wx parent owns the
rendering window.
New device UI goes inside `MonitorPanel` / `StatusPanel`, reads state on the timer, makes no network
call on the UI path, and stops its timers on hide.
## Web-based UI
| Host | Use |
|---|---|
| `WebView::CreateWebView(parent, url)` (`Widgets/WebView.hpp`) | the sanctioned way to make a browser: backend choice, handlers, user agent, `"wx"` script handler once per view, registration for `WebView::RecreateAll()` theming, a `FakeWebView` stub instead of null on failure. A raw `wxWebView::New` view gets none of these (no theming on colour change, no null safety) |
| `WebViewPanel` (`WebViewDialog.hpp`, `LazyInstance`) | the Home tab |
| `PrinterWebView` (`LazyInstance`) | the web Device tab (Fluidd/Mainsail/printer UIs) |
| `get_bool(section, key)` | `get(section, key) == "true" \|\| get("app", key) == "1"` — the `"1"` is read from **`"app"`** |
| `set(key, value)`, `set(section, key, value)`, `set_str(section, key, value)`, `set(section, key, bool)` | marks dirty only when the value changes; the `bool` overload writes `"true"`/`"false"`, and a bare `const char*` value selects it — pass a `std::string` or use `set_str` (`references/strings-i18n-files.md` §AppConfig) |
| `set_bool(key, value)` | `"true"`/`"false"` in `"app"` |
| `has(section, key)`, `dirty()`, `save()` | `save()` throws `CriticalException` off the main thread |
Persistence: the app idle handler saves whenever `dirty()` after `post_init`, and `MainFrame::shutdown`
saves if dirty, so a `set` persists on its own; an explicit `save()` is for immediate persistence
(Preferences rows, dark-mode init). Keys use both conventions — `set_bool` keys hold `"true"/"false"`,
others hold `"1"/"0"` (`dark_color_mode`, `default_page`, `sys_menu_enabled`) — so compare with the
key's own convention.
- **Rule:** Read a non-`"app"` boolean with `get(section, key)` and an explicit comparison.
**Why:**`get_bool(section, key)` accepts `"1"` only from the `"app"` section.
```cpp
// Wrong: false when section/key holds "1"
bool on = app_config->get_bool("section", "key");
// Right
bool on = app_config->get("section", "key") == "1";
```
- **Rule:** Write `app_config` on the main thread only.
**Why:** the storage map is unsynchronised and `save()` throws off the main thread.
```cpp
// Wrong: on a worker or agent thread
wxGetApp().app_config->set("key", value);
// Right
wxGetApp().CallAfter([value] { if (!wxGetApp().is_closing()) wxGetApp().app_config->set("key", value); });
```
Cite: `AppConfig::save`.
## Where new code goes
| You add | Put it | Must also |
|---|---|---|
| Modal dialog | `GUI/<Name>Dialog.hpp/.cpp`, `class X : public DPIDialog` | follow the dialog recipe in `references/windows-dialogs.md` (parent fallback `wxGetApp().mainframe`, `on_dpi_changed`, `SetSizerAndFit`, `UpdateDlgDarkUI` last) |
| Message / confirm box | `MessageDialog`, `RichMessageDialog`, `WarningDialog`, `ErrorDialog`, `InfoDialog` (`MsgDialog.hpp`), or `show_error` / `show_info` (`GUI.hpp`) | never `wxMessageBox`; `show_error` is asynchronous (an app `CallAfter` around an `ErrorDialog`), so pass a parent that outlives the call or none; `show_info` is a synchronous modal `MessageDialog` — `references/windows-dialogs.md` |
| Top-level tab | `LazyPage<Panel>` + `TAB_ID_*` in `MainFrame::init_tabpanel` | panel derives `LazyInstance<Panel>` (+ `StagedBuild` if heavy); fan-out hooks with `when_built`; statics outside `MainFrame`; no focus off screen |
| Heavy dialog owned by the frame | `Lazy<Dlg>` member of `MainFrame`, `Dlg : LazyInstance<Dlg>` | register in `prebuild_pages_when_idle` to prebuild; `if_built()` in the colour fan-out |
| Sidebar control | `Sidebar::Sidebar`, state in `Sidebar::priv` | `SidebarProps` spacing; add to `Sidebar::msw_rescale`, `sys_color_changed`, `update_mode` |
| Docked pane | `Plater::add_dock_pane` | window is a plater child; stable name; know that `remove_dock_pane` skips `on_close` |
| Print / filament / printer setting | def in `PrintConfig.cpp`, row in `Tab*::build` | the full checklist in `references/orca-settings-ui.md` |
| Overlay or tool UI in the 3D view | gizmo `on_render_input_window` or `GLCanvas3D::_render_overlays` | ImGui + `_u8L`; redraw via `set_as_dirty()` / `request_extra_frame()`; GL only in the canvas's current context |
| Transient message about the 3D view | `NotificationManager::push_notification` | UI thread; new `NotificationType` only to close/update by type |
| Main-menu item | the shared `wxMenu` in `MainFrame::init_menubar_as_editor` / `generate_help_menu` | `append_menu_item`, or `append_shortcut_item` when it has a shortcut — `references/popups-menus.md` |
| Context-menu item | `MenuFactory` (`GUI_Factories.cpp`) | show with `Plater::PopupMenu` |
| Speed Dial command | the `NativeCommands` catalog (`NativeCommands.cpp`, `NativeCommand{key, title, group, input, icon, runner}`) | `ActionRegistry` stores and dispatches it |
| Keyboard shortcut | `Shortcut` enum + `shortcut_table` (`Shortcuts.cpp`) | handle in the context's dispatcher; labels from the registry — `references/mouse-keyboard-focus.md`, `docs/HLSD/keyboard-shortcuts.md` |
| Preference | a `create_item_*` row in `PreferencesDialog::create_items`; default in `AppConfig::set_defaults` | effects in the row handler; post-close effects in `GUI_App::open_preferences` |
| Background task | `Job` subclass in `GUI/Jobs/` | UI only in `finalize` / `call_on_main_thread`; poll `was_canceled()` — `references/threads-timers-app.md` |
| Device UI | inside `MonitorPanel` / `StatusPanel` | pull from `DeviceManager::get_selected_machine()` on the timer; no network calls on the UI path |
| Reusable control | `GUI/Widgets/` | `references/orca-widgets.md`, `references/painting-custom-widgets.md` |
| Tests for wx-free GUI logic | `tests/slic3rutils/test_<subsystem>.cpp` (as `test_lazy.cpp`, `test_shortcuts.cpp`) | list the file in that suite's `CMakeLists.txt` (`tests/AGENTS.md`) |
## Build registration
`src/slic3r/CMakeLists.txt` defines `SLIC3R_GUI_SOURCES`, the list compiled into `libslic3r_gui`. It
covers everything under `src/slic3r` (`GUI/`, `GUI/Widgets/`, `GUI/Jobs/`, `Utils/`, `Config/`,
`plugin/`), with paths relative to `src/slic3r`. Add a new `.cpp`/`.hpp` pair on consecutive lines
(the list is only roughly alphabetical). Additional places:
| macOS Objective-C++ (`.mm`) and their headers | `if (APPLE) list(APPEND SLIC3R_GUI_SOURCES …)` |
| Design/CAD UI | the `if (SLIC3R_CAD) list(APPEND …)` block; shared code that references it is guarded with `#ifdef SLIC3R_CAD` (the root `CMakeLists.txt` adds the definition) |
| `GUI/DeviceCore/`, `GUI/DeviceTab/` | their own `CMakeLists.txt`, included with `add_subdirectory`, which `list(APPEND SLIC3R_GUI_SOURCES …)` and re-export it with `PARENT_SCOPE` |
- **Rule:** Keep platform-only sources out of the shared list.
**Why:** an `.mm` file or a Win32-only header in the shared list breaks the other platforms' builds.
```cmake
# Wrong: in the shared set(SLIC3R_GUI_SOURCES …) list
GUI/GUI_UtilsMac.mm
# Right
if (APPLE)
list(APPEND SLIC3R_GUI_SOURCES
GUI/GUI_UtilsMac.mm
)
endif ()
```
## Design docs (docs/HLSD)
Per `AGENTS.md`, the high-level design of a subsystem goes in `docs/HLSD/<subsystem>.md`, describes
the design as it stands (no phases or before/after framing), and is updated in the same PR when a change
invalidates it. Planning output stays in the gitignored `docs/superpowers/`. GUI-relevant documents:
| Doc | Covers |
|---|---|
| `docs/HLSD/deferred-page-construction.md` | `Lazy`, `LazyPage`, `StagedBuild`, `IdleScheduler`, `PrebuildQueue`, GL-resource prebuild; rules for reaching lazy objects, unit size, order |
| `docs/HLSD/keyboard-shortcuts.md` | `KeyChord`, `Shortcut` / `shortcut_table`, `ShortcutRegistry`, contexts and dispatchers, labels, the shortcuts dialog, "Adding a shortcut" |
1. Declare a setting once, in `PrintConfigDef::init_*_params()`, with `label`, `category`, `tooltip`, `sidetext`,
`mode`, limits and a default; mark every user-visible string with `L()`, never `_L()`. → [ConfigOptionDef](#configoptiondef-the-data-side)
2. Put the member in the `PRINT_CONFIG_CLASS_DEFINE` block of the scope it belongs to; the class decides whether
the setting can be overridden per object, part or layer range. → [Config classes](#config-classes-preset-lists-and-variants), [Overrides](#per-object-part-layer-and-plate-overrides)
3. List the key in the `Preset` option list of its preset type; without it the key is absent from the tab's
config and the row cannot be built. → [Config classes](#config-classes-preset-lists-and-variants)
4. A per-variant key is a vector option listed in the matching `*_options_with_variant` set, appended with index
`0`, and toggled with the variant index. → [Variants](#per-extruder-variant-options)
5. Add the row with `optgroup->append_single_option_line(key, wiki_path)` in `TabX::build()`; widget, tooltip,
undo/system icons, dirty tracking and search all come from the def. → [Groups](#optionsgroup-option-and-line)
6. Page titles, group titles, `Line` labels and tooltips are English `L()` strings; translation happens at
Enums need three parts: the `enum class` plus `CONFIG_OPTION_ENUM_DECLARE_STATIC_MAPS(Name)` in `PrintConfig.hpp`,
a `static t_config_enum_values s_keys_map_Name` plus `CONFIG_OPTION_ENUM_DEFINE_STATIC_MAPS(Name)` in
`PrintConfig.cpp`, and in the def `enum_keys_map = &ConfigOptionEnum<Name>::get_enum_values()` with parallel
`enum_values` (serialized keys) and `enum_labels` (`L()` display labels) — see `wall_generator`
(`PerimeterGeneratorType`).
| Field | Meaning for the GUI |
|---|---|
| `type` | `coFloat/coFloats/coInt/coInts/coString/coStrings/coPercent(s)/coFloatOrPercent(s)/coBool(s)/coEnum(s)/coPoint(s)/…`; picks the `Field` when `gui_type` is `undefined`. |
| `gui_type` | `GUIType {undefined, i_enum_open, f_enum_open, color, select_open, slider, legend, one_string, plugin_picker, plugin_config, printer_agent_select}`; checked first by `build_field`. `slider` is marked "currently unused" in `Config.hpp`. |
| `gui_flags` | `"serialized"`: a vector edited as one `;`-separated string; `"show_value"`: show the value even when enum labels exist. |
| `label` / `full_label` | `label` is the short row label (a sub-label inside a multi-option row). `full_label`, when set, names the setting on its own: sidebar search titles, the per-object "Add Settings" menu, the compare dialogs. The Speed Dial titles a setting with the label its row draws and keeps `full_label`/`label` only as a search alias (`ActionRegistry`). |
| `category` | English group name for per-object settings (`SettingsFactory::get_bundle`, the settings menus, the ObjectList settings item) and the transfer view's fallback in `UnsavedChangesDialog`. Empty = left out of the ObjectList settings item and the "Add Settings" menus (the model tabs still show the key). Search uses the tab page title instead. |
| `tooltip`, `sidetext` | Translated at display; `sidetext` is drawn inside most inputs (see [Field](#field-and-the-value-flow)). |
| `min`, `max`, `max_literal` | `Field` clamps to `[min, max]` with "Value is out of range."; `max_literal` bounds the absolute (non-%) value of `coFloatOrPercent` keys whose `sidetext` contains `"mm "`. Both `min`/`max` bounded → "Range:" line in the tooltip. |
| `mode` | `comSimple < comAdvanced < comExpert < comDevelop`; a row shows when its **first** option's mode ≤ the tab's mode. |
| `nullable` | Vector values may hold nil ("N/A"); used by model-scope overrides. |
| `multiline`, `full_width`, `is_code`, `height`, `width`, `readonly` | Text box shape (in em units for `height`/`width`); `is_code` sets `normal_font()` (not a monospace font, despite the `Config.hpp` comment) and adds the "Edit Custom G-code" button when the group has `edit_custom_gcode`; disabled control (`readonly`). |
| `ratio_over` | For `coFloatOrPercent`: the key a percentage refers to. |
| `aliases`, `shortcut` | Legacy names; one value expanding to several keys. |
| `plugin_type` | Makes the option plugin-backed (`is_plugin_backed()`). |
An unadorned `coBool` def becomes a checkbox with no GUI code at all.
Pitfalls:
- **Rule:** Give `coFloatOrPercent` defs a `sidetext` of the form `"mm or %"` / `"mm/s or %"` and a sensible
`max_literal`.
**Why:**`Field::get_value_by_opt_type` asks "Is it N% or N mm?" by matching the English `sidetext`: a unitless
value above `max` when it contains `"mm/s"`, above `max_literal` when it contains `"mm "` (only that form also
clamps literal values to `max_literal`); another unit text skips the check.
Cite: `Field::get_value_by_opt_type`.
- **Rule:** Give every key that can be overridden per object a non-empty `category`.
| `DropDown` | native combo popup, menu used as a list | `PopupWindow` | `DropDown(parent, std::vector<Item>& items, style = 0)` | holds `items`**by reference**; `Invalidate()` after edits. §ComboBox |
Field classes (`Field.cpp`) build their editors from these widgets (`TextCtrl` → `::TextInput`, `CheckBox` →
`::CheckBox`, `SpinCtrl` → `SpinInput`, `Choice` → `::ComboBox`); the field machinery is in
`references/orca-settings-ui.md`.
## Event semantics
| Widget | Emits on user action | Id / event object | Programmatic setters | Bind |
|---|---|---|---|---|
| `Button` | `wxEVT_BUTTON` (on release inside, Space/Enter) | button id / button | `SetValue(bool)` (Checked look) silent | on the button (Rule 6) |
| `::CheckBox`, `SwitchButton` | `wxEVT_TOGGLEBUTTON` | native id / widget | `SetValue`, `SetHalfChecked` silent (`interface/wx/tglbtn.h:98`) | on the widget; never `wxEVT_CHECKBOX` |
| `::ComboBox` | `wxEVT_COMBOBOX` (int = index, string = text) from popup pick, arrow keys (read-only combo), `SelectAndNotify`; `wxEVT_COMBOBOX_DROPDOWN`/`_CLOSEUP` | COMBOBOX: combo's auto id / combo; DROPDOWN/CLOSEUP: **id 0, no object** | `SetSelection`, `SetValue` silent for COMBOBOX; editable or `CB_NO_TEXT`: they send `wxEVT_TEXT` | on the combo |
| `::TextInput` | inner `wxEVT_TEXT` (propagates); `wxEVT_TEXT_ENTER`, `wxEVT_KILL_FOCUS` re-sent to the wrapper only | TEXT: **inner** id/object; ENTER/KILL_FOCUS: wrapper id, inner object | `GetTextCtrl()->SetValue` sends `wxEVT_TEXT`; `ChangeValue` silent | TEXT: wrapper, inner, or ancestor by type; ENTER/KILL_FOCUS: wrapper or inner |
| `SpinInput` | `wxEVT_SPINCTRL` (a plain `wxCommandEvent`, no int) on commit; `EVT_SPINCTRL_TEXT` (int + string) per parseable keystroke; inner `wxEVT_TEXT` | spinner id / spinner | `SetValue` sends `EVT_SPINCTRL_TEXT` + `wxEVT_TEXT`, no `wxEVT_SPINCTRL` | on the spinner; handler takes `wxCommandEvent&` |
| `RadioGroup` | `wxEVT_RADIOBOX` (int, string) | group id / **no object** | **every**`SetSelection(i)` emits, same index included | on the group |
| `TabCtrl` | `wxEVT_TAB_SEL_CHANGING` (int = old index, not vetoable), then `wxEVT_TAB_SEL_CHANGED` (int = new) | ctrl id / ctrl | `SelectItem(i)` emits both when `i` changes | on the ctrl |
| `Notebook` | tab click posts `wxCUSTOMEVT_NOTEBOOK_SEL_CHANGED` (id = page index), then `wxEVT_NOTEBOOK_PAGE_CHANGING`/`_CHANGED` | — | `SetSelection` emits PAGE_*; `ChangeSelection` silent | on the notebook |
| `MultiSwitchButton` | `wxCUSTOMEVT_MULTISWITCH_SELECTION` (int, string) | id / widget | `SetSelection` emits when the index changes | on the widget |
| `SwitchBoard` | `wxCUSTOMEVT_SWITCH_POS` (int: 1 = left half, 0 = right half), **posted** | id 0 / none | — | on the board |
| `DropDown` (standalone) | `wxEVT_COMBOBOX` (int, string), `EVT_DISMISS` | COMBOBOX: popup id / popup; EVT_DISMISS: id 0, no object | — | on the DropDown (popups block propagation on MSW/macOS, not on GTK) |
| `TempInput` | `wxCUSTOMEVT_SET_TEMP_FINISH` on commit, **posted to the parent** (int = the ctor's `type`, string = the `TempInputType` number) | id 0 / none | — | on the **parent**; tell inputs apart by the int |
Consequences that recur:
- **Ids.**`ComboBox` calls `TextInput::Create`, which takes no id and creates the window with `wxID_ANY`, so the
combo's `id` argument is ignored; `TextInput` and
`SpinInput` have no id parameter. `parent->Bind(wxEVT_COMBOBOX, h, ID_MY_COMBO)` never fires. Bind on the widget,
or call `SetId()` after construction if an id filter is unavoidable.
- **Handler order.** Dynamic handlers run most-recently-bound first, before static event tables
(`interface/wx/event.h:592-593`). Widgets bind their internal handlers in the constructor (or use event tables),
so yours run first: a handler that does not `Skip()` cuts off the widget's own behaviour (bitmap refresh, commit,
click). A handler taking the event **by value** cannot `Skip()` the real event (`references/events.md §Skip
discipline`).
- **Setters vs wx.** wx's rule is that setters are silent (`interface/wx/ctrlsub.h:102-103`); `wxTextEntry::SetValue`
and `wxBookCtrlBase::SetSelection` are the documented exceptions (`references/controls-dataview.md §Events from
| `Parameter` | size and min 120×26, radius 4 | `Body_14` | buttons next to parameter boxes |
| `Icon` | padding 5×5, size and min 26×26, radius 4 | — | icon-only; create with `iconSize = 16` and a 16 px icon |
| `Expanded` | min height 32, padding 12×8, radius 4 | `Body_14` | full-width buttons, e.g. inside a static box |
All values go through `FromDIP`. `ButtonStyle::{Regular, Confirm, Alert, Disabled}` picks the background, border and
text `StateColor`s from the `btn_regular/btn_confirm/btn_alert/btn_disabled` tables in `Button.cpp` (palette colours
where the dark map should apply; the focus-border colour is chosen per theme when `SetStyle` runs). `ButtonStyle::Disabled` is a **look**; it does not call `Enable(false)`.
```cpp
auto* export_btn = new Button(this, _L("Export"));
`references/windows-dialogs.md`). Yes/No/Confirm/custom buttons never close on their own.
- **Rule:** Every custom label must be in the translation catalog: write it as `L("Skip for Now")` in the vector, or
make sure the same string appears in a `_L()` elsewhere.
