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 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>
* 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
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>
* 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
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.
* 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.
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.
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.
* fix: bounds-check the toolchange flush-volume and HRC per-filament lookups
GCode::set_extruder's toolchange flush-volume lookup and
GCodeProcessor::update_slice_warnings's HRC check index per-filament and
per-extruder arrays (flush_volumes_matrix, the filament map, the nozzle list)
by filament/extruder id. When a config leaves one of those arrays shorter than
the filament count (partial or legacy multi-extruder projects, minimal
configs), the reads run off the end: silent on a normal STL, a hard abort under
_GLIBCXX_ASSERTIONS.
Route both reads through bounds checks: the flush lookup falls back to no flush,
matching the existing unknown-old-filament branch beside it, and the HRC check
skips an unmapped filament, mirroring the required_nozzle_HRC guard on the line
above. When the arrays are sized to the filament count the values are unchanged,
so correctly-specified configs are unaffected.
* ci: retrigger checks
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.
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.
build: clear eleven single-site clang-cl warning categories
Each of these is the last site left in its category, and every one is the
compiler saying it cannot tell what the code meant. Nothing here changes
defined behavior.
- OrcaSlicer_app_msvc.cpp printed a DWORD with %d
- StackWalker.cpp ran delete[] through an LPVOID
- ToolOrdering.cpp used a bare ; as a deliberate skip loop's body
- WipeTower.cpp had finish_block_tcr = finish_block_tcr, so the branch that
reached it did nothing. Folding the condition into the enclosing if leaves
the other branch untouched
- GCodeProcessor.cpp had an else binding to the inner if while the outer if
carried no braces
- AmsMappingPopupUpdate.cpp wrote >= 1 || <= 3 where its own comment says &&
- CalibrationWizardPresetPage.cpp left max_decimal_length unset through a
pair of conditions that cover every value but not visibly so
- DevManager.cpp bound map elements to pair<K, V> rather than
pair<const K, V>, copying every one
- SyncAmsInfoDialog.cpp had extraneous parentheses around a comparison
- Http.cpp had if (speed > 0.01) speed = speed;. speed now starts at 0 as
well, because curl_easy_getinfo leaves the target untouched when it fails
and the value reaches Progress either way
- SnapmakerPrinterAgent.cpp truncated npos into an unsigned int, so the
!= npos guard was always true. A colour with no # still yields 0, because
the wrap produced 0 as well
Nine categories go to zero. -Wtautological-overlap-compare and
-Wsometimes-uninitialized reach zero when #15583 merges their second site.
Every edit makes the precedence the compiler already applies explicit. None
of them regroups an expression, so behavior is unchanged at all eight sites.
Strip parentheses and whitespace from the diff and the token stream matches.
GCodeProcessor.cpp:1472 tests == where the symmetric clause below tests !=,
which reads like a typo and is not one. A comment now explains why.
OrcaSlicer.cpp:4760 was the only judgment call. Its leading !is_seq_print is
bare while both operands are parenthesized, so the written form matches what
the compiler does. Kept rather than guessed at.
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
* build: enable /Zc:lambda for MSVC
MSVC keeps its legacy lambda processor under /std:c++17, which rejects
reading a constexpr constant inside a lambda that does not capture it
(C3493). No other compiler requires that capture, and clang reports it as
an unused one, so the two cannot both be satisfied without the flag.
/Zc:lambda selects the conforming lambda parser that clang and GCC
already use. It is implied by /std:c++20 and /permissive-, so it is only
needed while we are on C++17. clang-cl is conforming already and does not
take the flag.
It requires VS2019 16.8, so build_release_vs.bat now says 16.8+.
* build: clear 237 unused lambda capture warnings
236 captures across 81 files, 142 of them `this`. Removing an unused
capture changes no behavior; clang does not report a capture whose type
has a non-trivial destructor, so nothing held only to extend an object's
lifetime is in this set.
Nine of them are the second half of the warning, "is not required to be
captured for this use", where the capture is a const or constexpr value
the body does read. Those depend on the /Zc:lambda change in the previous
commit. One of them, in FillRectilinear.cpp, had been worked around with
an #ifndef __APPLE__ guard around the capture list, which is now gone.