**Why:** the constructor's `_L(label)` translates a variable, which xgettext cannot see, so a label used only there
never reaches `OrcaSlicer.pot` and always shows in English. `L()` is a no-op marker that xgettext extracts
(`references/strings-i18n-files.md`).
```cpp
new DialogButtons(this, {"Download and Install", "Skip for Now"}); // Wrong: untranslatable
new DialogButtons(this, {L("Download and Install"), L("Skip for Now")}); // Right
```
- **Rule:** Don't call `UpdateButtons()` (or `SetLeftAlignedButtonsCount`) repeatedly in your code.
**Why:** each call re-binds `wxEVT_KEY_DOWN` on every button (wx `Bind` does not de-duplicate) and the handler
`Skip()`s, so arrow-key focus moves repeat once per accumulated binding; DPI changes already add one each.
- **Rule:** Don't expect the row to take the dialog's background.
**Why:** the panel paints `darkModeColorFor("#FFFFFF")`, re-applied on every `UpdateButtons()` (DPI change); on a
non-white dialog the row stands out.
## Label and HyperLink
`Label` (`Widgets/Label.cpp`) is a `wxStaticText` with font `Body_14` (or the font passed), foreground `#262E30` and
the parent's background. Style bits (no clash with `wxST_*`):
| Bit | Value | Effect |
|---|---|---|
| `LB_HYPERLINK` | `0x20` | underlined font, `#009688`, hand cursor — **only** when set through `SetWindowStyleFlag` |
| `LB_PROPAGATE_MOUSE_EVENT` | `0x40` | left-down/up are forwarded to the parent's handler with `ProcessEventLocally` (label-relative position, label as event object) and consumed |
| `LB_AUTO_WRAP` | `0x80` | `Label::Wrap(GetSize().x)` on every `wxEVT_SIZE` |
`Label::Wrap(width)` is Orca's own wrapper over the stored text and has no width cache, unlike 3.3.2's
- `GetDropDown()` exposes the popup (`SetUseContentWidth(true[, limit])`, `SetAlignIcon`, colours).
`SetKeepDropArrow(true)` keeps the arrow and shows the item icon as a second icon. `ForceDropdownOpen()` opens it
programmatically (data-view editors use this).
- Opening: click (debounced by `DropDown::HasDismissLongTime()`, ≥ 20 ms since the last dismissal, so the click that
dismissed the popup does not reopen it), Enter/Space. Up/Down/Left/Right on a focused read-only combo step the
selection and send `wxEVT_COMBOBOX` (in an editable combo the keys go to the inner text control). Mouse-wheel selection is disabled (handler commented out of the event table). Scrolling an
ancestor `wxScrollHelper` hides the popup.
Events: see §Event semantics. Picking an item sends `wxEVT_COMBOBOX` even when it is already selected. Typing into
an editable combo sends `wxEVT_TEXT` per keystroke and **no**`wxEVT_COMBOBOX`: the commit is `wxEVT_TEXT_ENTER`/
`wxEVT_KILL_FOCUS` on the combo, after `OnEdit()` has matched the text to an item (exact, case-sensitive) and re-set
it (one more `wxEVT_TEXT`).
Pitfalls:
- **Rule:** Use `SelectAndNotify(n)` when listeners must react; `SetSelection(n)` is silent for `wxEVT_COMBOBOX` and
returns early when `n` is already selected.
**Why:** matches wx (`interface/wx/ctrlsub.h:102-103`), but in an **editable** combo (or one with `CB_NO_TEXT`)
`SetSelection`/`SetValue` go through `GetTextCtrl()->SetValue()` and send a propagating `wxEVT_TEXT`; guard
`wxEVT_TEXT` handlers during sync.
- **Rule:** Store only untyped `void*` client data and free it yourself.
**Why:** every `Append` overload calls `SetClientDataType(wxClientData_Void)`, and declaring them hides every base
`wxItemContainer::Append` (the `wxClientData*` and `wxArrayString` ones included); `DeleteOneItem()` bypasses the
base client-object reset. Typed `wxClientData` ownership (`references/controls-dataview.md §Item containers`) does
not apply.
- **Rule:** After `Clear()`, `Insert()`, `Set()` or `Delete()`, call `SetSelection`/`SetValue`.
**Why:** they reset the selection to -1 (`DropDown::Invalidate(true)`) but leave the shown text.
- **Rule:** Bind `wxEVT_COMBOBOX_DROPDOWN`/`_CLOSEUP` on the combo without an id filter.
**Why:** they are created with id 0 and no event object; any ancestor bound without a filter receives them from
every combo below it.
`DropDown` (`Widgets/DropDown.cpp`) standalone: a `PopupWindow` (`wxPU_CONTAINS_CONTROLS`, `wxBG_STYLE_PAINT`,
`wxBufferedPaintDC`) drawing `items`, which it holds **by reference** — the vector must outlive the popup, and
`Invalidate()` must follow any edit of it. `Popup()` on GTK sets the toplevel as transient parent explicitly (a
data-view editor can get focus before wxGTK infers one); `Dismiss()` refuses while its sub-dropdown is shown;
`OnDismiss()` sends `EVT_DISMISS` and stamps the dismissal time; `ShouldDismissOnTopWindowDeactivate()` keeps chained
dropdowns open on Wayland, where mapping a grabbing child popup deactivates the toplevel; on macOS it binds an empty
`wxEVT_IDLE` handler to stop wx's idle-time capture release/re-capture ([source] `src/common/popupcmn.cpp:108-111,
439-472`). It captures without a `HasCapture()` guard, and its capture-lost handler replays release, which can commit
the hovered item. Popup mechanics: `references/popups-menus.md`.
## SpinInput and TempInput
`SpinInput` (`Widgets/SpinInput.cpp`) is a `wxNavigationEnabled<StaticBox>` with an inner `TextCtrl` validated by
`wxTextValidator(wxFILTER_DIGITS)`, two arrow `Button`s (not keyboard-focusable) and a repeat `wxTimer`.
- `text` (if it parses) overrides `initial`; `label` is a painted side label; `style` goes to the inner control
(`wxTE_PROCESS_ENTER` added).
- `SetValue(int|wxString)` clamps to `[min, max]` and stores; `GetValue()` returns the **last committed** value, not
the text being typed; `SetRange(min, max)` stores the bounds without re-clamping; `SetStep`.
- Commits — parse, clamp, `wxEVT_SPINCTRL` — on Enter, on kill-focus and on Up/Down keys (which do not step past a
bound) only if the value changed, but on **every** arrow-button press and auto-repeat tick (while held), even when
the value is pinned at `min`/`max` (`SpinInput::createButton`, `SpinInput::onTimer`). Mouse-wheel stepping is
disabled (handler commented out of the event table).
- `EVT_SPINCTRL_TEXT` (int + string) fires on every keystroke that leaves a parseable integer.
- Enter and kill-focus are re-dispatched to the spinner with `ProcessEventLocally`, as in TextInput. Unlike
TextInput, `SpinInput::onTextLostFocus` never `Skip()`s the inner control's `wxEVT_KILL_FOCUS` (it skips a local
copy), although focus handlers should (`interface/wx/event.h:3410-3412`); a `KILL_FOCUS` handler bound on
`GetTextCtrl()` runs before it and must `Skip()` to keep the commit.
Pitfalls:
- **Rule:** Bind `wxEVT_SPINCTRL` with a `wxCommandEvent&` handler and read `GetValue()` from the spinner.
**Why:**`wxEVT_SPINCTRL` is declared with `wxSpinEvent` (`include/wx/spinctrl.h:22`) but `SpinInput::sendSpinEvent`
sends a plain `wxCommandEvent` without `SetInt`; a `wxSpinEvent&` handler reads `GetPosition()` == 0 from an object
| What | a bare tab **bar**: one `Button` per tab in a `StaticBox` | `wxBookCtrlBase` with a `ButtonsListCtrl` header (one `Button` per tab) | standard stacked-page container, no UI |
- **Rule:** To change the label, recreate the box.
**Why:** the label is captured and measured in `Create()`; there is no `SetLabel` override, so `SetLabel` changes the
native text the widget never paints.
- **Rule:** Don't expect a disabled look.
**Why:** its `StateColor`s list `Normal` before `Disabled`, and the first match wins
(`references/colours-dark-mode.md §StateColor`), so the disabled colours never apply.
`StaticLine` (`Widgets/StaticLine.cpp`): horizontal or vertical separator with optional label and icon;
`SetLineColour(wxColour)`, `SetLabel`, `SetIcon`, `Rescale()`; line and text colours go through `darkModeColorFor` at
paint time. Use it instead of `wxStaticLine`.
## ProgressBar
`ProgressBar` (`Widgets/ProgressBar.cpp`) is a `wxWindow` with a rounded track and fill (MSW draws through a memory
DC + `wxGCDC` for anti-aliasing). `SetValue(step)` (re-enables after `Disable(text)`), `SetProgress(step)` (ignores
negatives), `Reset()`, `ShowNumber(bool)`, `SetRadius`, `SetHeight(h)` (sets min height and radius `h/2`), and
`Disable(wxString text)`, which draws an orange "disabled" bar with `text` and hides `wxWindow::Disable()` (name
hiding: `bar->Disable()` does not compile).
Pitfalls:
- **Rule:** Pass dark-mapped colours and re-set them on theme change.
**Why:** track, fill and text are plain `wxColour`s painted as given; nothing maps them.
```cpp
bar->SetProgressBackgroundColour(StateColor::darkModeColorFor(wxColour("#009688"))); // sets the FILL
```
- **Rule:** Mind the swapped setter names.
**Why:**`SetProgressForedColour` sets the **track** (`m_progress_background_colour`) and
`SetProgressBackgroundColour` sets the **fill** (`m_progress_colour`).
- **Rule:** Use `SetHeight(FromDIP(h))` for a bar thinner than 14 px.
**Why:**`ProgressBar::SetMinSize` returns without doing anything (width included) when the height is below its
`miniHeight` of 14 — so `SetMinSize(wxSize(w, -1))` is ignored too.
- **Rule:** Use `ShowNumber` only with `max == 100`.
**Why:** it draws the raw step followed by `%`.
- `Rescale()` is empty: re-set the height in `on_dpi_changed`.
## ScrolledWindow
`ScrolledWindow` (`Widgets/ScrolledWindow.cpp`) is a `wxScrolled<wxWindow>` that hides the native scrollbars and
draws `MyScrollbar`s (`Widgets/Scrollbar.cpp`) of `scrollbarWidth` in a `marginWidth` strip; content goes on
`GetPanel()` (the scroll target). Its internal panels are sized from the constructor `size`, so pass a real size
(`FromDIP`/`em`-based, as `Search.cpp` does). Use raw `wxScrolledWindow` for ordinary scrolling.
- **Rule:** Create it with `wxVSCROLL`.
**Why:** the mouse-wheel handler forwards to the vertical bar unconditionally and `SetBackgroundColour` touches the
vertical-bar panels unconditionally, so a `wxHSCROLL`-only instance dereferences null pointers; with neither flag
`GetPanel()` is null. With both flags only the vertical bar is built.
- `MyScrollbar` paints on a `wxClientDC` and forces `Refresh(); Update();` — do not copy its painting
(`references/painting-custom-widgets.md`).
## PopupWindow
`PopupWindow` (`Widgets/PopupWindow.cpp`) derives `wxPopupTransientWindow` and adds per-platform dismissal hooks.
Use it for every transient popup, knowing which parts are automatic:
| Port | Added behaviour |
|---|---|
| GTK (X11 and Wayland) | `Create` binds `wxEVT_ACTIVATE` on the first top-level window strictly above the parent (`GetTopParent` starts at the parent's parent, so a popup parented directly to a dialog watches the dialog's own parent); deactivation → `DismissAndNotify()` unless `ShouldDismissOnTopWindowDeactivate()` (virtual) returns false — `DropDown` uses that to keep chained popups open on Wayland |
| MSW | dismissal on toplevel deactivate/iconize/hide is **opt-in**: call `BindUnfocusEvent()` (MSW-only member); the destructor unbinds. Its activate handler does not `Skip()`, so while the popup exists the toplevel's earlier-bound `wxEVT_ACTIVATE` handlers and wx's focus save/restore (`wxTopLevelWindowMSW::OnActivate`, `src/msw/toplevel.cpp:1326`) do not run [source] |
| macOS | with `wxPU_CONTAINS_CONTROLS` it hit-tests and forwards mouse and enter/leave events to child controls (`OnMouseEvent2`); wx documents the flag as MSW focus behaviour only (`interface/wx/popupwin.h:17-26`) |
Command events from controls inside the popup stop at the popup on MSW and macOS ([source]
`src/common/popupcmn.cpp:135`) but bubble on to the popup's parent on GTK, whose `wxPopupWindow::Create` never sets
`wxWS_EX_BLOCK_EVENTS` (`src/gtk/popupwin.cpp`; `references/events.md §5`); bind them on the popup or its children,
as `ComboBox` binds `drop`'s `wxEVT_COMBOBOX`, and don't let an ancestor's unfiltered handler also act on them. When a click on the opener both
dismisses and reopens the popup, gate the reopen on `DropDown::HasDismissLongTime()` or an equivalent timestamp.
Dismissal contracts, per-port mechanics and parenting rules: `references/popups-menus.md`.
## ProgressDialog, WebView and device-page composites
- **`Slic3r::GUI::ProgressDialog`** (`Widgets/ProgressDialog.cpp`): a `wxDialog` copy of wx's generic progress dialog
with Orca styling and a `Button` for Cancel; wx-compatible API (`Update(value, msg, &skip)`, `Pulse`,
`WasCancelled`, `WasSkipped`, `Resume`, `SetRange`). `Update` yields to the event loop
(`YieldFor(wxEVT_CATEGORY_UI | wxEVT_CATEGORY_USER_INPUT)`), so handlers can re-enter; usage rules and the Jobs
| Linux | **wxGTK3** | `deps/CMakeLists.txt` declares `option(DEP_WX_GTK3 "Build wxWidgets against GTK3" ON)` (default ON since 026499c5b7, #10294), which gives `-DwxBUILD_TOOLKIT=gtk3`. The root `CMakeLists.txt` sets `SLIC3R_GTK "3"`, so `src/CMakeLists.txt` finds wx through `wx-config --toolkit=gtk${SLIC3R_GTK}` and `src/slic3r/CMakeLists.txt` links `GTK${SLIC3R_GTK}`. Flatpak uses gtk3 too. |
**GTK2 is an opt-out, not a default.** `wxWidgets.cmake` starts from `_gtk_ver 2` and switches to 3
when `DEP_WX_GTK3` is on, so you get GTK2 only by passing `-DDEP_WX_GTK3=OFF` (and `SLIC3R_GTK=2`).
A GTK2 build loses the following (wx `build/cmake/init.cmake`, `include/wx/features.h`):
- EGL: `wxUSE_GLCANVAS_EGL` is forced OFF unless GTK3 and EGL are both found (init.cmake:559-561), and
`wxHAS_EGL` is set only on GTK3 (:531-540). No EGL means no native Wayland GL.
- WebKit2: GTK2 gets WebKit1 (init.cmake:568-569); GTK3 uses webkit2gtk-4.1 and falls back to 4.0
(:571-577).
- DIP pixels: `wxHAS_DPI_INDEPENDENT_PIXELS` is defined only for `__WXGTK3__ || __WXMAC__ || __WXQT__`
(`include/wx/features.h:115-120`), so GTK2 uses physical pixels and gets no `wxEVT_DPI_CHANGED`.
- Backend detection: Orca's `wxHAVE_GDK_*` macros are not defined, so `get_linux_display_backend()`
| `wxUSE_NANOSVG` | `OFF` | no SVG in wx; see [No SVG](#no-svg-in-orcas-wx). It was disabled to avoid duplicate symbols with Orca's own NanoSVG, which carries an `nsvgRasterizeXY` extension (7658cf9076) |
| `wxUSE_GLCANVAS_EGL` | `ON` | takes effect only on GTK3 with EGL found; with both EGL and GLX built, wx picks EGL even on X11 unless `PreferGLX()` is called → `references/webview-gl-aui-media.md` §EGL vs GLX |
| `wxUSE_OPENGL` | `ON` | `wxGLCanvas` |
| `wxUSE_WEBVIEW` / `wxUSE_WEBVIEW_EDGE` / `wxUSE_WEBVIEW_IE` | `ON` / `ON` only `if (MSVC)` / `OFF` | Edge (WebView2) on Windows, WKWebView on macOS, WebKit2GTK on Linux. `wxUSE_WEBVIEW_CHROMIUM` keeps its default `OFF` (`build/cmake/options.cmake:303`), so Chromium-backend notes never apply. `wxUSE_WEBVIEW_EDGE_STATIC` keeps its default `OFF` (`build/cmake/options.cmake:514`), so the root `CMakeLists.txt` ships `WebView2Loader.dll` from `deps/WebView2/lib/win-<arch>` next to the executable → `references/webview-gl-aui-media.md` §wxWebView backends |
| `wxUSE_WEBREQUEST` | `ON` | `wxWebSession`/`wxWebRequest` are available (used only for a few image downloads) |
| `wxUSE_MEDIACTRL` | `ON` | kept for `wxMediaState`; the camera view is Orca's `wxMediaCtrl3`, not a `wxMediaCtrl` |
| `wxUSE_PRIVATE_FONTS` | `ON` | `wxFont::AddPrivateFont` for the bundled fonts → `references/dpi-bitmaps-fonts.md` |
| `wxUSE_AUI` | `ON` | Plater docking and `BBLTopbar` (a `wxAuiToolBar`) |
| `wxUSE_STC` | `OFF` | no `wxStyledTextCtrl` |
| `wxUSE_DETECT_SM` | `OFF` | no X11 session-manager detection |
| `__WXGTK20__` | GTK ≥ 2.0, **also defined in GTK3 builds** (`build/cmake/setup.cmake:61-72` defines every version macro up to the toolkit version) | `defined(__WXGTK20__) \|\| defined(__WXGTK3__)` in `GUI_App.cpp` is the same as `__WXGTK__` |
| `__WINDOWS__` | defined by wx when `_WIN32`, `__WIN32__` or `__WXMSW__` is defined (`include/wx/platform.h:87-91`) | wx's own Windows checks; Orca's dark-mode code uses it |
| `_WIN32`, `__APPLE__`, `__linux__` | compiler | OS APIs (Win32, Cocoa frameworks, `/proc`), and all of `src/libslic3r` |
| `wxHAS_EGL`, `wxHAS_GLX` | wx `setup.h` (`build/cmake/setup.h.in:1144-1147`) when built with EGL / GLX | GL backend code; test them only after a wx header is included |
| `wxHAVE_GDK_WAYLAND`, `wxHAVE_GDK_X11` | Orca's `cmake/modules/FindGTK3.cmake` (`check_symbol_exists(GDK_WINDOWING_WAYLAND/X11 "gdk/gdk.h" …)`), passed by `src/slic3r/CMakeLists.txt` as PRIVATE definitions of `libslic3r_gui` | only `LinuxDisplayBackend.cpp` needs them; wx headers do not define them |
| `GTK_CHECK_VERSION(a,b,c)` | GTK headers, compile time | branching on GTK API version. `gtk_check_version()` and wx's internal `wx_is_at_least_gtk3(n)` check the runtime version |
The toolkit macros come from wx's compile definitions: `wxTOOLKIT_DEFINITIONS` in
`build/cmake/toolkit.cmake:57-87,157`, exported through the CMake targets and through
`wx-config --cxxflags`. They are therefore defined in every `libslic3r_gui` source, even before the
first wx include; `LinuxDisplayBackend.hpp` relies on this. `src/libslic3r` does not link wx and uses
none of them. Prefer the toolkit macro whenever the difference comes from wx: a later toolkit change
(for example wxGTK on another OS) then keeps the right branch.
- **Rule:** Never call the detectors from static initialisers or before `wxEntry` has initialised GTK.
**Why:**`gdk_display_get_default()` is null then, and the function-local cache keeps `Unknown`
forever, which silently disables every Wayland or X11 branch.
## Wayland gaps
wx has no dedicated Wayland document; the documented limits are scattered across the interface
headers. Rows marked [source] or "protocol" come from the implementation or from Wayland itself.
| Area | What happens on Wayland | Cite | Orca handling / owner |
|---|---|---|---|
| Global pointer position | `wxGetMousePosition()` / `wxGetMouseState()` call `gdk_device_get_position` [source], but Wayland gives clients no global pointer position, so the result is not a screen position | `src/gtk/window.cpp``wxGetMousePosition` | use event coordinates + `ClientToScreen` (`BBLTopbar`); dismiss popups by focus tracking (`SearchDialog::Dismiss`) → `references/mouse-keyboard-focus.md` |
| Top-level position | the compositor places windows, and `SetPosition()` / `Move()` on a TLW have no effect (protocol) | — | `GUI_App::window_pos_restore` restores only size and maximised state; moves go through `gtk_window_begin_move_drag` |
| Window icon | `SetIcon()` / `SetIcons()` "doesn't do anything when using Wayland … create a `.desktop` file" | `interface/wx/toplevel.h:517-521, 538-542` | `src/dev-utils/platform/unix/com.orcaslicer.OrcaSlicer.desktop` (`Icon=OrcaSlicer`, `StartupWMClass=orca-slicer`) |
| App id | `wxAppConsole::SetClassName()` is the xdg `app_id` with GTK ≥ 3.24.22 (and the AUMID on Windows); it must be set before any TLW. wx applies it when a TLW is mapped, and only if it is non-empty [source: `wxTopLevelWindowGTK::GTKHandleMapped`] | `interface/wx/app.h:765-812` | Orca calls only `SetAppName(SLIC3R_APP_KEY)`, so GTK's default applies. On Windows, `SetClassName` would also change shell behaviour (MRU, Shift+middle-click) |
| `wxClientDC` | deprecated in 3.3 ("please use wxInfoDC instead for obtaining information", `interface/wx/dcclient.h:43-46`). Drawing through it "simply doesn't have any effect" on GTK3/Wayland or wxOSX. `CanBeUsedForDrawing()` [source]: false on Wayland only for wxGTK (`src/gtk/dc.cpp`), always false on wxOSX, always true on wxMSW (`include/wx/{osx,msw}/dcclient.h`), although its doc also lists "wxMSW when using double buffering" | `interface/wx/dcclient.h:48-53, 80-88` | draw only in `wxPaintDC` after `Refresh()`/`RefreshRect()` → `references/painting-custom-widgets.md` |
| `wxWindow::Update()` | "doesn't do anything in wxGTK port when using Wayland". [source] wx skips the GDK update calls there because they broke later updates (#25036) | `interface/wx/window.h:2405-2407` | never rely on `Update()` to paint synchronously |
| `WarpPointer()` | works only if the compositor implements the pointer-warp protocol; mutter also needs a mouse button held | `interface/wx/window.h:3902-3907`; `docs/changes.txt:294` | Orca never warps the pointer |
| `wxUIActionSimulator` | "currently doesn't work when using Wayland with wxGTK" | `interface/wx/uiaction.h:20` | not used; Orca also builds `wxUSE_XTEST=OFF` |
| OpenGL | only EGL; `PreferGLX()` has no effect. Without EGL in the build, wxGTK's `wxGLCanvas` shows a fatal message and refuses to work [source: `src/gtk/glcanvas.cpp``IsAvailable`]. The EGL surface is a subsurface over the canvas, ready only after map and a frame callback [source: `src/unix/glegl.cpp`] | `interface/wx/glcanvas.h:1094-1095` | `CLI::run` forces X11 when wx lacks EGL; overlays are drawn in GL/ImGui, never as wx children over the canvas → `references/webview-gl-aui-media.md` |
| AUI | the doc note "live resize is always used … for wxOSX and wxGTK3 when using Wayland" is obsolete: "As of wxWidgets 3.3.0 this function always returns false", and `wxAUI_MGR_LIVE_RESIZE` is in the default flags. Floating panes need global positions | `interface/wx/aui/framemanager.h:336-345` | `Plater::priv::priv` clears `wxAUI_MGR_ALLOW_FLOATING`; `sanitize_window_layout_for_wayland` strips floating state from the saved layout → `references/webview-gl-aui-media.md` §wxAuiManager |
| Popups | a GTK popup is an `xdg_popup` only for COMBO/DROPDOWN/POPUP_MENU hints, so wx creates popups with `GDK_WINDOW_TYPE_HINT_COMBO` [source]. A chained popup's parent must be the mapped popup, and mapping it with a grab deactivates the toplevel (Orca's comments in `DropDown::mouseMove`, `DropDown::ShouldDismissOnTopWindowDeactivate`) | `src/gtk/popupwin.cpp:110-114` | `DropDown` transient-for chain, `ShouldDismissOnTopWindowDeactivate` → `references/popups-menus.md` |
| Fractional scale | arrives as an integer GDK scale | `docs/doxygen/overviews/high_dpi.md:348-351` | → `references/dpi-bitmaps-fonts.md` |
| Window decorations | some desktop environments draw a title bar on undecorated windows anyway | — | [Undecorated windows](#wayland-undecorated-top-level-windows-splash) |
| `wxWebViewChromium` | X11 only | `interface/wx/webview_chromium.h:133-146` | not built (`wxUSE_WEBVIEW_CHROMIUM` OFF) |
Wayland history in the change logs (`docs/changes_32.txt`): already in 3.1.5, Orca's previous wx,
were two-finger scrolling (703, 3.1.3), the EGL-based `wxGLCanvas` for Wayland (497) and `wxMediaCtrl`
support (498, both 3.1.5). New with the upgrade: "Many bug fixes for Wayland-specific problem" (397)
and a `wxMediaCtrl` fix (402) in 3.1.6, GDK errors from `PopupMenu()` avoided (317, 3.1.7),
`wxCURSOR_SIZING` fixed (261, 3.2.0). In 3.3, `WarpPointer()` on supported compositors
(`docs/changes.txt:294`) and the EGL/Wayland high-DPI scale fix (`changes.txt:538`).
## Window decoration and custom title bars
Only `MainFrame` replaces the native title bar. Dialogs keep native decorations, chosen through their
style flags (`references/windows-dialogs.md`). `MainFrame` is created with
- *Best size* is derived from content. *Min size* is "normally explicitly set by the programmer"; most
controls also take it from a non-default ctor size. *Initial size* is the ctor size; a partly specified
size such as `wxSize(150, -1)` is completed from the best size. *Virtual size* is the scrollable extent.
- `GetEffectiveMinSize()` merges the best size into the min size: "This is the value used by sizers to
determine the appropriate amount of space to allocate for the widget" (`interface/wx/window.h:1393-1402`). It is the min size with unspecified components filled
from the best size (`wincmn.cpp:868`). Once `SetMinSize(wxSize(w, h))` sets both components, the
content no longer affects the sizer allocation; pass `-1` for the component that must follow content.
- "The best size respects the minimal and maximal size explicitly set for the window" (`window.h:1342-1350`;
`wincmn.cpp:879`: raised to min, lowered to max). It is **cached only when the window has no sizer**
**[source]** (`wincmn.cpp:881`). Containers with sizers recompute every time; leaf controls and custom
widgets return the cache until `InvalidateBestSize()` (`window.h:1645-1651`; the cache is documented for
`DoGetBestClientSize`, `window.h:4432-4435`).
- `InvalidateBestSize()` also invalidates the parent chain, stopping at a top-level window **[source]**
(`wincmn.cpp:636`). It does not lay anything out (see [Re-layout](#re-layout-after-content-or-visibility-changes)).
- The default `DoGetBestSize()` uses the sizer's min size; without a sizer, the bounding box of visible
children; with no children, the min size or (1,1) (`wincmn.cpp:649`).
- `SetInitialSize(size)` sets the min size to `size`, merges it with the best size and resizes
(`window.h:1742-1757`, `wincmn.cpp:937`). A ctor size therefore becomes the min size: a fixed height
pins the minimum height.
- `SetMinSize` "doesn't prevent the program from making the window explicitly smaller … by calling
SetSize(), it just ensures that it won't become smaller than this size during the automatic layout"
(`window.h:1808-1815`). Top-level windows are the documented exception: their size hints also stop the
program's own `SetSize()` (`interface/wx/toplevel.h:576-603`). **[source]** On GTK, top-level windows and
`wxPopupWindow` clamp `SetSize` to min/max (`src/gtk/toplevel.cpp:1365``ConstrainSize`,
`src/gtk/popupwin.cpp:171`).
- On a top-level window, `SetMinSize`/`SetMaxSize` go through `SetSizeHints(min, max)`
(`src/common/toplvcmn.cpp:197-205`), so a min larger than the current max (or the reverse) is dropped
silently.
- `wxWindow::SetSizeHints` on a non-top-level window "is discouraged. Please use SetMinSize() and
- keeps the original text in `m_text`; `Label::Wrap(int)`**hides** the non-virtual `wxStaticText::Wrap`,
wraps `m_text` with Orca's `wxTextWrapper2` (breaks between ideographs above U+4E00 and hard-breaks
space-less runs) and sets the result through the base `SetLabel` with `m_skip_size_evt` set; it has no
width cache;
- `LB_AUTO_WRAP` binds `wxEVT_SIZE` (with `Skip()`) and re-wraps at the new width; `Label::SetLabel`
re-wraps at `GetSize().x` under `LB_AUTO_WRAP` and returns early when the text is unchanged;
- the label needs a real width: a ctor size `wxSize(FromDIP(w), -1)`, min = max width, or `wxEXPAND` in a
vertical sizer whose container width is fixed;
- `Label::split_lines(dc, width, text, out, max_count)` exposes the same wrapper for owner-drawn text.
Orca's wrapping is preferred over `wxST_WRAP` for CJK text.
**Pitfalls**
- **Rule:** Re-wrapping after `SetLabel` must defeat the wrap cache.
```cpp
st->SetLabel(msg); st->Wrap(w); // Wrong when w is unchanged: stays unwrapped
st->SetLabel(msg); st->Wrap(-1); st->Wrap(w); // Right
// or: Label with LB_AUTO_WRAP / label->Wrap(w) through a Label*
```
- **Rule:** Call `Wrap` through a `Label*`, never through a `wxStaticText*` that points at a `Label`.
**Why:** the base `Wrap` sets the wrapped text through the virtual `Label::SetLabel`, which overwrites
`m_text` with the wrapped copy (and re-wraps it under `LB_AUTO_WRAP`).
- **Rule:** Never compute or commit a size or wrap from a width that has not been laid out; guard with a
sanity check on the client width.
**Why:** a plain `wxWindow`/`wxPanel` child created with `wxDefaultSize` is wx's 20×20 placeholder on
every port until the first sizer layout **[source]** (`include/wx/window.h:1914-1915``WidthDefault`/
`HeightDefault`; native controls instead take their best size through `SetInitialSize` in `Create()`),
and a minimized top-level window reports a (0,0) client size. Wrapping against that width commits a garbage min size, which `SetSizeHints` then locks in.
```cpp
void UpdateMinSize() {
int cWidth = GetClientSize().GetWidth();
if (cWidth < 50) return; // not laid out yet: don't commit a size
/* wrap at cWidth, SetMinSize(wxSize(-1, h)), GetParent()->Layout() */
}
```
Cite: 5ede9711f5 (`CenteredMultiLinePanel::UpdateMinSize` and `::OnPaint` in `TroubleshootDialog.hpp`).
- **Rule:** Give a wrapping label a fixed width and `-1` height, then wrap.
**Why:** the ctor size becomes the min size (`SetInitialSize`), so a fixed height clips the text when a
translation is longer.
```cpp
new wxStaticText(this, wxID_ANY, txt, wxDefaultPosition, wxSize(FromDIP(490), FromDIP(40))); // Wrong
m_action_line = new wxStaticText(this, wxID_ANY, wxEmptyString, wxDefaultPosition,
- **Rule:** An ellipsized label needs a width limit.
**Why:** its best size is the full text, so the sizer gives it the full width and nothing ellipsizes.
## Layout on DPI change
**What wx does per platform.**
| Platform | When `DPIAware::on_dpi_changed` runs | What wx already rescaled |
|---|---|---|
| MSW (per-monitor v2 manifest) | `wxEVT_DPI_CHANGED`; also `wxEVT_MOVE_END` when the scale differs | **[source]** before the event, `MSWUpdateOnDPIChange` rescales every window's min/max size, the fonts of non-top-level windows (the TLW keeps its font), every sizer item's border, spacer and nested-sizer min sizes, and invalidates best sizes (`src/msw/window.cpp:5002-5085`; detail in `references/dpi-bitmaps-fonts.md` §wxEVT_DPI_CHANGED) |
| GTK3 | `wxEVT_DPI_CHANGED` when the integer GDK scale changes (GTK ≥ 3.10, `src/gtk/toplevel.cpp:336-350`) | nothing; logical pixels are DIPs, so DIP sizes stay valid |
| macOS | never: wx does send `wxEVT_DPI_CHANGED` on backing-scale changes **[source]** (`src/osx/cocoa/nonownedwnd.mm``windowDidChangeBackingProperties`), but `DPIAware` binds it only `#ifndef __WXOSX__` | nothing needed; Cocoa scales |
A top-level window's default `wxEVT_DPI_CHANGED` handler resizes it, and a handler that does not `Skip()`
suppresses that (`interface/wx/event.h:3572-3584`). `DPIAware`'s handler does not `Skip()`, so on MSW wx
skips its own resize of the top-level window (`src/msw/nonownedwnd.cpp:284-318``HandleDPIChange`): on MSW
the dialog must resize itself in `on_dpi_changed`. `DPIAware::rescale` (`src/slic3r/GUI/GUI_Utils.hpp`) runs
`Freeze(); update_em_unit(); on_dpi_changed(rect); Layout(); Thaw();`. The em machinery: see
| implicit `std::string` / `const char*` → `wxString` | **on**, decoded with the current locale (`wxNO_IMPLICIT_WXSTRING_ENCODING` is not defined) | `include/wx/string.h:1215-1217, 1323-1325` |
| wx's own `_()` macro | not defined (`-DWXINTL_NO_GETTEXT_MACRO`); Orca defines its own | `CMakeLists.txt`; `include/wx/translation.h:47-49`; `src/slic3r/GUI/I18N.hpp` |
| debug level | wx built with `wxBUILD_DEBUG_LEVEL=0`; `libslic3r_gui` gets `wxDEBUG_LEVEL=0` under `SLIC3R_STATIC` (default ON): `wxASSERT`, `Format` type checks, `wxLogDebug`, `wxLogTrace` compile out; `wxCHECK*` still return, silently | `deps/wxWidgets/wxWidgets.cmake`, `src/slic3r/CMakeLists.txt`; `include/wx/log.h:62-77` |
| MSVC source charset | `/utf-8`, so narrow literals hold UTF-8 bytes | `CMakeLists.txt` (`add_compile_options(... /utf-8)`) |
| clipboard, DnD, secret store, config, single-instance checker, IPC | compiled in (all `wxUSE_* 1`); Orca uses its own config and messaging, and the wx checker only on Windows | installed setup.h |
Because asserts are compiled out, every misuse described below fails silently in Orca: wrong or
empty text, a dropped call, a leaked object. wx would assert in a debug build; Orca never shows it.
## wxString encodings and conversions
**Contract.** A narrow `char*`, `std::string` or `std::string_view` given to wxString "supposes that
the string contains data in the current locale encoding, use FromUTF8() if the string contains
UTF-8-encoded data instead"; the implicit constructors are "dangerous … the resulting string will be
empty if the conversion from the current locale encoding fails" (`interface/wx/string.h:58-89`).
In the other direction `c_str()`/`mb_str()` are "potentially destructive … an empty string is
returned if the conversion fails", and `ToStdString()` loses data unless given `wxConvUTF8`; use
`utf8_string()` (`interface/wx/string.h:108-120, 843-863`). `FromUTF8`: "If s is not a valid UTF-8
string, an empty string is returned" (`interface/wx/string.h:2037`). The Unicode overview calls
`FromUTF8()` followed by `c_str()` "a recipe for disaster … may work perfectly well during testing on
Unix systems using UTF-8 locale but completely fail under Windows" (`docs/doxygen/overviews/unicode.h:314-320`).
| Platform | Narrow ↔ wide conversion | Consequence for UTF-8 data |
|---|---|---|
| MSW | `wxMBConv_win32` with `CP_ACP`, the system ANSI code page (`src/common/strconv.cpp:2600-2607`), independent of the UI language Orca picks; Orca's manifest does not opt into a UTF-8 active code page | non-ASCII text becomes mojibake (single-byte code pages) or an empty string (DBCS code pages). A developer machine with Windows' system-wide "Use Unicode UTF-8" option hides the bug |
| macOS | `wxMBConvLibc` (`mbstowcs`, follows `LC_CTYPE`) | the C runtime stays in the `"C"` locale until `GUI_App::load_language` calls `wxLocale::Init`; in `"C"` macOS maps each byte to one character (mojibake) [tested]; region locales afterwards decode UTF-8 [tested] |
| GTK | `wxMBConvLibc` | wxGTK calls `gtk_disable_setlocale()` (`src/gtk/app.cpp:526-537`), so the locale is `"C"` until `load_language`; `wxUILocale` then prefers a UTF-8 codeset (`src/unix/uilocale.cpp``TryCreateLocaleWithUTF8`, `wxSetlocaleTryUTF8`) |
The implicit conversions therefore work on macOS and Linux after startup and break on Windows, so a
bug of this class survives testing on a Mac.
**What compiles in Orca and what it does:**
| Expression | Compiles? | Encoding |
|---|---|---|
| `wxString w = s;` or `f(const wxString&)` called with `std::string`/`const char*` | yes | locale (CP_ACP on MSW): mojibake or empty for UTF-8 |
| `wxString w(sv)` from `std::string_view` | only explicitly: the ctor is `explicit` (`include/wx/string.h:1327-1328`) although `interface/wx/string.h:71` calls it implicit | locale |
| `std::string s = w;`, `const char* p = w;` | **no** | — |
| `from_u8(const std::string&)` | `wxString::FromUTF8(str.c_str())` | stops at an embedded NUL, `""` on invalid UTF-8; for binary-safe input use `wxString::FromUTF8(s)` (the `std::string` overload passes the length, `include/wx/string.h:1782-1783`) |
Cite: `GUI_App::import_model` (model opener with `from_u8` default dir); `CMakeLists.txt`
`wxNO_UNSAFE_WXSTRING_CONV` comment.
- **Rule:** get UTF-8 out of a `wxString` with `into_u8`/`utf8_string()`.
**Why:**`ToStdString()`, `mb_str()` and `c_str()` use `wxConvLibc` and return `""` when a character
is not representable in the ANSI code page (`include/wx/string.h:4292-4303`).
```cpp
// Wrong:
std::string path = dlg.GetPath().ToStdString();
// Right:
boost::filesystem::path path = into_path(dlg.GetPath()); // or into_u8(dlg.GetPath())
```
## Formatting
**`wxString::Format`/`Printf` contract.** Variadic templates that normalise each argument; positional
`%2$d %1$d` is supported because `wxUSE_PRINTF_POS_PARAMS` is 1 (`interface/wx/string.h:1533-1546`).
The specifier/argument type check is a `wxASSERT_MSG` (`include/wx/strvararg.h:336-347`), so at
debug level 0 nothing checks it: `%d` with `size_t`, `%s` with an `int` or a missing argument is
silent undefined behaviour. `const char*`, `std::string` and `std::string_view` arguments are
decoded with `wxConvLibc` (`wxArgNormalizerWchar<const char*>`, `include/wx/strvararg.h:612-624,
772-781`), so UTF-8 data garbles on MSW exactly as in the table above.