GUI_ObjectTableSettings.cpp captured its reset button only to read it
inside #ifdef __WXOSX_MAC__. That branch now takes the button from the
event it is already handling.
* build: fail configure on MSVC older than 19.28 instead of dropping /Zc:lambda
cl.exe answers an unrecognized /Zc: sub-option with warning D9002 and keeps
going, so on VS2019 before 16.8 the flag is silently ignored and the build
instead dies with C3493 in FillRectilinear.cpp, nowhere near the cause.
* fix: delete three locals that are now unused
Their only remaining use was the lambda capture this branch removed. The
Clang builds set -Wno-unused-variable, so the build never flagged them.
---------
Co-authored-by: Rodrigo Faselli <162915171+RF47@users.noreply.github.com>
* Normalize the junction direction vector over XYZE
calc_vmax_junction_deviation() treats the dot product of two jd_unit_vec as a
cosine, but the vectors were scaled by 1 / block.distance, which is the XYZ
length. On an extruding move the E component then pushes the 4D norm above 1 and
the dot product below -1, so the corner reads as straighter than it is and is
planned too fast -- the more so the higher the flow. Measured on a 6 degree
corner at scv 5: 86.9mm/s with no extrusion, 94.4mm/s at 0.029mm/mm, 150.0mm/s
at 0.1mm/mm.
Neither firmware does that. Marlin normalizes over XYZE for any extruding move
(planner.cpp: `if (... || esteps > 0) normalize_junction_vector(unit_vec)`) and
Klipper leaves E out of the cosine entirely, dotting only axes_r[0..2]
(toolhead.py::Move.calc_junction). Normalizing satisfies both: with E normalized
in, the cosine differs from the XYZ-only one by ~1e-5 at printing flow rates.
This is a deliberate divergence from PrusaSlicer, which still scales by
1 / distance -- it carries an older Marlin's behaviour.
Travel moves are unaffected, their vector was already unit length.
Reported by Copilot in review of #15304.
* Test that extrusion rate does not change corner planning
The junction deviation tests were all travel-only, which is exactly why the E
component of the junction vector went unchecked. Cover it: the same corner has
to be planned the same whether nothing, an ordinary 0.42 x 0.2 line, or a fat
large-nozzle line is extruded through it, on both Klipper and Marlin 2.
Reported by Copilot in review of #15304.
* Plan corners with junction deviation where the firmware uses it
The time estimator only ever had the classic per-axis jerk model, which limits a
corner by the largest single-axis component of the velocity change. That is
anisotropic: the same corner is allowed sqrt(2) more speed on a diagonal than on
an axis, which paints a four-lobed ripple around every circular wall in the
actual speed and actual flow views, worst on small parts whose walls are made of
short segments.
Klipper has no classic jerk at all and Marlin 2 has none while M205 J is in use;
both plan corners with junction deviation, which sees only the corner angle. Add
that model and use it for those machines:
- Klipper: derived from the square corner velocity, as the firmware does
(jd = scv^2 * (sqrt(2) - 1) / max_accel), reading the scv from
machine_max_jerk_x, where process_SET_VELOCITY_LIMIT() already stores
SQUARE_CORNER_VELOCITY.
- Marlin 2: machine_max_junction_deviation, which was already loaded into the
machine limits but never reached the planner.
- Every other flavor keeps the classic jerk path unchanged.
The model has no per-axis jerk floor, so this also drops the hard slow spot the
estimator drew at the start of every loop from machine_max_jerk_e.
Toolpaths are unaffected: on a full export the only lines that change are M73.
The junction deviation maths, including Marlin's JD_HANDLE_SMALL_SEGMENTS arc
approximation, is ported from PrusaSlicer's src/libslic3r/GCode/GCodeProcessor.cpp.
The Klipper mapping is not in PrusaSlicer, which ignores SET_VELOCITY_LIMIT.