**OrcaSlicer: `format_wxstr` and `GUI::format`** (`src/slic3r/GUI/format.hpp`, on top of
`Slic3r::format` in `src/libslic3r/format.hpp`) wrap `boost::format`:
| Helper | Returns | Arguments |
|---|---|---|
| `format_wxstr(fmt, args...)` | `wxString`, built with `wxString::FromUTF8(result)` | `fmt` as `const char*`, `std::string` (UTF-8) or `wxString`; args of any streamable type |
| `GUI::format(fmt, args...)` | UTF-8 `std::string` | same |
dlg.GetPaths(paths); // GetPath() returns "" with wxFD_MULTIPLE
for (const wxString& p : paths)
load(into_path(p));
```
**Contract** (`interface/wx/filedlg.h`):
- `wxFD_OPEN` and `wxFD_SAVE` are exclusive; `wxFD_MULTIPLE`/`wxFD_FILE_MUST_EXIST` are open-only,
`wxFD_OVERWRITE_PROMPT` save-only (`:163-200`). Contradictory styles are only asserted
(`src/common/fldlgcmn.cpp:776-787`), i.e. ignored silently in Orca.
- Style bits are reused between classes: `wxPD_APP_MODAL` (0x0002) equals `wxFD_SAVE`
(`include/wx/progdlg.h:21`, `include/wx/filedlg.h:45-46`), so a foreign flag can change the dialog
type. Pass only `wxFD_*`.
- `GetPath()`/`GetFilename()` "can't be used with dialogs which have the `wxFD_MULTIPLE` style"
(`:343-347, 380-384`); they return `""` there (`docs/changes_32.txt:119-120`). Use
`GetPaths()`/`GetFilenames()`.
- Wildcard format `"Desc (*.a;*.b)|*.a;*.b|Desc2 (*.c)|*.c"` (`:106-115`); the default wildcard is
`"*.*"` on MSW and `"*"` elsewhere (`:31-36`).
- `SetFilename()` in wxGTK has "little effect unless a default directory has previously been set"
(`:452-456`).
- `SetExtraControlCreator()` forces old XP-style dialogs on MSW; `SetCustomizeHook()` is native
(`:131-155`). New-style MSW dialogs need a single-threaded COM apartment (`:157-161`).
- `wxDirDialog`: `wxDD_NEW_DIR_BUTTON` is `0` ("deprecated, on by default now",
`interface/wx/dirdlg.h:14`); the "Create new directory" button is shown exactly when
`wxDD_DIR_MUST_EXIST` is absent (`interface/wx/dirdlg.h:42-46`); on macOS 10.11+ there is no title
bar, the `message` argument is what the user sees (`interface/wx/dirdlg.h:59-62`).
**Platforms:**
| | Behaviour |
|---|---|
| MSW | New-style `IFileDialog`: wx calls `SetDefaultExtension` with the selected filter's first extension and no longer calls `AppendExtension` itself (`src/msw/filedlg.cpp:1644-1656, 1738-1740`; `docs/changes.txt:556`). `SetExtraControlCreator` → XP-style dialog. |
| macOS | Open dialogs show **no filter choice** and apply all wildcards at once unless `wxSystemOptions::SetOption(wxOSX_FILEDIALOG_ALWAYS_SHOW_TYPES, 1)`, and even then non-matching files are only greyed (`interface/wx/filedlg.h:117-128`). Matching compares the lower-cased last path extension (`wxOpenSavePanelDelegate panel:shouldEnableURL:`, `src/osx/cocoa/filedlg.mm:58-88`) [source]: case-insensitive, and a multi-dot pattern such as `*.gcode.3mf` or `*.zip.amf` never matches by itself. `wxFD_OVERWRITE_PROMPT` is always on (`:172-175`); `wxFD_OPEN` always behaves as `wxFD_FILE_MUST_EXIST` (`:184-189`). The save panel replaces the initial file name's extension with the first one in the wildcard (Orca's comment on `file_wildcards`). The native panel runs `runModal` and does not re-raise the parent dialog afterwards (`src/osx/cocoa/filedlg.mm``ShowModal`); see `references/windows-dialogs.md` for the deferred re-raise. |
| GTK3 | `GtkFileChooserNative` (portal-capable, e.g. Flatpak) when GTK ≥ 3.20 at runtime and neither `wxFD_PREVIEW` nor an extra control/customize hook is used (`src/gtk/filedlg.cpp:265-274, 437-443`; `src/gtk/dirdlg.cpp:122-130`; `docs/changes.txt:393`). It runs through `gtk_native_dialog_run`: there is no wx window, so size/position calls do nothing (`wxFileDialog::DoSetSize` is empty). Patterns go to `gtk_file_filter_add_pattern` per token and are **case-sensitive** (`src/gtk/filectrl.cpp:169`). The chooser is transient for the parent left after `GetParentForModalDialog`, which replaces a null (or hidden, dying or `wxWS_EX_TRANSIENT`) parent with the active top-level window, else the app's main top window (`src/gtk/filedlg.cpp:206, 223-225`; `src/common/dlgcmn.cpp:180-203``DoGetParentForDialog`) [source]. `wxFD_PREVIEW` is GTK-only (`interface/wx/filedlg.h:195-197`). |
| GTK2 (opt-out build) | never uses the native chooser; patterns case-sensitive. |
**OrcaSlicer:**
- `file_wildcards(FileType, custom_extension)` (`GUI_App.cpp`) builds every filter from
`file_wildcards_by_type`: translated title, and an upper-case twin of every extension
(`*.stl;*.STL`) because GTK patterns are case-sensitive. A `custom_extension` is put first because
the macOS save panel substitutes the first extension into the initial file name.
- Openers to copy: `GUI_App::import_model`, `GUI_App::import_zip` (default dir `from_u8(...)`).
Last directories come from `AppConfig::get_last_dir()`/`get_last_output_dir()`; the app stores
them itself (`update_config_dir`, `update_last_output_dir`) instead of using `wxFD_CHANGE_DIR`.
- `Plater::priv::get_export_file` appends the expected extension on `__WXMSW__` when the returned name
lacks it and then asks its own overwrite question, because the native overwrite prompt only checked
the name the user typed.
- `CheckboxFileDialog` (`GUI_Utils.hpp/.cpp`) uses `SetExtraControlCreator`, which costs the native
dialog on MSW and GTK3; prefer `SetCustomizeHook` for new extra controls.
**Pitfalls:**
- **Rule:** give file and directory dialogs the top-level window they belong to as parent.
**Why:** with `nullptr`, wxGTK (`wxFileDialog::Create`, `wxDirDialog::Create`) and wxMSW
(`wxFileDialog::ShowModal`) substitute whatever top-level window is active at that moment, else
the app's main top window (`wxDialogBase::DoGetParentForDialog`, `src/common/dlgcmn.cpp:180-203`)
[source], so the chooser becomes transient for and modal over an arbitrary window (a modeless
dialog or web window that happened to be active), or gets no parent when none qualifies. Gizmo
code has no `this` window to hand, which is where `nullptr` creeps in.
on Windows (`explorer /select,<path>` for `desktop_open_any_folder`), `openFolderForFile`
(any folder) or `open` (data dir) on macOS, `xdg-open` on Linux (of the containing folder when
`path` is a file) with the AppImage variables (`APPIMAGE`, `APPDIR`, `LD_LIBRARY_PATH`,
`LD_PRELOAD`, `UNION_PRELOAD`) removed and `OWD` as the working directory. A bare `wxExecute("xdg-open …")` or
`wxLaunchDefaultApplication` from an AppImage passes Orca's bundled libraries to the file manager.
- Project attachments: `desktop_open_project_attachment` accepts only a regular file inside the
project's auxiliary temp folder (`is_absolute_path_within_root`) and asks before opening anything
`is_safe_to_open_file_name` (`libslic3r/utils.cpp`) does not allow-list, because the desktop would
run a script or executable without a download warning. Reuse it for any file that arrived inside a
project or from the network.
## Clipboard
**Contract** (`interface/wx/clipbrd.h`): `Open()` "should be tested"; keep the clipboard open "only
momentarily" (`:24-29, 145-155`). `SetData` replaces any previously set object, so several formats
need one composite object; "After this function has been called, the clipboard owns the data"
(`:157-170`). `Flush()` keeps the data after exit; implemented on MSW and GTK only, on GTK only for
the CLIPBOARD selection and with a clipboard manager running (`:56-61, 100-115`).
`UsePrimarySelection(true)` makes every operation fail on platforms without a PRIMARY selection
(`:172-187`). `wxClipboardLocker` (`include/wx/clipbrd.h:167-191`) opens in its ctor, closes in its
dtor, and `!lock` tests `IsOpened()`.
**Platforms** [source]:
| | Behaviour |
|---|---|
| MSW | `SetData`/`GetData` do not check `Open()` (`src/msw/clipbrd.cpp``wxClipboard::SetData`), so code that forgets it works here only. `wxTextDataObject::SetData()` size must now include the 2-byte NUL (`docs/changes.txt:62-66`); use `SetText()` or the ctor. |
| macOS | without `Open()`, `SetData`/`AddData`/`GetData` return false through `wxCHECK_MSG(m_open, …)` without taking the object (`src/osx/carbon/clipbrd.cpp:84-105, 140-147`); a successful write is flushed to the pasteboard immediately (`:110-114`). |
| GTK | same `wxCHECK_MSG(m_open, …)` early return (`src/gtk/clipbrd.cpp:643-647, 775`). `GetData`/`IsSupported` are asynchronous underneath: `wxClipboardSync` spins `YieldFor(wxEVT_CATEGORY_CLIPBOARD)` until GTK answers and forbids re-entrancy (`src/gtk/clipbrd.cpp:68-92`). PRIMARY selection exists. Under Wayland, Wayland MIME types are advertised next to the X11 atoms (`src/gtk/clipbrd.cpp:655-700`). |
```cpp
wxClipboardLocker lock;
if (!lock)
return;
wxTheClipboard->SetData(new wxTextDataObject(text)); // clipboard owns it on success
```
**OrcaSlicer.** Models: the copy button in `TroubleshootDialog` (`wxClipboardLocker`) and
`ImGuiWrapper::clipboard_set`/`clipboard_get` (tested `Open()`, UTF-8 via `wxString::FromUTF8`/
`into_u8`). The 3D scene's copy/paste is an internal clipboard (`Selection::Clipboard`,
`Selection::copy_to_clipboard`/`paste_from_clipboard`), not the wx one.
- **Rule:** open the clipboard through a checked `wxClipboardLocker` before `SetData`/`GetData`, and
test the result.
**Why:** on GTK and macOS an unopened clipboard makes `SetData` return false without taking the
object (leak, nothing copied) and `GetData` return false (nothing pasted); MSW does not check, so
the bug passes Windows testing.
```cpp
// Wrong:
wxTheClipboard->Open();
wxTheClipboard->SetData(new wxTextDataObject(t));
wxTheClipboard->Close();
// Right:
wxClipboardLocker lock;
if (lock)
wxTheClipboard->SetData(new wxTextDataObject(t));
```
## Drag and drop
**Contract.** `SetDropTarget`: "If the window already has a drop target, it is deleted"
(`interface/wx/window.h:3651-3657`); the window also deletes its target in its destructor
(`src/common/wincmn.cpp:515`), so the window owns it. `DragAcceptFiles` "Cannot be used together
with SetDropTarget() on non-Windows platforms" (`interface/wx/window.h:3659-3672`). The
`wxDropTarget` destructor deletes its data object and `SetDataObject` deletes the previous one
false to refuse) → `OnData` (`interface/wx/dnd.h:44-45, 75-127`). `wxFileDropTarget::OnDropFiles(x, y,
filenames)` returns true to accept (`interface/wx/dnd.h:395-422`). `wxDropSource::SetData` "will not
delete any previously associated data", i.e. the source does not own it (`interface/wx/dnd.h:335-338`).
`DoDragDrop(flags)` "blocks the program until the user releases the mouse button", and the target
cannot change the result code the source gets (`docs/doxygen/overviews/dnd.h:44-50, 84-90`).
`wxDragMove` is reported on MSW only (`interface/wx/dnd.h:26`).
**Platforms** [source]:
| | Behaviour |
|---|---|
| MSW | `SetDropTarget` revokes and deletes the old target immediately (`src/msw/window.cpp:1753-1761`). `wxDropTarget::MSWUpdateDragImageOnLeave()` (undocumented, `include/wx/msw/ole/droptgt.h:73`) hides the shell drag image. |
| GTK | `SetDropTarget` unregisters and deletes immediately (`src/gtk/window.cpp:6605-6616`). `DoDragDrop` returns `wxDragNone` unless a mouse button is down after a mouse event (`src/gtk/dnd.cpp:835-850`), then runs a nested `gtk_main_iteration` loop with `g_blockEventsOnDrag` set (`:906-916`). Under Wayland it also hooks button/motion events because "drag-end" may never arrive (`:945-955`). |
| macOS | `SetDropTarget` deletes the old target immediately (`src/osx/window_osx.cpp:616-622`). |
**wxDataViewCtrl DnD** (`interface/wx/dataview.h`): `EnableDropTargets` is fully implemented in the
generic and native macOS versions, wxGTK uses only the first format (`:1376-1386`); `SetDragFlags` is
honoured only by the generic control (MSW), not native GTK/macOS (`:4010-4024`); `GetDropEffect`
returns `wxDragNone` on native GTK and macOS (`:4026-4041`); `GetProposedDropIndex` works from
`ITEM_DROP` everywhere and from `ITEM_DROP_POSSIBLE` except native GTK (`:4053-4064`). The control
itself is in `references/controls-dataview.md`.
**OrcaSlicer:**
- Model-file drops: `PlaterDropTarget : wxFileDropTarget` (`Plater.cpp`), installed with
`q->SetDropTarget(new PlaterDropTarget(...))` (the window takes ownership); `SetDefaultAction(wxDragCopy)`.
`OnDropFiles` calls `MSWUpdateDragImageOnLeave()` under `WIN32` before showing UI, raises the main
frame, switches to the Prepare tab, routes a single `.svg` to the SVG gizmo (`GLGizmoSVG::create_volume`
at the drop point) and otherwise calls `Plater::load_files`. `Preview::set_drop_target(target)` is
the helper for putting a target on the preview window.
1. Free heap-allocated windows with `Destroy()`, never `delete`; a modal dialog on the stack is freed by its scope and must never be `Destroy()`ed. §2
2. Never `Destroy()` a child window (or an ancestor of it) from inside that child's own event handler: child `Destroy()` is a synchronous `delete this`. Defer with `wxTheApp->ScheduleForDestruction(w)` or a guarded `CallAfter`. §2
3. A close handler either vetoes or destroys/ends the window; prompts and vetoes only when `CanVeto()`; when `!CanVeto()` it must not veto. §2
4. Do irreversible teardown in a close handler only once every veto is decided — the default top-level handler still vetoes a vetoable close while any modal dialog is open. §2
5. Give every dialog and frame a live top-level parent (`parent ? parent : wxGetApp().mainframe`); a parentless window that only hides on close keeps the process alive. §1, §2
6. Never construct a window to reach computation; make the logic `static` or free. §1
7. To ask "is this top-level window still alive", check `wxTheApp->IsScheduledForDestruction(w)` as well as `IsBeingDeleted()`; `wxWeakRef` nulls only at the very end of destruction; cross-thread code uses a `std::shared_ptr<std::atomic<bool>>` alive flag. Re-check after every nested event loop. §3
8. A `wxEVT_DESTROY` handler on a parent fires for every descendant: compare `GetEventObject()` and `Skip()`. For a top-level window it runs after the derived destructor — do derived cleanup in the destructor. §3
9. Accessors reachable from teardown-time events return `nullptr` instead of dereferencing a pimpl or child; the real fix is still to stop the events first. §3
10. Bring a top-level window forward with `if (!w->IsShown()) w->Show(); w->Raise();` — `Raise()` never shows, and a redundant `Show()` is not inert on GTK3. §4
11. On macOS, after a native modal (file/dir dialog, native message box) or a generic progress dialog opened from a secondary top-level window, re-raise that window with a deferred, liveness-guarded `Raise()`. §4
12. Remove style bits with `& ~flag`, never `!flag`. §5
13. Always pass the style to a `DPIDialog`: the `DPIAware` default is `wxDEFAULT_FRAME_STYLE` (resizable, min/max boxes). §5, §7
14. A `wxFRAME_FLOAT_ON_PARENT` frame takes a non-null top-level parent (parent to `mainframe`, not to a panel). §5
15. End a modal dialog with `EndModal(wxID_*)` (or `EndDialog` from inside the class); never with `Hide()`, `Show(false)` or `Destroy()` while its loop runs. A heap modal is `Destroy()`ed after `EndModal()` — normally by the caller once `ShowModal()` returns. §6
16. Return codes are `wxID_*` ids, never the `wxOK/wxCANCEL/wxYES/wxNO/wxCLOSE` style bits (the `MsgDialog` "Go to" button's `wxFORWARD` is the one deliberate exception). §6, §9
17. Nested modal dialogs end innermost first; `DPIDialog::EndModal` refuses (logs, leaves the dialog open) on a dialog that is not the innermost `DPIDialog`. §6, §7
18. Never call `EndModal()` on a modeless dialog; use `Hide()`, `Close()`, `Destroy()`, or `EndDialog(rc)` inside the class. §6
19. Never show a modal dialog directly from a mouse-button or motion handler; defer with `CallAfter`. §6
20. Run cancel cleanup on every close path: ESC and the title-bar close box go through `Close()` and `wxID_CANCEL`, never through an Orca `Button`'s handler. §6, §8
21. Dialogs derive `DPIDialog`, implement `on_dpi_changed`, call `CenterOnParent()` after fitting, and call `wxGetApp().UpdateDlgDarkUI(this)` last; an override of `ShowModal()` calls `DPIDialog::ShowModal()`. §7, §8
22. The bottom row is `DialogButtons` with untranslated labels; bind every button whose default wx handling is not what you want — Yes/No/Confirm/custom buttons never close the dialog by themselves. §8
23. Messages to the user go through the `MsgDialog` family (`MessageDialog`, `RichMessageDialog`, `WarningDialog`, `ErrorDialog`, `InfoDialog`, `show_error`/`show_info`), never `wxMessageBox`/`wxMessageDialog` once the GUI exists; treat `wxID_CANCEL` (ESC/close box) as "no". §9
24. `show_error` is asynchronous and captures its parent pointer; pass a long-lived parent or `nullptr`. §9
---
## 1 Creating and parenting windows
**Contract.**
- Child windows are created shown; top-level windows (frames, dialogs) are created hidden "to allow
you to create their contents without flicker" (`interface/wx/window.h:3140-3158`).
- "Child windows are deleted from within the parent destructor. This includes any children that are
themselves frames or dialogs, so you may wish to close these child frame or dialog windows explicitly
from within the parent close handler" (`docs/doxygen/overviews/windowdeletion.h:96-100`).
- `wxDialog` with a NULL parent "will be owned by the application's top window, if any. Use
`wxDIALOG_NO_PARENT` style to really make dialog not owned by any window" (`interface/wx/dialog.h:166-172`).
The style doc adds: orphan dialogs are "not recommended for modal dialogs" (`interface/wx/dialog.h:122-126`).
- **[source]** Only the *native* owner is substituted; `GetParent()` stays NULL. Owner resolution
(`wxDialogBase::DoGetParentForDialog`, `src/common/dlgcmn.cpp:180-204`): the given parent's
top-level parent → the active window's top-level parent → `wxApp::GetMainTopWindow()`, each
rejected if it is pending delete, being deleted, has `wxWS_EX_TRANSIENT`, or — for modal use — is
not `IsShownOnScreen()` (`CheckIfCanBeUsedAsParent`, `src/common/dlgcmn.cpp:130-178`).
**Two-step creation.** "the underlying window must be created exactly once… if you use the default
constructor… you *must* call Create() before using the window and if you use the non-default
constructor, you can *not* call Create()" (`interface/wx/window.h:411-417`). Methods may be called
between the C++ constructor and `Create()` — `Hide()` to build invisibly, `Enable(false)`,
`SetExtraStyle()` for create-time extra styles (`interface/wx/window.h:419-428, 3120-3126`; `interface/wx/dialog.h:127-133`).
```cpp
auto* panel = new wxPanel(); // C++ object only
panel->Hide(); // legal before Create
panel->Create(parent, wxID_ANY); // native window, still hidden
/* build children */ panel->Show();
```
**[source]** `Destroy()` on a never-created window skips `wxEVT_DESTROY` (`src/common/wincmn.cpp:559-573`),
and a never-created top-level window is deleted immediately (`src/common/toplvcmn.cpp:102-112`; not on
- Defaults (`docs/doxygen/overviews/windowdeletion.h:63-73`): wxDialog's close handler simulates `wxID_CANCEL`; the cancel
handler hides a modeless dialog or `EndModal(wxID_CANCEL)`s a modal one — "the dialog *is not*
destroyed (it might have been created on the stack)". wxFrame's default close handler calls `Destroy()`.
**[source]** The top-level default (`wxTopLevelWindowBase::OnCloseWindow`, `src/common/toplvcmn.cpp:535-546`)
**vetoes** a vetoable close while `wxModalDialogHook::GetOpenCount() > 0` (any app-modal wx or native
dialog open), otherwise calls `Destroy()`.
- Deleting from inside a handler: "it may be unsafe for an event handler to delete the object which
generated the event because more events may be still pending for the same object. In this case the
handler may call ScheduleForDestruction() instead" (`interface/wx/app.h:191-212`); without an event
loop it deletes immediately.
**When does `Destroy()` actually free the window?** **[source]**
| Window | `Destroy()` does | Cite |
|---|---|---|
| Child (control, panel) | sends `wxEVT_DESTROY`, then `delete this`**synchronously**; the window detaches from its containing sizer in `~wxWindowBase` | `src/common/wincmn.cpp:559-573` |
| Top-level, normal case | appends to `wxPendingDelete`, hides unless it is the last visible top-level window; deleted by `DeletePendingObjects()` at the next idle — which also runs inside full yields and modal loops (`wxYield`, `ShowModal`), not inside a masked `YieldFor` (progress-dialog `Update`) | `src/common/toplvcmn.cpp:102-142`; `src/common/appbase.cpp:643-662` |
| Top-level, parent already being deleted, or never created (not macOS) | deleted immediately | `src/common/toplvcmn.cpp:102-112` |
| Top-level, MSW | base behaviour plus `wxWakeUpIdle()` so iconized windows still get deleted | `src/msw/toplevel.cpp:771-784` |
| Any child of a dying parent, top-level children included | `DestroyChildren` calls the non-virtual `wxWindowBase::Destroy()`: deleted **immediately**, not queued | `src/common/wincmn.cpp:586-607` |
| `wxPopupTransientWindow` | deferred to idle (not hidden); a second `Destroy()` fails a `wxCHECK` — see `references/popups-menus.md` | `src/common/popupcmn.cpp``wxPopupTransientWindowBase::Destroy` |
**Close-handler shapes** (pick one per window, never mix; `confirm_discard`/`cleanup` stand for your code):
```cpp
// destroy-on-close (a frame, or a modeless dialog the user owns)
| inside `wxEVT_DESTROY` handlers and the wx base destructors | true | false (already removed) | **non-null** until `~wxTrackable` |
| inside a top-level window's own derived destructor (deleted at idle or by `delete`) | **false** — `SendDestroyEvent()` runs later, in `~wxFrameBase` or the port's top-level/dialog destructor (`src/common/framecmn.cpp:200`, `src/msw/toplevel.cpp:522`, `src/gtk/toplevel.cpp:975`, `src/osx/dialog_osx.cpp:90`) | false | non-null |
**`wxWeakRef<T>`** auto-resets "when the object pointed is destroyed"; works for `wxEvtHandler`/`wxWindow`
(`interface/wx/weakref.h:40-99`). **[source]** The reset happens in `~wxTrackable`, the last base
destructor (`include/wx/tracker.h`) — non-null throughout the destroy sequence and while a top-level window
sits in the pending list. The tracker list is unlocked: main thread only. Orca uses it for the splash
(`wxWeakRef<SplashScreen> scrn` in `GUI_App::on_init_inner`), `g_delay_webviews` (`Widgets/WebView.cpp`),
`DockPanel`, `GuideFrame`.
**`wxWindowPtr<T>`** is a shared pointer that calls `Destroy()` at refcount 0
(`interface/wx/windowptr.h:10-26`). It does not track: if the window dies another way (parent deletion,
default frame close) the last release calls `Destroy()` on freed memory. Use it only for parentless,
self-owned top-level windows — the documented `ShowWindowModalThenDo` idiom (§6).
**`wxEVT_DESTROY`** (`wxWindowDestroyEvent`, `interface/wx/event.h:4525-4553`): for top-level windows it
is sent "by wxFrame or wxDialog destructor, i.e. after the destructor of the derived class was executed";
for children "just before deleting the window from wxWindow::Destroy()… or from the window destructor if
operator delete was used directly". **[source]** It derives from `wxCommandEvent` and **propagates to the
parent** unless the parent is being deleted (`wxWindowBase::TryAfter`, `src/common/wincmn.cpp:3499-3522`), so a
parent's handler fires for every descendant destroyed before it. Derived-class code that must run at
destruction goes in the derived destructor (or call `SendDestroyEvent()` there).
**Pending events and nested loops.** **[source]** `ProcessPendingEvents` runs queued events and
`CallAfter`s of a top-level window that is already in `wxPendingDelete` (`src/common/appbase.cpp:561-601`); only the
actual deletion discards them (`~wxEvtHandler` → `DeletePendingEvents`, `src/common/event.cpp`). Idle
events are skipped for pending-delete windows (`src/common/appcmn.cpp:405-428`). "Deferred" does not mean "after the
current handler returns": `ShowModal()`, `wxYield()` and `wxSafeYield()` run pending events *and*
`DeletePendingObjects()` (`src/common/evtloopcmn.cpp:172-192`), so a raw pointer to a `Destroy()`ed
window dies across any of them, and a lambda queued before a modal opens can run while the caller is
still inside `ShowModal()`. A masked `YieldFor` — `wxProgressDialog::Update`/`Pulse` — runs no idle pass
and only the queued events its categories allow (none on GTK): `references/threads-timers-app.md`
§Event categories and yields. `CallAfter` liveness mechanics (self-queued calls are dropped
with their handler; `wxGetApp().CallAfter` calls are not): `references/events.md` §CallAfter. Yields:
`references/threads-timers-app.md`.
**OrcaSlicer.**
- Alive flag for anything that may run after the window died, including worker threads:
`std::shared_ptr<std::atomic<bool>> m_alive = std::make_shared<std::atomic<bool>>(true);`, set false in the
destructor (`PluginsDialog::~PluginsDialog`), captured by value, checked inside the lambda. Unlike
`wxWeakRef` it flips at the *start* of the derived destructor and is thread-safe.
- App shutdown gate: `wxGetApp().is_closing()`; background-to-UI callbacks check it before posting and
again inside the `CallAfter` lambda (`references/threads-timers-app.md`).
**Pitfalls.**
- **Rule:** For a top-level window that something may have `Destroy()`ed, test both flags.
```cpp
// Wrong: if (!tlw->IsBeingDeleted()) use(tlw);
// Right: if (!wxTheApp->IsScheduledForDestruction(tlw) && !tlw->IsBeingDeleted()) use(tlw);
```
- **Rule:** Don't trust `wxWeakRef` inside teardown code; combine it with `IsBeingDeleted()` /
`IsScheduledForDestruction()` or an alive flag.
- **Rule:** In a `wxEVT_DESTROY` handler bound on a parent, compare the event object and `Skip()`.
**Show / Hide.** `Show()` returns false when nothing changed (`interface/wx/window.h:3140-3158`).
`IsShownOnScreen()` = shown and every parent up to the top-level window shown (`interface/wx/window.h:3100-3106`).
**[source]** GTK3: `wxTopLevelWindowGTK::Show` calls `GTKSendSizeEventIfNeeded()` even when nothing
changes, so a redundant `Show(true)` can flush a pending size event into layout handlers
(`src/gtk/toplevel.cpp:1258-1268`). X11 (GTK2/GTK3 without client-side decorations): the first `Show()`
may be deferred until `_NET_FRAME_EXTENTS` arrives — `IsShown()` is already true but the window is not
mapped (`src/gtk/toplevel.cpp:1141-1243`).
**Raise.** "only requests the window manager to raise this window… If the window is currently hidden,
this function does *not* show it", top-level windows only (`interface/wx/window.h:3013-3033`); true on all ports
since 3.3 (`docs/changes.txt:144-146`). **[source]** MSW = `::SetForegroundWindow`, subject to the
foreground lock — Windows may only flash the taskbar button (`src/msw/toplevel.cpp:650-655`); GTK =
`gtk_window_present` only if shown (`src/gtk/toplevel.cpp:1301-1310`; during a deferred X11 first show it
already counts as shown); macOS = `makeKeyAndOrderFront` only if shown (`src/osx/nonownedwnd_osx.cpp:289-295`, `src/osx/cocoa/nonownedwnd.mm:897-899`),
which also makes a `wxPopupWindow` the key window — never `Raise()` a popup (`references/popups-menus.md` §5, §10).
**Enable.** `Enable(false)` on a parent disables children logically: `IsEnabled()` reflects ancestors,
`IsThisEnabled()` the window's own flag (`interface/wx/window.h:3060-3070, 3116-3138`). **[source]** On MSW/macOS wx
propagates through `NotifyWindowOnEnableChange` → `DoEnable` on non-top-level children — not through the
virtual `Enable()` — and skips children entirely when a top-level window is disabled, so a modal dialog
does not grey the frame (`src/common/wincmn.cpp:1150-1191`); GTK relies on native sensitivity. Orca widgets update
their painted state only from their own `Enable()` override, so disabling an ancestor leaves them
looking enabled: `references/orca-widgets.md`.
**State and geometry** (`interface/wx/toplevel.h`):
| API | Contract / platform note |
|---|---|
| `Iconize()`, `Maximize()` | on wxGTK "the change… is not immediate" (`:260-274, 335-346`); **[source]** MSW on a hidden window only records the state for the next show (`src/msw/toplevel.cpp:661-735`) |
| `Restore()` | on wxGTK does not unmaximize — call `Maximize(false)` (`:394-404`) |
| `ShowFullScreen(show, style)` | also shows a hidden window (`:730-750`) |
| `EnableFullScreenView()` | wxOSX only; then `ShowFullScreen` uses the native Spaces API and only `wxFULLSCREEN_NOTOOLBAR|NOMENUBAR` apply (`:697-728`); `wxEVT_FULLSCREEN` is macOS-only, only with it, and not generated by `ShowFullScreen()` (`interface/wx/event.h:2391-2412`) |
| `RequestUserAttention()` | documented for Win32 (taskbar flash) and wxGTK (`:375-392`); **[source]** macOS bounces the dock icon (`src/osx/cocoa/nonownedwnd.mm:1306`) |
| `SetIcon/SetIcons` | MSW needs a 16×16 or 32×32 icon; no effect on Wayland — ship a `.desktop` file (`:506-546`); **[source]** no macOS override: stored, never shown |
| `SetSizeHints/SetMinSize/SetMaxSize` | on a top-level window they also constrain programmatic `SetSize()` (`:576-603`) |
| `SetTransparent()` | on wxGTK call it before the first show (`:632-648`) |
| `EnableMaximizeButton/EnableMinimizeButton` | MSW and macOS only (`:172-203`) |
| `EnableCloseButton` | all ports, but its result is unreliable on X11, GTK included (`:160-170`) |
| `wxEVT_MOVE_START/END` | wxMSW only (`:80-87`) |
| `wxEVT_SHOW` | not sent for iconize/restore on MSW (`interface/wx/event.h:4905-4914`) |
| `SaveGeometry/RestoreToGeometry`, `wxPersistentTLW` | serializer-based persistence (`:406-492`; `interface/wx/persist/toplevel.h`) — not used by Orca |
**Wayland.** `SetIcon(s)` do nothing (`interface/wx/toplevel.h:518-521, 539-542`); the app id comes from
`SetClassName` with GTK ≥ 3.24.22 (`interface/wx/app.h:765-771`). `SetPosition()`/`Move()` on a
top-level window is a no-op (the compositor places windows), so `CentreOnParent` and saved positions are
ignored (observed; recorded in `GUI_App::window_pos_restore`, not documented by wx). Detection: `Slic3r::GUI::is_running_on_wayland()`; see `references/platforms.md`.
**OrcaSlicer geometry persistence.** Orca uses `AppConfig`, not `wxPersistentTLW`:
- `wxMINIMIZE_BOX`, `wxMAXIMIZE_BOX`, `wxCLOSE_BOX` implicitly enable `wxCAPTION` "on most systems"
(`interface/wx/dialog.h:94-114`, `interface/wx/frame.h:61-80`). `wxMAXIMIZE_BOX` is ignored on wxGTK without
`wxRESIZE_BORDER` (`interface/wx/frame.h:75-78`). The `wxMAXIMIZE` style works on Windows and GTK only; `wxICONIZE`/
`wxMINIMIZE` on Windows only (`interface/wx/frame.h:59-74`).
- `wxSTAY_ON_TOP`: above all other windows. `wxFRAME_FLOAT_ON_PARENT`: above its parent only, "must have a
non-null parent". `wxFRAME_NO_TASKBAR`: no taskbar button on Windows/GTK (GTK only with
`_NET_WM_STATE_SKIP_TASKBAR` support). `wxFRAME_TOOL_WINDOW`: small title bar, no taskbar button
(`interface/wx/frame.h:82-106`).
- Borders: `wxBORDER_NONE/SIMPLE/SUNKEN/RAISED/STATIC(MSW)/THEME`; `wxTRANSPARENT_WINDOW` is obsolete and
does nothing (`interface/wx/window.h:191-220`).
- Extra styles (`SetExtraStyle`, some must precede two-step `Create`): `wxWS_EX_BLOCK_EVENTS`,
`wxWS_EX_TRANSIENT` ("Don't use this window as an implicit parent… risk of creating a dialog/frame with
this window as a parent, which would lead to a crash"), `wxWS_EX_PROCESS_IDLE`,
`wxWS_EX_PROCESS_UI_UPDATES` (`interface/wx/window.h:262-290`). Dialogs set `wxWS_EX_BLOCK_EVENTS` by default
(`src/common/dlgcmn.cpp:124-127`): command events from inside a dialog never reach its parent frame;
frames do not block.
**Platforms. [source]**
| Style | MSW | macOS | GTK |
|---|---|---|---|
| MIN/MAX/CLOSE_BOX | force `WS_CAPTION` (`src/msw/toplevel.cpp:132-137`) — a custom title bar must strip it | — | — |
| frame with a parent, no `FLOAT_ON_PARENT` | unowned; gets its own taskbar button (`WS_EX_APPWINDOW`) unless `wxFRAME_NO_TASKBAR` (`src/msw/toplevel.cpp:193-244`) | — | — |
| `wxFRAME_FLOAT_ON_PARENT` | native owner = `GetHwndOf(parent)` as given (`MSWGetParent`, `src/msw/toplevel.cpp:212-240`) — pass a top-level window | `NSFloatingWindowLevel`, a *global* level: floats above other apps' windows too (`src/osx/cocoa/nonownedwnd.mm:835-876`) | `transient_for` the parent's top-level window, set only in `Create` (`src/gtk/toplevel.cpp:774-782`) |
| `wxFRAME_TOOL_WINDOW` | small caption, no taskbar | `NSFloatingWindowLevel` | no taskbar |
| `Hide()`/`Show(false)` on a modal dialog | exits the loop; `ShowModal()` returns the current code (0 if none set) | **does not exit**: clears the modality, orders the window out; `ShowModal()` stays blocked behind an invisible dialog, and wx's `EndDialog` path then takes the `Hide()` branch | calls the virtual `EndModal(wxID_CANCEL)`, overwriting any code |
| `EndModal()` on a modeless dialog | sets the code and hides (assert compiled out) | sets the code, hides, raises the parent | sets the code, then `wxFAIL` + return: the dialog stays visible |
| `IsModal()` between `EndModal()` and the return of `ShowModal()` | true (`m_modalData`, `include/wx/msw/dialog.h:48`) | false | false |
| `EndModal()` raises the parent | no | yes ("Prevent app frame from taking z-order precedence") | no |
| native owner / transient | fixed at `Create` (§1) | none (not a child window) | `transient_for` set in `ShowModal()` |
| how other windows are blocked | `wxWindowDisabler` in a generic `wxModalEventLoop` (`src/msw/dialog.cpp:70`) | Cocoa modal session (`src/osx/cocoa/evtloop.mm:424-456`) | `gtk_window_set_modal` grab; mouse capture released first |
**Buttons, ESC and the close box. [source]**
- `SetAffirmativeId(id)` (default `wxID_OK`): that button runs `Validate()` + `TransferDataFromWindow()` and
closes with the id (`interface/wx/dialog.h:480-494`). `SetEscapeId(id)`: default `wxID_ANY` = the `wxID_CANCEL`
button if present, else the affirmative one; `wxID_NONE` = ignore ESC; native dialogs cannot be
customized (`interface/wx/dialog.h:496-515`). `CreateStdDialogButtonSizer` makes `wxButton`s and sets the affirmative
id (`interface/wx/dialog.h:291-304`) — Orca uses `DialogButtons` instead (§8).
- Routing is by **id**, not type: `wxDialogBase::OnButton` (static table, `src/common/dlgcmn.cpp:455-480`) handles any
`wxEVT_BUTTON` that propagates to the dialog — affirmative id → `AcceptAndClose()` (`Validate()` +
`TransferDataFromWindow()` then `EndDialog(id)`), `wxID_APPLY` → validate + transfer, no close, escape id
or `wxID_CANCEL` → `EndDialog(wxID_CANCEL)`, anything else skipped. An Orca `Button` with `wxID_OK` or
`wxID_CANCEL` and no handler (or a handler that `Skip()`s) closes the dialog by itself; a handler bound
on the button that does not `Skip()` suppresses it.
- ESC → `wxDialogBase::OnCharHook` → `SendCloseButtonClickEvent()`: tries the escape id (`wxID_CANCEL`), then
the affirmative id, through `EmulateButtonClickIfPresent`, which does
`wxDynamicCast(FindWindow(id), wxButton)` and requires the button enabled and shown
(`src/common/dlgcmn.cpp:387-453`). Orca's `Button` derives from `StaticBox : wxWindow`, so emulation finds nothing:
in a plain `wxDialog` with Orca buttons ESC does nothing; in a `DPIDialog`, `DPIAware`'s own char hook
turns ESC into `Close()` first (§7).
- Close box (and `Close()`) → `wxDialogBase::OnCloseWindow` (`src/common/dlgcmn.cpp:525-559`): if shown, try
`SendCloseButtonClickEvent()`; when that finds no `wxButton`, `EndDialog(wxID_CANCEL)`. The Cancel
button's own handler never runs on this path.
**Modeless dialogs.** `Show()`; the close box only hides them (above). Long-lived modeless windows (monitor
pages, progress dialogs, plugin dialogs) either destroy themselves in a close handler or are deliberately
hide-on-close and reused (§2 shapes).
**Modality helpers.**
- `wxWindowDisabler(winToSkip, winToSkip2)` disables all *shown and enabled* top-level windows except the
skipped ones and re-enables them in its destructor (`interface/wx/utils.h:58-111`). MSW: a skipped window
that appears in the taskbar lets the user close the whole app from the taskbar (`interface/wx/utils.h:93-99`).
**[source]** The destructor re-enables every top-level window not recorded as skipped — including ones
created after the disabler (`src/common/utilscmn.cpp:1527-1546`); on macOS the constructor begins a Cocoa
modal session and **shows**`winToSkip` if it is not on screen (`src/osx/cocoa/evtloop.mm:458-514, 573-584`).
Keep disablers scoped and strictly nested.