* Add tests for junction deviation corner planning
Cover the three properties the change rests on:
- a right angle on Klipper is planned at exactly the square corner velocity,
the identity that makes the scv to junction deviation mapping correct, and a
shallow corner is planned far faster than per-axis jerk allows;
- junction deviation gives the same speed whatever the corner's orientation,
while classic jerk keeps its sqrt(2) spread, which is the four-lobed ripple;
- machines that do not plan with junction deviation are provably untouched,
including a Marlin 2 printer that has it disabled.
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.
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.
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).
- WipeTower: use filament_ramming_volumetric_speed(_nc) for ramming, falling back to
max_vol_speed only when nil; gate precool temps on enable_pre_heating
- ToolOrderUtils: disable the inter-layer forecast in the per-nozzle base reorder so
H2D/H2C ordering is unchanged
- PrintConfig: stop stripping filament_prime_volume in handle_legacy; document that
prime_volume drives the Type2 wipe tower and filament_prime_volume the Type1 one
- GCodeProcessor: exclude post-print end-gcode M400 dwells from the M73 estimate and
drop the dead air-filtration state
- Trim verbose BambuStudio source-location comments across the port
End gcode contains firmware-conditional M400 waits for air purification,
timelapse capture, and sound notification that are post-print operations.
These were incorrectly included in M73 total time, inflating the estimate.
The fix detects MACHINE_END_GCODE_START tag during the streaming parse
(process_tags) and sets m_skip_end_gcode_delays=true. process_M400 then
skips timed delays (S/P params) in the end gcode scope.
BBS achieves the same effect by dropping leftover in calculate_time
(is_final=true). We skip at the source instead, which is more surgical
and leaves calculate_time behavior unchanged for all printers.
Affects all BBL printers with MACHINE_END_GCODE_START tag.
Non-BBL printers are unaffected (no tag = no skip).
The time estimator's speed/acceleration limits were indexed by time
mode only, reading slot 0 of the per-(extruder x volume-type) arrays
the multi-extruder profiles already carry (H2C 0.4: 8 entries, H2D
0.4: 10). Every move was therefore modelled with the first machine
slot's limits regardless of which nozzle variant was printing -
estimation fidelity only, since emitted feedrates/accelerations are
decided on the slicing side.
Now the estimator resolves the machine slot of the nozzle currently
mounted in the active extruder: the nozzle grouping context is handed
to the processor BEFORE the streaming replay (new member + setter -
deliberately separate from the post-stream result-field handover that
gates the richer change-time model, whose timing is unchanged), the
occupancy recorder is populated on every filament change (bookkeeping
decoupled from the gated time model; recorder writes have no time
effect), and get_machine_config_idx maps (volume type x extruder type
x extruder) to the slot via the printer's variant layout, newly
carried on the processor result. The feedrate/acceleration getters
gain a slot parameter indexing [slot*2 + mode]; jerk and the
print/travel/retract accelerations stay mode-only. Reloaded sliced
projects re-estimate with the result's saved grouping context;
imported bare g-code degrades to slot 0 - the historical read.
M201/M203 write the parsed value into EVERY slot's mode entry (a
firmware envelope change is global), which keeps per-slot reads in
lockstep with the mode-only reads they replace: the fleet emits
envelope lines before any motion, so estimates - hence the estimated
time header, M73 lines, and every other byte - are unchanged (20/20
pinned-slice byte gate bit-identical, incl. the sequential repro
sliced twice). Fidelity improves where envelope emission is off or a
migrating per-layer plan moves filaments across variants.
Tests: a stub-driven processor case proving the slot follows the
active nozzle through the exact production path (T..H.. commands,
fallback recorder bookkeeping, 4x time ratio on the slow variant),
that emitted M201/M203 reach every slot, and that a missing context
degrades to slot 0. Suites green (libslic3r 48998/169, fff_print
667/62).
When a per-layer nozzle grouping migrates a filament across nozzle
variants, the write-back turns two groups of config arrays from
filament-indexed into column-indexed: the per-variant filament options
(one column per variant a filament uses) and the merged extruder
retract overrides (resized to the column count by apply_override).