- `wxModalDialogHook::GetOpenCount()` (since 3.3.0, `interface/wx/modalhook.h:118-126`) counts every open
modal — generic and native message/file/colour/font/print dialogs all use `WX_HOOK_MODAL_DIALOG`. Use it
for "is any modal open" in new code.
- `wxFrame::SetWindowModality(wxWindowMode)` (new in 3.3.2): call before showing; `AppModal`/`WindowModal`
(`interface/wx/frame.h:318-329`). **[source]** Applies immediately in the call, ends at the first hide, adds
`wxFRAME_NO_TASKBAR` and removes `wxMINIMIZE_BOX` (`src/common/framecmn.cpp:159-195, 304-339`); it does not
| scale factor, normal font | from `get_dpi_for_window(this)`; `SetFont(m_normal_font)`**except on macOS** (`#ifndef __WXOSX__`, avoids name cutting in `ObjectList`) | `references/dpi-bitmaps-fonts.md` |
| `CenterOnParent()` | runs **before any content exists** — re-centre after fitting | §8 |
| `SetupInspectorAccelerator(this)` | wxInspector toggle (Ctrl+Shift+I) in builds without `WXINSPECTOR_DISABLE`; a later `SetAcceleratorTable()` on the window replaces it | `references/orca-widgets.md` |
| MSW `update_dark_ui(this)` | no-op (its body only reads the dark flag) | `references/colours-dark-mode.md` |
| bind `wxEVT_DPI_CHANGED` (not macOS) | rescales (`Freeze` → font/em → `on_dpi_changed` → `Layout` → `Thaw`) when the scale changed and no monitor drag is in progress, and does **not**`Skip()` — so wx's default top-level handler (scale the window size by the DPI ratio) never runs (`interface/wx/event.h:3572-3583`) | `references/dpi-bitmaps-fonts.md` |
| bind `wxEVT_MOVE_START/END` (MSW-only events) | defer rescale while the window is dragged between monitors | `references/dpi-bitmaps-fonts.md` |
| `RichMessageDialog` | `(parent, message, caption = "", style = wxOK)` | adds `ShowCheckBox(text, checked)` / `IsCheckBoxChecked()` (a "Don't show again" check box added in its `ShowModal()`); its `SetYesNoLabels`/`SetYesNoCancelLabels`/`SetOKLabel`/`SetOKCancelLabels`/`SetHelpLabel` only store strings and never relabel a button |
| `InfoDialog` | `(parent, title, msg, is_marked = false, style = wxOK | wxICON_INFORMATION)` | caption is always "<app> information"; `title` is passed as the base's headline, which is not displayed |
| `DeleteConfirmDialog` | `(parent, title, msg)` | a plain `DPIDialog`, not a `MsgDialog`: Delete → `wxID_OK`, Cancel → `wxID_CANCEL` |
`DownloadDialog` and `FilamentWarningDialog` are single-purpose `MsgDialog` subclasses; read their
constructors before reusing them.
**Behaviour** (`MsgDialog.cpp`):
- Window style is always `wxDEFAULT_DIALOG_STYLE`; the `style` argument only selects buttons and icon. A NULL
parent becomes `wxGetApp().mainframe`.
- `MsgDialog::apply_style`: `wxOK` → OK, `wxYES` → Yes, `wxNO` → No, `wxCANCEL` → Cancel; Orca's use of
`wxFORWARD` adds a "Go to <forward_str>" button and turns OK into "Later" (`wxID_CANCEL`). Every button's
handler is `EndModal(btn_id)`, so compare with `wxID_OK/wxID_YES/wxID_NO/wxID_CANCEL` — and with
`wxFORWARD` (the style bit `0x2000`, used as the button id) for "Go to". OK/Yes/Go-to are Confirm-styled and
focused; `wxNO_DEFAULT`, `wxCANCEL_DEFAULT`, `wxHELP` and `wxSTAY_ON_TOP` are ignored.
- Icon from the style: `wxAPPLY` → "completed", `wxICON_WARNING` → "exclamation", `wxICON_INFORMATION` →
"info", `wxICON_QUESTION` → "question", otherwise the app logo; `wxICON_ERROR` greys it.
- ESC and the close box return `wxID_CANCEL` whatever the buttons (§7).
- Content (`add_msg_content`): plain text → a wrapped `Label` in a scrolled window, so `&` is a mnemonic —
escape user text (`references/strings-i18n-files.md`); with `link_text`/`link_callback`, `is_marked`,
code excerpts, or a message containing `<tr>` → a `wxHtmlWindow` with the text `xml_escape`d (`is_marked`
keeps `<`/`>` so markup works) and `\n` → `<br>`.
- Base helpers: `SetButtonLabel(wxID_*, label, set_focus = false)` relabels a created button;
`AddButton(id, label, set_focus = false)` appends a choice button that also ends with `EndModal(id)`;
`show_dsa_button(title = {})` adds a "Don't show again" `CheckBox` that posts `EVT_CHECKBOX_CHANGE`
(int = checked) to the dialog; `get_checkbox_state()`.
- `finalize()` (called by each subclass constructor): `SetSizeHints`, `Layout`, `Fit`, `CenterOnParent`,
`UpdateDlgDarkUI` — the recipe of §8 in one call; a custom `MsgDialog` subclass ends its constructor with it.
**Free helpers** (`GUI.hpp`): `show_error(parent, msg, has_code_excerpts = false)` is **asynchronous** — it
shows an `ErrorDialog` from `wxGetApp().CallAfter`, capturing the raw `parent`; `show_info(parent, msg,
title)` and `warning_catcher(parent, msg)` are synchronous `MessageDialog`s. The `const char*`/`std::string`
overloads decode UTF-8.
**Pitfalls.**
- **Rule:** Relabel buttons with `SetButtonLabel`, not the `RichMessageDialog::Set*Labels` methods.
```cpp
// Wrong: dlg.SetYesNoLabels(_L("Discard"), _L("Keep")); // stored, never shown
| `wxImage` ctor from XPM data is `explicit` | :225-226 | write `wxImage(xpm)` |
| `wxWindow::DoGetBorderSize()` removed | :228-229 | use `GetWindowBorderSize()` |
| MSVC 7 unsupported | :231-232 | none |
| 3.1.7: `wxImageFileProperty` members and internal `wxPropertyGridPageState` functions removed | :282-290 | no propgrid in Orca's own code (only the `wxInspector` dev dependency links it, on MSW/macOS) |
### 2.2 New APIs and behaviour in 3.1.6–3.2.0
| Change | Cite | Orca relevance |
|---|---|---|
| `wxBitmapBundle` added and used "throughout the entire API" (3.1.6) | `changes_32.txt:366` | setters take bundles (`wxBitmap` converts implicitly); **overrides** of virtuals that took `wxBitmap` break (5f365b5c6b). Bundles and sizing: `dpi-bitmaps-fonts.md` |
| Bitmap logical/DIP API: `CreateWithDIPSize` (3.1.6), `CreateWithLogicalSize` (3.3.0), `GetLogicalSize`; `CreateScaled` and `GetScaledSize/Width/Height` are "older synonyms … use the new function in the new code" | `interface/wx/bitmap.h:514, 551, 563-579, 706, 770-787` | new code uses the new names; Orca's older call sites (`BitmapCache`, `ScalableBitmap`) still compile |
| `wxDC::GetContentScaleFactor()` returns the effective DPI factor for window DCs (e.g. 1.5 on MSW), unlike `wxWindow::GetContentScaleFactor()` (always 1 on MSW), "since wxWidgets 3.1.6" | `interface/wx/dc.h:153-168` | do not divide bitmap sizes by the DC scale (`painting-custom-widgets.md`) |
| MSW: TLW resizing on DPI change improved and overridable — a handler that sizes the TLW itself does not `Skip()` (3.2.0) | `changes_32.txt:269`; `interface/wx/event.h` (`wxDPIChangedEvent`) | Orca's `DPIAware` handler does not `Skip()` (`dpi-bitmaps-fonts.md`) |
| `wxUILocale` (3.1.6); `wxLocale::IsAvailable` is now implemented through it | `changes_32.txt:348`; `src/common/intl.cpp:740-781` | see `wxUILocale::IsSupported` in §3 |
| wxOSX: "Allow user input in `wxPopupTransientWindow`" (3.1.7) — the popup's child holds mouse capture while the cursor is outside it and releases it inside, toggled from `OnIdle` [source] | `changes_32.txt:334`; `src/common/popupcmn.cpp:438-471` | how wxOSX transient popups track the mouse; the macOS gap-dismissal worked around in `SidePopup::Popup` is not traced to a specific wx change (§8; `popups-menus.md`) |
| wxGTK: Wayland fixes (3.1.6), no GDK errors from `PopupMenu()` on Wayland (3.1.7), `wxCURSOR_SIZING` on Wayland (3.2.0) | `changes_32.txt:397, 317, 261` | `platforms.md` |
| MSW: all native modal dialogs are app-modal (3.1.6) | `changes_32.txt:409` | `windows-dialogs.md` |
| Also new: `wxKeyEvent::IsAutoRepeat()`, `wxSpinCtrl::GetTextValue()/SetIncrement()`, `wxTopLevelWindow::SetContentProtection()`, `wxEVT_SPLITTER_SASH_POS_RESIZE`, OSX full-screen view options, OSX `wxEVT_CHAR` from `wxDataViewCtrl` | `changes_32.txt:364-436` | available API |
### 2.3 Older 3.x changes that still trip prior knowledge
These were already in force in 3.1.5, so the upgrade did not change them, but code written from
3.0-era knowledge gets them wrong (`changes_32.txt` line):
| Contract | Line |
|---|---|
| MSW `wxYield()` generates `wxEVT_IDLE` (idle handlers can run inside a yield) | :25-27 |
| A 0-width or 0-height `wxBitmap` fails on every port | :29-30 |
`changes.txt` "Changes in behaviour not resulting in compilation errors" (:11-146). Each row: the
change, its line, and what it means for Orca.
| Change | Line | Orca relevance |
|---|---|---|
| wxMSW needs Windows 7+ | :14-15 | drop XP/Vista assumptions (they need wx 3.2) |
| Fatal-error exit code is 255 everywhere (was 127 with MSVC); `wxApp::SetErrorExitCode()` (virtual, `interface/wx/app.h:886`), static `SetFatalErrorExitCode()` (:922) | :17-19 | update anything (CI, crash reporting, wrapper scripts) that matches the old code |
| `wxGLCanvas` no longer multisamples by default | :21-23 | pitfall below |
| `wxFileConfig` on Unix defaults to XDG `~/.config/appname.conf` (old file still used; `wxCONFIG_USE_XDG`/`wxCONFIG_USE_HOME`, `MigrateLocalFile()`) | :25-29 | none: Orca uses `AppConfig`, no `wxConfig` (`strings-i18n-files.md`) |
| `wxColourDatabase` uses CSS values; `UseScheme()` restores the old ones (`interface/wx/gdicmn.h:845, 973-996`) | :31-33 | none: stock colours (`*wxGREEN`, …) are fixed RGB (`src/common/gdicmn.cpp:789-830`) and Orca builds colours from hex/RGB. Do not build colours from names |
| `wxAuiNotebook` default art is the new flat art; `wxAuiNativeTabArt` (or `"native"` in XRC) keeps the old look | :35-37 | none: no `wxAuiNotebook` in Orca |
| `wxTHREAD_WAIT_DEFAULT` is `wxTHREAD_WAIT_BLOCK`: `wxThread::Delete()/Wait()` no longer pump events | :39-41 | Orca uses std/boost threads (`wxThread` only for `IsMain()`); a new `wxThread` whose exit needs the main loop would deadlock |
| `wxDocument::OnCloseDocument()` runs once, after the views are destroyed (it used to run twice when the document was closed from the menu); an override must not rely on any view existing | :43-46 | none: no doc/view |
| `wxGrid::FreezeTo()` asserts on out-of-range counts, and freezes even when the grid is too small | :48-51 | silent `false` in Orca; it also refuses reordered or drag-movable rows/columns [source] (`src/generic/grid.cpp:5742-5750`). Clamp arguments (`controls-dataview.md`) |
| Invalid `wxImageList` calls assert | :53-56 | silent in Orca; `wxCHECK` returns (generic `Add` → -1). Create the list with a valid size before use |
| `wxTRANSPARENT_WINDOW` does nothing (`#define wxTRANSPARENT_WINDOW 0`, `include/wx/defs.h:1449`); MSW code that needs it can set `WS_EX_TRANSPARENT` | :58-60 | pitfall below; 026b105dcb |
| MSW `wxTextDataObject::SetData()` size includes the 2-byte NUL (consistent with `GetDataSize()`); an old-style size chops the last character | :62-66 | none: Orca uses the `wxTextDataObject(text)` ctor; use `SetText()`, never `SetData()` |
| `wxListCtrl::EditLabel()` asserts without `wxLC_EDIT_LABELS` | :68-69 | silent no-op in Orca; add the style where editing is intended |
| MSW `wxSystemAppearance::IsDark()` reports the app's own mode; `AreAppsDark()`/`IsSystemDark()` report the OS | :71-73 | pitfall below |
| Unix `wxUILocale::IsSupported()` no longer falls back to another region of the same language; pass just `"fr"` to accept any `fr_XX` | :75-79 | `wxLocale::IsAvailable` builds the region-qualified tag (`GetCanonicalWithRegion()`) and calls `IsSupported()` [source] (`src/common/intl.cpp:740-781`); `GUI_App::load_language` relies on it, so on Linux a language whose canonical locale (e.g. `fr_FR`) is not installed reports unavailable. Keep the fallbacks in `load_language` |
| Deprecated `wxPGCellRenderer::DrawCaptionSelectionRect()` overload not called; override the overload taking `wxWindow*`, or enable 3.0 compatibility | :81-83 | none: no propgrid in Orca's own code |
| `wxImageList` size is in physical pixels | :85-88 | pitfall below |
| Mac `wxWebRequest` no longer uses persistent storage. The entry names `wxWebRequest::EnablePersistentStorage()`; the real API is `wxWebSession::EnablePersistentStorage(bool)` (`interface/wx/webrequest.h:1612-1629`; also `wxWebSessionSync`, :1835), macOS-only, before the first request | :90-92 | none: Orca uses `wxWebRequest` only for image downloads (`wxWebSession::GetDefault().CreateRequest` in `StatusPanel`, `SliceInfoPanel`, `ReleaseNote`, `DeviceErrorDialog`); network and login agents use libcurl (`Slic3r::Http`) |
| MSW `wxBitmap::Create(size, dc)` no longer multiplies by the DC's content scale; the size is physical | :94-96 | remove compensating scaling; use `CreateWithLogicalSize` for logical sizes |
| `wxIMAGE_QUALITY_NEAREST` has a new value and is no longer `wxIMAGE_QUALITY_NORMAL`; `NORMAL` (the `Scale`/`Rescale` default) is bilinear + box average (`interface/wx/image.h:31-67`) | :98-99 | never store or compare the enum numerically; default-quality thumbnails and icons look smoother than under 3.1.5; pass `wxIMAGE_QUALITY_NEAREST` for pixel-exact scaling |
| `wxTextCtrl::{Save,Load}File()` treat `.rtf` as RTF | :101-104 | pass `wxTEXT_TYPE_PLAIN` to keep plain text |
| `wxClientDC`/`wxPaintDC` offset their origin by a `wxFrame` toolbar on every port | :106-109 | none: Orca frames have no native toolbar (`BBLTopbar` is a child `wxAuiToolBar`) |
| `wxTextCtrl::SetLabel()` does nothing and asserts on every port (MSW used to act as `SetValue`) | :111-113 | pitfall below |
| `wxAuiGenericTabArt` subclasses (also via `wxAuiMSWTabArt`) override `DrawPageTab()`/`GetPageTabSize()` instead of `DrawTab()`/`GetTabSize()`; direct `wxAuiTabArt` subclasses still work | :115-120 | none: no Orca tab art |
| `wxAuiNotebook` page index is logical (reorder-independent); `GetPagePosition()` gives the screen position | :122-126 | none; index math under `wxAUI_NB_TAB_MOVE` is what breaks |
| `wxListbook`/`wxChoicebook` interpret mnemonics in page titles "just as the other wx*book classes already did" | :128-130 | Orca uses neither (`BedShapeDialog` uses `wxSimplebook` + a combo); every book interprets `&` (`interface/wx/bookctrl.h:141-147`) → rule 11 |
| `wxAUI_MGR_HINT_FADE` is not in the default `wxAuiManager` style | :132-133 | Orca's `AuiMgr` keeps the default flags (minus `wxAUI_MGR_ALLOW_FLOATING` on Wayland, `AuiMgr::init`), so the docking hint no longer fades; add the flag if wanted |
| `wxPrintDialogData::SetAllPages(false)`/`SetSelection(false)` changed meaning | :135-137 | none |
| `wxGetTranslation()` returns `wxString` by value | :139-142 | pitfall below |
| `wxWindow::Raise()` no longer shows a hidden window on any port | :144-146 | pitfall below; ba867cc534 |
**Pitfalls**
- **Rule:** Request multisampling explicitly; never rely on `wxGLAttributes::Defaults()` or a null
attribute list for MSAA.
**Why:** 3.1.5's `Defaults()` was RGBA, double buffer, depth 16 plus `SampleBuffers(1).Samplers(4)`
on every port; 3.3's is `RGBA().Depth(16).DoubleBuffer()` (`include/wx/glcanvas.h:175-178`), also used for a null
attribute list (`src/common/glcmn.cpp:159-165`). The change-log line has the arguments swapped:
`SampleBuffers(n)` is "number of sample buffers, usually 1" and `Samplers(n)` the samples per pixel
(`interface/wx/glcanvas.h:233-246`). The plater canvas is unaffected:
`changes.txt` "Changes in behaviour which may result in build errors" (:149-252).
| Change | Line | Orca relevance |
|---|---|---|
| 3.0-deprecated symbols disabled by default (`WXWIN_COMPATIBILITY_3_0=1` at wx build time re-enables them), 2.8-deprecated removed | :152-154 | Orca builds `WXWIN_COMPATIBILITY_3_0 0`, `_3_2 1`: port off 3.0-deprecated API |
| `wxUSE_UNICODE=0` unsupported | :156 | none |
| `wxUSE_STD_CONTAINERS=1` by default ("Container Classes" overview); building wx with 0 keeps the old containers | :158-161 | Orca builds with 1 → pitfall below |
| `wxUSE_STL` gone; implicit `wxString` → `std::[w]string` only with `wxUSE_STD_STRING_CONV_IN_WXSTRING=1` at wx build time | :163-166 | Orca builds 0: convert explicitly (`into_u8`, `ToStdString`, `ToUTF8()`) |
| MSW links `gdiplus.lib`, `msimg32.lib` | :168-172 | automatic with MSVC/wx-config; only static non-MSVC builds add them |
| wxMotif, wxGTK1 removed | :174-175 | none |
| Private containers (e.g. `wxSimpleDataObjectList`) removed; object arrays (`wxImageArray`) compare values in `Index()` | :177-183 | use `std::vector`/`std::list` |
| Operators on wx types are hidden (not global) | :185-189 | pitfall below |
| `wxString` from `std::string_view` makes `wxstr = {"Hello", 2}` ambiguous | :191-194 | write `wxString{"Hello", 2}` |
| 64-bit DLLs carry an `x64` suffix in all build systems | :202-204 | packaging scripts matching DLL names; Orca links wx statically (Flatpak builds it shared in its own manifest) |
| CMake config installs to `lib/cmake/wxWidgets-3.3`; plain `find_package(wxWidgets)` is unaffected, hard-coded paths break | :206-211 | Orca: `find_package(wxWidgets 3.3 CONFIG …)` on Windows/macOS, `wx-config --toolkit=gtk${SLIC3R_GTK}` on Linux (Orca's `src/CMakeLists.txt`) |
| Memory-tracing options removed | :213-216 | use ASan |
| Experimental MSW dark mode (#23028), with `wxApp::SetAppearance()` (#24461) | pitfall below; Orca kept NppDarkMode and only calls `MSWEnableDarkMode(DarkMode_Auto)` |
| Chromium `wxWebView` backend (#706) and `wxEVT_WEBVIEW_CREATED` | Chromium is not built; `wxEVT_WEBVIEW_CREATED` is the documented ready signal for Edge (`webview-gl-aui-media.md`) |
| WebP images (#25205) | Orca builds `wxUSE_LIBWEBP=builtin` and calls `wxInitAllImageHandlers()` (`GUI_App::on_init_inner`), so `wxImage` decodes WebP (e.g. downloaded thumbnails) |
| Pinned and multi-row AUI tabs (#25187, #25076) | none: no `wxAuiNotebook` |
| Synchronous `wxWebRequest` (#24760) | available (`wxWebSessionSync`); "must not be used from the main thread of GUI applications" (`interface/wx/webrequest.h:614-615`) |
| Raw touch events (#17077); wxGrid accessibility (#24368) | available |
| Unix power events and blockers (#22396, #23717) | could keep a Linux system awake during long work; Linux covers only `wxPOWER_RESOURCE_SYSTEM` and needs systemd ≥ 183 (`interface/wx/power.h:163-171`); not used by Orca |
| Native GTK file dialogs when possible (#24486, #25104) | the portal dialog is used only with GTK ≥ 3.20 at runtime, without `wxFD_PREVIEW` and without an extra control [source] (`src/gtk/filedlg.cpp:265-273, 438-443`); Orca's `CheckboxFileDialog` (`SetExtraControlCreator`, `GUI_Utils.hpp`) therefore gets the non-native GTK dialog |
| `wxWebRequest`: base URL (#24769), proxy (#24762), repeated headers (#24878); `wxWebSession::EnablePersistentStorage()` (#23743) | see §3 for storage |
| `wxString`: move operations (#23215), `std::string_view` ctor (#23711), faster and more robust `To/FromCDouble()` (#23287), `errno` preserved (#23113), `wc_string()` (#23463), `wxWARN_UNUSED` on the class (#24833, unused-variable warnings for `wxString` locals) | settings fields do **not** parse through `ToCDouble`: `Field` normalises the decimal separator and calls `wxString::ToDouble`, and `double_to_string` uses `wxNumberFormatter::ToString`; `ToCDouble`/`FromCDouble` appear only in a few helpers (e.g. `PreferencesDialog::create_camera_orbit_mult_input`). The `wxNumberFormatter` changes (3.3.1, 3.3.2) matter more for `Field` |
| Lambdas with `Bind()` without RTTI (#14850); move-only `wxMessageQueue` (#25026); `wxLogXXX(string)` safe with a single string (#25414) | available |
| Environment variables use UTF-8 (#25101) | check round-tripping of non-ASCII paths through `wxGetEnv`/`wxSetEnv` |
| Improved locale matching (#24855); thread-safe `wxPlatformInfo::Get()` (#25459); `wxXmlParseError` from `wxXmlDocument::Load()` (#24215); customisable error exit code (#24770) | available |
**All (GUI)** (:444-500)
| Change | Orca relevance |
|---|---|
| High-DPI wave: `wxCursorBundle` (#25374), animations (#23817), generic `wxListCtrl` (#22916), print preview (#24666), AUI dock art after DPI change (#23420), `wxBusyInfo` bitmaps (#23813) | AUI dock-art size metrics are DIP-like now: `wxAuiManager` reads them through `GetMetricForWindow()` (since 3.3.0), which scales them by the window DPI. The docs exempt `wxAUI_DOCKART_SASH_SIZE` and `wxAUI_DOCKART_PANE_BORDER_SIZE` (`interface/wx/aui/dockart.h:274-294`), but the default implementation exempts only `wxAUI_DOCKART_PANE_BORDER_SIZE` (and the non-pixel gradient type) and does scale the sash size [source] (`src/aui/dockart.cpp:199-226`). Pass DIP values to `SetMetric` (Plater's `SetMetric(wxAUI_DOCKART_CAPTION_SIZE, 18)`), never `FromDIP(…)` |
| Dark-mode colours in XRC (#23571); CSS colour names (#23518) | none: no XRC, no colour names |
| `wxTextCtrl::SearchText()` (#24756) and RTF (#24626); locale-aware date/time pickers (#23965, display format changes); `wxSearchCtrl` derives from `wxTextEntry` on all ports (#23686) | available |
| Non-live resize restored in wxAUI and `wxSplitterWindow` (#24193) | `wxAUI_MGR_LIVE_RESIZE` is in `wxAUI_MGR_DEFAULT` since 3.3.0 (`interface/wx/aui/framemanager.h:59-66`). The style table's "always enabled in wxGTK3 and wxOSX ports as non-live resizing is not implemented in them" (:199-206) is as stale as the `AlwaysUsesLiveResize()` note: that function "always returns false" as of 3.3.0 (:345; `src/aui/framemanager.cpp:711-714`), and the flag decides on every port [source] (`HasLiveResize`, :716-719). See `webview-gl-aui-media.md` §AUI docking |
| `wxInfoBar::ShowCheckBox()` (#25394); printing multiple page ranges (#25030); `wxGrid::CopySelection()` (#24124); new default flat AUI tab art (#25316); `wxApp::SetAppearance()` (#24461) | available |
| `wxGLCanvas` scale fixed with EGL/Wayland in high DPI (#23733) | Orca has no compensating hack: GTK3 `RetinaHelper::get_scale_factor` returns `GetContentScaleFactor()` |
| `wxKeyEvent::GetKeyCode()` fixed for non-US layouts (#23379) | shortcuts match through `KeyChord::from_event` (`mouse-keyboard-focus.md`) |
| Total window size with GNOME on X11 (#25348) | TLW geometry |
| `libwebkit2gtk-4.1` support (#23633) | wx's CMake build prefers 4.1 and falls back to 4.0 (`build/cmake/init.cmake:571-577`) |
| `libsecret` not required at runtime (#25355) | `wxSecretStore` loads it on demand: always check `IsOk()` (`interface/wx/secretstore.h:197-201`), as `OrcaCloudServiceAgent` does |
| `wxDC::DrawRoundedRectangle()` radius limited to half the smaller side (#24327) | the clamp is in the GTK2 GDK DC (`src/gtk/dcclient.cpp:874-875`, built only in the GTK2 opt-out, `GTK2_LOWLEVEL_SRC` in `build/files`) and also in the common `wxGraphicsPathData::AddRoundedRectangle` (`src/common/graphcmn.cpp:439-440`, absent in 3.1.5) that `wxGraphicsContext::DrawRoundedRectangle` uses, so every `wxGCDC` clamps: GTK3 (Cairo) and macOS window DCs, and Orca's memory-DC + `wxGCDC` paint paths on all ports [source]. A radius larger than half the smaller side is now clamped instead of drawing overlapping arcs |
| "Enable double buffering for all windows" (#22851) | **reverted in 3.3.2** (#25808) → pitfall below |
| `wxOverlay` reimplemented with layered windows (#23261) | none: no `wxOverlay` |
| `wxBG_STYLE_TRANSPARENT` implemented (#23412) | as `WS_EX_TRANSPARENT` on non-TLW children [source] (`src/msw/window.cpp:1581-1582`); set before `Create()` (`painting-custom-widgets.md`) |
| `wxCAPTION` turned on when min/max/close boxes are set (#23575) | `WS_CAPTION` is added to the style (`src/msw/toplevel.cpp:132-135`); origin of the MainFrame workaround (§8) |
| `wxButton` default size larger in high DPI (#25297) | raw `wxButton` layouts may grow; Orca dialogs use `Button`/`DialogButtons` |
| Markup in `wxStaticText` (#25000); `wxHyperlinkCtrl` colour changeable (#23549) | `SetLabelMarkup` now works on MSW (single line) |
| Extended-length paths (#25033); `wxFileDialog` no unwanted extension (#24949); UTF-8 build fixes (#23313); non-BMP strings (#25128) | available |
| `wxDisplay` invalidated on display change (#25396); TLW with one child resized on DPI change (#22983); Aero-snapped geometry saved | multi-monitor geometry |
| Modern default `wxTreeCtrl` look (#23844); RTL fixes for `wxOverlay`/`wxScrolled` (#25413) and GDI+ (#25431); `wxBitmap::UseAlpha()` returns `bool` (#23919) | available |
**wxOSX** (:585-592) — `wxEventLoop::OnExit()` is always called (#25409); TLW cursor setting fixed
(#25131); Cmd-C no longer activates a "Close" button (#25346); `wxCursor` loadable from resources
(#24374); `wxUIActionSimulator` works (#23692), usable for GUI tests.
**Pitfalls**
- **Rule:** Do not replace Orca's MSW dark mode with wx's (`SetAppearance`/`MSWEnableDarkMode`)
unless live theme switching is kept.
**Why:** wx's MSW dark mode cannot change once any top-level window exists or a mode was chosen:
| All | Persistence for `wxCheckBox` (#25515) and `wxRadioButton` groups (#25530); `wxAuiPaneInfo::FloatingClientSize()` (#25483); PNG "Description" chunk (#25556) | available |
| All | Settable app id (#25548): `wxAppConsole::SetClassName()` — the Windows AppUserModelID and the Wayland `app_id` (wxGTK ≥ 3.24.22), unused elsewhere; call it before any TLW, typically in the app ctor; on Windows it also changes shell behaviour (shift-middle-click new instance, shell MRU) (`interface/wx/app.h:760-812`) | Orca calls only `SetAppName`; setting a class name would change Windows taskbar grouping and jump lists |
| All | `wxDataViewCtrl::Collapse()` safe from event handlers (#25631); no `wxEVT_GRID_SELECT_CELL` at `wxGrid` creation (#25498); empty `wxGridSizer` no longer asserts (#25641); `wxPropertyGrid` compatibility (#25627); `wxNumberFormatter` (#25614, #25635); reproducible static Unix builds (#25502) | `ObjectList`/`ObjectGrid` event handlers; `Field` number formatting |
| wxGTK | Crash sorting a `wxDataViewCtrl` with a single leaf (#25625); `wxListCtrl` contents lost after `AppendColumn()` (#25519) | native GTK `wxDataViewCtrl` |
| wxMSW | Dark-mode fixes: disabled `wxButton` bitmaps (#25575), disabled `wxStaticText` (#25574), `wxComboCtrl` (#23766), `wxTE_RICH``wxTextCtrl` (#25602), selected toolbar buttons (#25616), `wxStaticBitmap` in `wxNotebook` crash (#25499), notebook background in high-contrast (#25542) | apply only where wx's own dark mode is active (`wxMSWDarkMode::IsActive()`; with Orca's `DarkMode_Auto`, whenever `ShouldAppsUseDarkMode()` reports dark, `src/msw/darkmode.cpp:205-227, 414-417`) |
| wxMSW | `wxDataViewCtrl` border in light mode (#25532); `wxTreeCtrl::EnsureVisible()` while frozen (#18435); preferred-languages buffer overrun (#25612); `wxAcceleratorTable` with 0 entries (#25517); date/time pickers on non-English Windows (#25511); per-window menu MDI crash (#25522) | available |
| wxOSX | Border look of `wxDataViewCtrl`, `wxListBox`, `wxTextCtrl` (#25570); startup crash with Farsi system language (#25561) | native macOS controls |
## 7 3.3.2 notable changes
`changes.txt:255-318`, relative to 3.2.10 (see §1 for the unlisted 3.2.9/3.2.10 fixes).
| All | `wxDC::DrawLabel()` bitmap position fixed in high DPI (#25888) | available |
| GUI | `wxGLContext::ClearCurrent()` (#25958) and `wxGLContext::GetProcAddress()` (#9215) — static members of `wxGLContext`, not `wxGLCanvas` as the change log says; `GetProcAddress` "is currently not implemented under macOS and always returns NULL" (`interface/wx/glcanvas.h:550-599`) | Orca loads GL with GLAD (through `eglGetProcAddress` on Wayland, `OpenGLManager::init_gl`) |
| GUI | `wxPersistentDVC` restores column positions (#26222); safer `wxTipWindow` close detection (#26070); wxDC-derived objects movable (#25726) | available |
| GUI | AUI: pane minimising (#23986), crash on hover after closing a notebook tab (#25959), notebook splitting (#26081) | Plater docking |
| GUI | LunaSVG option (#25902); `wxSVGFileDC` improvements (#25723); generic `wxCalendarCtrl` DPI-aware (#25713); `wxTextEntryDialog::SetHint()` (#26176); `wxStyledTextCtrlMiniMap` (#25887) | LunaSVG and STC are off in Orca's build |
| GUI | Many RTL layout fixes in wxMSW and wxGTK (#25426) | RTL languages |
| wxGTK | GLX and EGL in the same program (#26023): wx defaults to EGL even on X11 unless `PreferGLX()` or `wx_opengl_egl=0`; Wayland is always EGL [source] (`src/unix/glcanvas.cpp:218-245`; `interface/wx/glcanvas.h:1082-1102`) | pitfall below |
| wxGTK | `WarpPointer()` on Wayland compositors with the pointer-warp protocol (#23778); mutter moves the pointer only while a button is pressed (`interface/wx/window.h:3895-3914`) | Orca does not call `WarpPointer` |
| wxMSW | `WS_EX_COMPOSITED` use from earlier 3.3 reverted (#25808) | §5 pitfall |
| wxMSW | Dark-mode rendering of several controls (#25835), toolbar (#25892), menus (#26182); accessibility: `wxCheckBox` in dark mode (#26184), full `wxCheckListBox` (#25948) and `wxStyledTextCtrl` (#25956), basic `wxRichTextCtrl` (#26202) | wx dark mode only |
The upgrade landed as 8248b06337 ("Updated wxWidgets to 3.3.2", #12941; build system 2d7e26292b).
Its PR kept Orca's own MSW dark mode to avoid broader changes and because wx's needs an app restart to
switch. Each row is a rule for new code; the owning file has the detail.
| Commit | Rule for new code | Where | Owner |
|---|---|---|---|
| 8248b06337 | No version-conditional code for wx < 3.3: the upgrade deleted the pre-3.1.3 DPI-event shim (`DpiChangedEvent`, `EVT_DPI_CHANGED_SLICER`), the luma fallback in `check_dark_mode`, the `wxCHECK_VERSION` guards in `I18N.hpp`/`GUI_App.hpp`, and the macOS 10.9.5 `wxGLContext` hack | `GUI_Utils.hpp``DPIAware`, `OpenGLManager` | this file |
| 8248b06337 | `MSWEnableDarkMode(DarkMode_Auto)` runs before `NppDarkMode::InitDarkMode()`, so NppDarkMode's `SetPreferredAppMode(ForceDark` or `ForceLight`, per Orca's setting) overrides wx's `AllowDark` at OS level in both directions; `GUI_App::dark_mode()` honours `dark_color_mode` before `check_dark_mode()` | `GUI_App::on_init_inner`, `GUI_App::dark_mode` | `colours-dark-mode.md` |
| 8248b06337 | `wxToolTip::GetToolTipCtrl()` is private in 3.3 (`include/wx/msw/tooltip.h:92`); wx applies dark mode to the tooltip window itself through `wxMSWDarkMode::AllowForWindow`, which follows wx's mode [source] (`src/msw/tooltip.cpp:321`; `src/msw/darkmode.cpp:454-457`) | `GUI_App::force_colors_update` (the `#if wxVERSION_NUMBER < 3300` block is dead under 3.3) | `colours-dark-mode.md` |
| 8248b06337 | Backend webview factories override `GetVersionInfo(wxVersionContext)` without a default argument (`include/wx/msw/webview_edge.h:150`); pass `wxVersionContext::RunTime` when calling through the concrete factory | `WebView::CheckWebViewRuntime` | `webview-gl-aui-media.md` |
| 8248b06337 | The GL canvas gets `wxBG_STYLE_PAINT`; on MSW `GLCanvas3D::on_paint` renders immediately because idle events are not dispatched inside the modal resize loop (c06a0223a7) | `OpenGLManager::create_wxglcanvas`, `GLCanvas3D::on_paint` | `webview-gl-aui-media.md` |
| 8248b06337 | On MSW, do not `SetFocus()` the GL canvas while `wxCurrentPopupWindow` is set: the focus change makes wx call `MSWDismissUnfocusedPopup` and closes the search dropdown. `wxCurrentPopupWindow` is a wx-internal global (`src/msw/popupwin.cpp`) that Orca declares `extern` itself, usable only because wx is linked statically | `GLCanvas3D::on_mouse` (`evt.Entering()` branch) | `popups-menus.md` |
| 6148ba16b3 (in 8248b06337), 988b500f33 | On GTK a `wxBitmapToggleButton`/`wxButton`-based widget sized to exactly its bitmap leaves no room for the theme's CSS padding (GTK "negative content width" criticals). Either strip the native button CSS with `Slic3r::GUI::RemoveButtonBorder` and size to the bitmap (`CheckBox`), or size to `GetBestSize()` grown to the bitmap (`IncTo`; `RadioBox` and `SwitchButton` do both) | `RemoveButtonBorder` (`GUI_Utils.cpp`), `CheckBox::Rescale`, `RadioBox::Rescale`, `SwitchButton::Rescale` | `platforms.md` §GTK native chrome |
| 8248b06337 | The macOS-only vertical text nudges in `Button::render` and `SideButton::dorender` were removed; do not re-add per-OS baseline offsets | `Button::render`, `SideButton::dorender` | `painting-custom-widgets.md` |
| 8248b06337 | `wxEXPAND` combined with `wxALIGN_*` in a box sizer was cleaned up; the combination is ignored | `Sidebar::priv::layout_printer`, `AMSControl::createAmsPanel` | `sizers-layout.md` |
| 8248b06337 | `GUI_App::on_init_inner` filters known-harmless GTK criticals (allocation on hidden widgets, events on unrealised widgets, style-context calls from `SetBackgroundColour` before realisation); check that filter before chasing such a message | `GUI_App::on_init_inner` | `platforms.md` |
| 5f365b5c6b | `wxBitmapComboBox` overrides take `wxBitmapBundle` (`OnAddBitmap(const wxBitmapBundle&)`, `include/wx/bmpcbox.h:89`); `m_bitmaps` became `m_bitmapbundles`; get a bitmap with `GetBitmap(GetDefaultSize())` | `BitmapComboBox::OnAddBitmap`, `BitmapComboBox::OnDrawItem` (both macOS-only overrides) | `dpi-bitmaps-fonts.md` |
| ed88cbe3f5 → d62aa42e61 | 3.3's `wxWebViewWebKit` has neither the default ctor nor the creating `(parent, id, url, …)` ctor of 3.1.5; its only ctor is `explicit wxWebViewWebKit(const wxWebViewConfiguration&, WX_NSObject request = nullptr)` (`include/wx/osx/webview_webkit.h:36`). ed88cbe3f5 switched macOS to `wxWebView::New()`, which bypassed the subclass destructor that calls `RemoveScriptMessageHandler("wx")`; d62aa42e61 restored `WebViewWebKit` via `wxWebView::NewConfiguration(wxWebViewBackendWebKit)`. Only Linux uses `wxWebView::New()` | `WebView::CreateWebView`, `WebViewWebKit` | `webview-gl-aui-media.md` |
| 1765d296a8, 2b3328c2b2 | `wxArrayString` via `Add()`; `wxString` first in mixed concatenation; `compatibility_iterator` for wx lists (§4) | `Plater::import_model_id`, `OptionsSearcher::search`, `SendMultiMachinePage` | this file |
| 7ea69199fd | `dynamic_cast`, not `wxDynamicCast`, on `wxComboPopup` (§4) | `combochecklist_get_flags`/`_set_flags` (`GUI.cpp`) | this file |
| ba867cc534 | `Show()` (only if `!IsShown()`) before `Raise()` (§3) | `Plater::priv::priv` handlers, `Plater::priv::bring_instance_forward` | `windows-dialogs.md` |
| 026b105dcb | No `wxTRANSPARENT_WINDOW` (§3); it is a no-op `0`, so this was cleanup, not a build fix | `MainFrame` (`ResizeEdgePanel`, side-tool panels) | this file |
| eefdabcd98, f70d30bf79 (#13074) | The auto-added `WS_CAPTION` (3.3.0, #23575; eefdabcd98's message says 3.3.2) made `DefWindowProc` subtract a caption from the maximised client area, and on Windows 10 left the native frame visible behind the custom title bar ("double window"). The `MainFrame` ctor strips `WS_CAPTION` from `GWL_STYLE` right after creation (`SetWindowPos(… SWP_FRAMECHANGED)`); `WM_NCCALCSIZE` computes border thickness with `GetWindowLongPtr(hWnd, GWL_STYLE) & ~WS_CAPTION` and strips the border overshoot itself when maximised; f70d30bf79 restored the `wxEVT_MAXIMIZE` handler that clamps the maximised frame to the display client area | `MainFrame::MainFrame`, `MainFrame::MSWWindowProc` (`WM_NCCALCSIZE`), `AdjustWorkingAreaForAutoHide` | `platforms.md` §MSW title bar |
| 46e47cec0a | `wxGrid::GetSelectedBlocks()` (unordered, possibly overlapping, `interface/wx/grid.h:5091-5109`) is an empty range after the user deselects; compare `begin()` with `end()` and fall back to the activating cell | `GridCellSupportEditor::DoActivate` (`GUI_ObjectTable.cpp`) | `controls-dataview.md` |
| 9a053f15eb (#12936) | Since the upgrade, on macOS the Slice/Print split-button's transient popup was dismissed the moment the cursor entered the gap between button and menu; on macOS anchor transient popups flush with (slightly overlapping, 2 px) their button | `SidePopup::Popup` (`__APPLE__` branch) | `popups-menus.md` |
| 1f2ed70288 (#13119) | Orca registers its own `kAEGetURL` handler so `orcaslicer://` links reach `MacOpenURL` | `register_mac_deep_link_handler` (`DeepLinkHandlerMac.mm`), called from `GUI_App::on_init_inner` | this file |
These post-upgrade regressions were fixed: 46e47cec0a (ObjectTable crash on cell deselect),
1f2ed70288 (macOS deep links), d62aa42e61 (WebView script-handler cleanup), eefdabcd98 (maximised
window not filling the desktop), f70d30bf79 (Windows 10 double window frame), c06a0223a7 (blank 3D
canvas during MSW resize). Watch for them to
recur when touching the same code.