Export-path readers that still indexed them with the raw filament id
read a neighbor's column for every filament ordered after a migrating
one: toolchange/standby temperatures (M104/M109), retraction lengths
and feedrates, wipe distance, z-hop types, air-filtration keys, and -
through the Extruder's cached flow term - the extrusion E of every
move.
Now every such read resolves its column through the existing
layer-aware resolver (get_filament_config_index ->
Print::get_filament_config_indx), which returns the raw filament id
whenever no per-layer grouping result is published, so static prints
are byte-inert by construction. The Extruder itself has no layer
knowledge, so it gains an injected config column (set_config_index,
default = filament id) that the generator refreshes at the only two
resolution-changing events - layer change and writer toolchange - and
that re-syncs the cached e_per_mm3 flow term. Old-filament reads
resolve at the current layer, which is safe because the per-layer maps
are gap-filled carry-forward. Whole-array placeholder copies
(toolchange temperature overrides) are rebuilt in filament order,
mirroring the existing per-variant placeholder remap. The resolvers
move to the public section so non-friend helpers (ooze prevention) can
resolve too.
Documented, deliberately unchanged: the wipe tower's per-filament
parameter rows (no layer dimension; tower x per-layer grouping is a
follow-up), travel_slope's physical-extruder read, estimator pre-heat
bookkeeping temps, and index-0 header diagnostics.
Verification: new Extruder column-injection scenario (defaults, column
follow + flow-cache rescale, filament-indexed reads unaffected, reset
semantics) and a migrating write-back case proving the column shift for
filaments ordered after a migrator and the resolver tracking it (11 +
14 assertions); suites green (libslic3r 48998/169, fff_print 655/61);
20/20 pinned-slice byte gate bit-identical (incl. sequential repro x2
deterministic).
- the g-code writer tracks the current layer id and resolves
FILAMENT_CONFIG/NOZZLE_CONFIG (plus every non-macro variant lookup,
toolchange placeholder scalars, and the change-filament flush
overrides) through Print's per-filament, per-layer config-index
resolvers instead of the filament->extruder collapse
- update_layer_related_config refreshes the per-layer
extruder/volume/nozzle maps in the writer config;
update_placeholder_parser_with_variant_params remaps the
filament-variant arrays into filament-id space for custom g-code
(Orca's flush placeholder computation moves inside it)
- the engine's concrete per-filament volume assignment now merges into
the config write-back (the temporary hold from the producer commit
is lifted together with these consumers), and the background process
reads the computed volume map back to the plate
- append_full_config dumps the resolved filament_map_2 slots
- update_used_filament_values gains a bounds guard
- tests: per-filament Hybrid slot resolution + null-result fallback
Result: on a Hybrid extruder, each filament's features slice with its
assigned sub-nozzle's variant values (speeds, volumetric limits,
retraction). Verified on a 4-filament H2C Hybrid project: outer walls
split into three feedrate populations (30/50/200 mm/s), toolpath
geometry byte-identical, deterministic across repeated slices. All 18
non-Hybrid reference fixtures stay byte-identical except the
filament_map_2 header value now showing the real slot. Auto grouping
ties (multiple zero-flush perfect matchings) may pick a different
filament-to-nozzle isolation than other slicers; verified co-optimal.
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.
perf(GCodeProcessor): stop recompiling std::regex on every g-code line
process_SET_VELOCITY_LIMIT() constructed three std::regex objects from
scratch on every call, and Klipper-flavor g-code contains
SET_VELOCITY_LIMIT on a large share of lines (8,834 of 103,549 lines for
a single 3DBenchy sliced for a Creality K2). perf attributes 6.4% of the
whole slicing run to this one function, almost all of it regex
compilation and the allocator traffic it generates.
process_SET_PRESSURE_ADVANCE() and the External_Purge_Tag handler had
the same per-call construction.
Hoist all five patterns to function-local static const std::regex so
they compile once. Generated g-code is byte-identical (modulo the
timestamp header); slicing a 16x Benchy plate for a K2 drops from
78.5s to 27.3s wall (2.9x) on a 16-core Linux box, single Benchy from
8.9s to 5.6s.
Co-authored-by: grant0013 <grant@harktech.co.uk>
Joseph Robertson