**Pitfalls**
- **Rule:** Keep Orca's own `kAEGetURL` registration in `GUI_App::on_init_inner`; do not rely on
wx's handler for `orcaslicer://` links.
**Why:** wx installs its handler in `applicationWillFinishLaunching:`
(`src/osx/cocoa/utils.mm:46-57`); it forwards to `MacOpenURL` only once `OSXInitWasCalled()` is
true and otherwise stores the URL with `OSXStoreOpenURL` (:191-200) [source]. After the upgrade
the wx handler stopped delivering deep links on macOS (#13119, reported on macOS 26): links from
Printables/Thingiverse opened a blank project. `NSAppleEventManager` keeps the last registration, so Orca's later registration wins
and routes straight to `MacOpenURL` → `start_download`. Removing it, or registering before wx does,
brings the regression back.
```cpp
// Right (GUI_App::on_init_inner, __APPLE__): register_mac_deep_link_handler();
```
Cite: 1f2ed70288; `DeepLinkHandlerMac.mm`.
- **Rule:** Do not reintroduce `wxTRANSPARENT_WINDOW` or a post-creation
`SetBackgroundStyle(wxBG_STYLE_TRANSPARENT)` to make a panel blend in; set its background colour,
dark-mode mapped, and re-apply it on theme change.
**Why:** see §3. The side-tool panels lost MSW transparency with the upgrade, and 8248b06337
(b9952b39ad) fixed them by giving them `StateColor::darkModeColorFor(wxColour("#3B4446"))`,
re-applied in `MainFrame::update_side_button_style`.
run:echo "::warning title=Older deps::No cache matches this PR's deps/, so clang-tidy parsed against ${{ steps.newest_cache.outputs.cache-matched-key }}. If it reports errors in code that uses a dependency, rerun clang-tidy once the base branch's Build all has saved the new deps."
# re-sliced on its own to see whether it changes the G-code
# harness - the GUI-vs-CLI parity harness (metrics only, never fails)
# Both test the latest successful build_all.yml Linux AppImage from main, with
# sources checked out at the commit that build was made from. Nothing here
# gates a build or a PR.
# sources checked out at the commit that build was made from; a manual run can
# name another branch, or pin one build by its run id. Nothing here gates a
# build or a PR.
name:Parity Nightly
on:
@@ -20,9 +21,13 @@ on:
required:false
default:"main"
build_branch:
description:"branch whose latest successful build_all artifact to test"
description:"branch whose newest successful build_all artifact to test (a PR build is the PR merged into its base; sources are checked out at the PR head)"
required:false
default:"main"
build_run_id:
description:"build_all run id to test instead of build_branch's newest (same PR caveat)"
# a page of 100 runs spans about a day and a half; a manual run may
# target a branch that last built weeks ago
pages=3
if [ "$SCHEDULED" != true ]; then pages=20; fi
found=""
for page in $(seq "$pages"); do
found=$(gh api "repos/$GH_REPO/actions/workflows/build_all.yml/runs?per_page=100&page=$page" --jq "$pick")
if [ -n "$found" ]; then break; fi
done
[-n"$found"]|| { echo "no successful $BRANCH build among the last $((pages * 100)) build_all runs; pass build_run_id to test an older one" >&2; exit 1; }
fi
read -r run_id head_sha created <<< "$found"
# the nightly fails rather than report on a stale build
- `#pragma once` for headers. Smart pointers and RAII preferred
- Include what you use: include the header for every symbol a file uses, and keep headers compilable on their own. Never rely on the precompiled header or a transitive include. The `clang-tidy` CI job enforces this on changed lines; run the same check locally with `scripts/run_clang_tidy.sh` (`scripts\run_clang_tidy.ps1` on Windows), which sets up everything it needs
- Parallelization via TBB — be mindful of shared state
- Always use `SetSizerAndFit(sizer)` instead of `SetSizer(sizer)` on top level window. Unless `SetSizer` must be called before the full layout is built, call `sizer->SetSizeHints(window)` afterwards in this case.
- Keep code concise and clear. Manually simplify AI generated bloated codes before review.
- Include targeted tests or documented verification for behavior changes, especially in slicing logic, profiles, formats, and GUI defaults.
- For profile changes (`resources/profiles/<Vendor>/**`), check that `version` in the sibling `resources/profiles/<Vendor>.json` was bumped.
- For new or changed UI strings, check they follow [Writing translatable strings in the source](#writing-translatable-strings-in-the-source): no concatenated fragments, plural forms for counts, contexts for ambiguous words.
- For translation changes (`localization/i18n/**/*.po`), check that recurring terms match the [Localization glossary](https://github.com/OrcaSlicer/OrcaSlicer_WIKI/blob/main/developer_reference/localization_glossary.md) for that language.
Catalogs live in `localization/i18n/<lang>/OrcaSlicer_<lang>.po`; the template is `OrcaSlicer.pot`.
See the [Localization guide](https://github.com/OrcaSlicer/OrcaSlicer_WIKI/blob/main/developer_reference/localization_guide.md) for the human-facing version of these principles.
### Writing translatable strings in the source
Translators see one msgid at a time and cannot change the code around it, so each message must be complete in a single msgid. Use the wrapper that fits instead of assembling text from pieces:
- **One msgid per sentence, with placeholders.** Never build a message by concatenating translated fragments (`_L("A") + name + _L(" is B")`) or by splicing a translated word such as `_L("left")` into another sentence: the translator can't reorder or inflect the parts. Write `format(_L("%1% is too close to others"), name)`.
- **A count needs a plural form.** Don't write `file(s)` or leave a verb unchanged after a number. Use `_L_PLURAL("%1% object", "%1% objects", n)`. libslic3r has no plural lookup, so word the message so it doesn't depend on a count, or build it in the GUI.
- **An ambiguous word needs a context.** When a short string means different things in different places (verb or noun, a mode name, a CAD term), give each sense its own context with `_L_CONTEXT`/`_u8L_CONTEXT`/`L_CONTEXT`. Keep the most common sense without a context so its existing translations stay valid.
- **Mark every user-visible string** (`_L`/`_u8L`, or `L` where it is translated later) and list new source files in `localization/i18n/list.txt`: files that aren't listed are never extracted.
- Add a `// TRN` comment above a new string whose meaning or placeholders aren't obvious.
- Never choose behavior by comparing translated text; compare ids or flags.
See `.claude/skills/orca-translations/references/source-audit.md` for the full checklist.
### Terminology
- Use the [Localization glossary](https://github.com/OrcaSlicer/OrcaSlicer_WIKI/blob/main/developer_reference/localization_glossary.md) as the source of truth for recurring terms, so the same English term is always rendered the same way within a language, and terms that must stay in English (brand/product names, acronyms, materials, file formats, G-code tokens, macros/variables/identifiers) are not translated.
@@ -89,6 +106,7 @@ See the [Localization guide](https://github.com/OrcaSlicer/OrcaSlicer_WIKI/blob/
- Plural entries: read `nplurals` from the catalog's `Plural-Forms` header (it is **not** always 2 — ja/ko/zh/th/vi use 1, ru/cs/pl/lt use 3, uk uses 4). Each form must be genuinely inflected for its quantity; repeating one sentence across all forms is a bug in Slavic/Baltic languages, though it is correct for Turkish and Hungarian.
- An entry whose `msgstr` equals its `msgid` is untranslated even though it is not empty; a plural entry with any empty form is likewise incomplete.
- Mark machine-produced translations with an `# AI Translated` translator comment. Don't add it to a human translation you didn't actually rewrite.
- When you can't be sure of a machine translation's meaning or UI wording, also add `# Needs human review: <what to check>`. Never use `fuzzy` for this — fuzzy entries are hidden from users.
- Don't reflow or re-wrap unrelated entries — keep the diff limited to the strings you changed.
set(_orca_runtime_error"CMake ${CMAKE_VERSION} did not find msvcp140.dll and vcruntime140.dll for MSVC ${MSVC_VERSION}. Update CMake to a release that supports this Visual Studio.")
Slice for belt / conveyor (infinite-Z) printers, with belt-aware supports and a tilted-bed preview. Contributed by [Joseph Robertson (@HarrierPigeon)](https://github.com/HarrierPigeon).
Print copies or mirror images of a part on every carriage of an IDEX or IQEX printer at once, with the mode chosen per plate and nozzle clearance zones shown on the bed. Contributed by [Clifford (@cgarwood82)](https://github.com/cgarwood82).
- Additional features can be found in the [change notes](https://github.com/OrcaSlicer/OrcaSlicer/releases/).
# Wiki
@@ -82,24 +86,14 @@ The [wiki](https://www.orcaslicer.com/wiki) aims to provide a detailed explanati
## Stable Release
📥 **[Download the Latest Stable Release](https://github.com/OrcaSlicer/OrcaSlicer/releases/latest)**
Visit our GitHub Releases page for the latest stable version of OrcaSlicer, recommended for most users.
Visit our GitHub Releases page for the latest stable version of OrcaSlicer, recommended for most users.
Windows release binaries are code signed; see our [Code signing policy](#code-signing-policy).
## Nightly Builds
🌙 **[Download the Latest Nightly Build](https://github.com/OrcaSlicer/OrcaSlicer/releases/tag/nightly-builds)**
Explore the latest developments in OrcaSlicer with our nightly builds. Feedback on these versions is highly appreciated.
### Belt Printer Builds
The [nightly release](https://github.com/OrcaSlicer/OrcaSlicer/releases/tag/nightly-builds) ships **two parallel builds**: the standard build and a belt-printer build. Both are attached to the same release — tell them apart by the filename suffix:
- **Standard** — no suffix (e.g. `OrcaSlicer_Windows_Installer_x64_nightly.exe`)
The `_belt` builds add **experimental support for belt / conveyor (infinite-Z) printers**, where the model is sliced against a tilted belt surface instead of a flat horizontal bed. They include ready-to-use belt printer profiles, the full belt slicing pipeline (mesh rotation and G-code transforms), belt-aware support generation, and a tilted-bed preview.
> ⚠️ Belt printer support is under active development and is **not yet merged into `main`** — it currently ships only in these parallel `_belt` builds, produced from the [`belt-printer`](https://github.com/OrcaSlicer/OrcaSlicer/tree/belt-printer) branch. See tracking PR [#14394](https://github.com/OrcaSlicer/OrcaSlicer/pull/14394) and the original documentation in [#12998](https://github.com/OrcaSlicer/OrcaSlicer/pull/12998).
# How to install
## Windows
@@ -251,6 +245,16 @@ OrcaSlicer began in that same spirit, drawing from BambuStudio, PrusaSlicer, and
The OrcaSlicer logo was designed by community member [Justin Levine](https://github.com/jal-co).
# Code Signing Policy
Free code signing provided by [SignPath.io](https://about.signpath.io), certificate by [SignPath Foundation](https://signpath.org).
Windows release binaries are built by GitHub Actions from this repository and signed through the SignPath Foundation program. Every signing request is manually approved.
**Team roles:** [SoftFever](https://github.com/SoftFever) is the committer, reviewer and approver for all signed releases.
**Privacy policy:** See the [OrcaSlicer privacy policy](https://www.orcaslicer.com/privacy-policy/).
# License
- **OrcaSlicer** is licensed under the GNU Affero General Public License, version 3.
echo" -u: install system dependencies (asks for sudo password; build prerequisite)"
echo" -l: use Clang instead of GCC (default: GCC)"
echo" -L [lld|mold]: use an alternate linker (if available) (default: lld)"
echo" -L: use ld.lld as linker (if available)"
echo" -- CMAKE_ARGS...: pass everything after '--' to the cmake configure step of the deps (-d) and OrcaSlicer (-s/-t) builds,"
echo" e.g. -DCMAKE_CXX_FLAGS=\"-O2 -march=native\", -DCMAKE_CXX_COMPILER=g++-14, -DCMAKE_EXE_LINKER_FLAGS=-fuse-ld=mold."
echo" They come last on the cmake command line, so they override what -l and -L set."
echo"For a first use, you want to './${SCRIPT_NAME} -u'"
echo" and then './${SCRIPT_NAME} -dsi'"
echo"For a GitHub Actions-like Linux build locally, use './${SCRIPT_NAME} -g -istrlL'"
echo"Use './${SCRIPT_NAME} -gF -istrlL' to rebuild the cached runner image first."
echo"Set ORCA_CONTAINER_CLI, ORCA_DOCKER_IMAGE, ORCA_DOCKER_BASE_IMAGE, or ORCA_DOCKER_CMAKE_VERSION to override the container runtime, cached image tag, base image, or CMake version."
echo"DEPS_EXTRA_BUILD_ARGS and ORCA_EXTRA_BUILD_ARGS add whitespace-separated cmake configure arguments to the deps and OrcaSlicer builds; prefer '--' for arguments that contain spaces."
}
SLIC3R_PRECOMPILED_HEADERS="ON"
DEPS_PARALLEL=""
unset name
BUILD_DIR=build
BUILD_CONFIG=Release
FORWARDED_ARGS=()
whilegetopts":1j:bcCdDeFghiprstulL" opt ;do
whilegetopts":1j:J:bcCdDeFghiprstulL" opt ;do
case${opt} in
1)
exportCMAKE_BUILD_PARALLEL_LEVEL=1
@@ -51,6 +57,10 @@ while getopts ":1j:bcCdDeFghiprstulL" opt ; do
exportCMAKE_BUILD_PARALLEL_LEVEL=$OPTARG
FORWARDED_ARGS+=("-j""$OPTARG")
;;
J )
DEPS_PARALLEL=$OPTARG
FORWARDED_ARGS+=("-J""$OPTARG")
;;
b )
BUILD_DIR=build-dbg
BUILD_CONFIG=Debug
@@ -115,24 +125,8 @@ while getopts ":1j:bcCdDeFghiprstulL" opt ; do
FORWARDED_ARGS+=("-l")
;;
L )
# -L takes an optional argument. getopts has no native support for
# this, so L is declared bare (no ':') in the optstring above, and we
# manually peek at the next unconsumed argv token via ${!OPTIND}. If
# it's a bare 'lld' or 'mold' (not another option, i.e. doesn't start
# with '-'), consume it as the explicit choice and advance OPTIND so
# getopts doesn't reprocess it as a new flag. Otherwise, leave it
# alone (it isn't meant for -L) and default to lld.
echo"Error: -L mold is not available in the Docker/Podman build image, so -g and -L mold cannot be combined."
echo"Omit -L mold when using -g (the container build defaults to GCC without mold), or drop -g and build with -L mold directly on a host with mold installed."
echo"Building up to ${DEPS_JOBS} dependencies at a time, ${CMAKE_BUILD_PARALLEL_LEVEL} jobs each: up to $(( DEPS_JOBS * CMAKE_BUILD_PARALLEL_LEVEL )) compile jobs at once."
message(AUTHOR_WARNING"Plural form ${tmp} found in config options of ${PREFIX}. This works as before but is now deprecated. Please only use singular forms INCLUDE_DIR and LIBRARY, and update your find scripts for LibFindMacros > 2.0 automatic dependency system (most often you can simply remove the PROCESS variables entirely).")
message(WARNING"The find module for ${PREFIX} does not provide version information, so we'll just assume that it is OK. Please fix the module or remove package version requirements to get rid of this warning.")
set(vars"${vars}You may use CMake GUI, cmake -D or ccmake to modify the values. Delete CMakeCache.txt to discard all values and force full re-detection if necessary.\n")
if(version_unsuitable)
set(msg"${PREFIX} ${${PREFIX}_VERSION} was found but")
if(exactver)
set(msg"${msg} only version ${findver} is acceptable.")
else()
set(msg"${msg} version ${findver} is the minimum requirement.")
endif()
else()
if(missing_headers)
set(msg"We could not find development headers for ${PREFIX}. Do you have the necessary dev package installed?")
elseif(some_files)
set(msg"We only found some files of ${PREFIX}, not all of them. Perhaps your installation is incomplete or maybe we just didn't look in the right place?")
if(findver)
set(msg"${msg} This could also be caused by incompatible version (if it helps, at least ${PREFIX} ${findver} should work).")
endif()
else()
set(msg"We were unable to find package ${PREFIX}.")
endif()
endif()
#FatalerroroutifREQUIRED
if(required)
set(msg"REQUIRED PACKAGE NOT FOUND\n${msg} This package is REQUIRED and you need to install it or adjust CMake configuration in order to continue building ${CMAKE_PROJECT_NAME}.")
message(FATAL_ERROR"${msg}\n${vars}")
endif()
#Otherwisejustprintanastywarning
if(NOTquiet)
message(WARNING"WARNING: MISSING PACKAGE\n${msg} This package is NOT REQUIRED and you may ignore this warning but by doing so you may miss some functionality of ${CMAKE_PROJECT_NAME}. \n${vars}")
set(DEP_DOWNLOAD_DIR${CMAKE_CURRENT_SOURCE_DIR}/DL_CACHECACHEPATH"Path for downloaded source packages.")
set(FLATPAKFALSECACHEBOOL"Toggles various build settings for flatpak, like /usr/local in DESTDIR or not building wxwidgets")
option(SLIC3R_CAD"Build the SolveSpace solver and OCCT ModelingAlgorithms module the parametric Design/CAD tab needs. Must match the main project's SLIC3R_CAD."ON)
The OpenGL Extension Wrangler Library (GLEW) is a cross-platform open-source C/C++ extension loading library. GLEW provides efficient run-time mechanisms for determining which OpenGL extensions are supported on the target platform. OpenGL core and extension functionality is exposed in a single header file. GLEW has been tested on a variety of operating systems, including Windows, Linux, Mac OS X, FreeBSD, Irix, and Solaris.
Loaded 100 of 13007 files, more files were not shown because too many files have changed in this diff.
Show more
Reference in new issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
SoftFever
