# 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>
* CLI: --inspect-paint — dump per-facet paint state as JSON
Reads the per-facet enforcer/blocker/extruder/fuzzy-skin state stored
on every ModelVolume (supported_facets / seam_facets /
mmu_segmentation_facets / fuzzy_skin_facets) and emits a structured
JSON summary to stdout. Machine-readable alternative to opening the
paint gizmos.
Per (object, volume, layer, state): facet count, surface area in
mm², and mesh-local bounding box. Empty layers collapse to
{"empty": true}. Summary at the top level rolls up totals.
One correctness detail worth calling out: FacetsAnnotation::
get_facets_strict returns an indexed_triangle_set whose `vertices`
array is the whole source mesh — only `indices` are filtered to the
painted triangles. A naive bounding_box(its) would report the whole
mesh's bbox even when only a few facets are painted. The helper
its_referenced_bbox() walks only the vertices actually indexed by
the painted triangles, so `bbox` correctly localizes the painted
region.
Rationale: every paint-driven workflow — GUI-painted .3mf verified
in CI, AI agents planning support enforcers, MMU color layout checks
— needs to know what's already painted on a model. Today that's a
GUI-only read. --inspect-paint closes that loop for scripted callers.
New file src/slic3r/Utils/PaintCLI.{hpp,cpp} (~215 lines). Depends
only on Model, TriangleMesh, TriangleSelector, FacetsAnnotation, and
nlohmann::json — all already in tree. No new dependencies, no
signature changes, no behavior change when the flag is absent.
Registered as an action (parallel to --info) so it satisfies the
"needs an action" check and bypasses the GUI fallback; control falls
through the normal post-action path to a clean exit 0.
Verification:
unpainted STL: every layer {"empty": true}, summary zero
GUI-painted .3mf: enforcer count / area / bbox match painter
clean JSON: parseable via jq
* CLI --inspect-paint: exit after printing, reject conflicting actions
- Finish like the end of CLI::run once the JSON is written, as the
tooltip says. The callback manager is Linux-only, so its use is
guarded.
- Reject actions that would otherwise be skipped without notice
(--slice, --export-3mf, ...) before loading. Load-time options such as
--uptodate are still accepted.
- Replace invalid UTF-8 in object names and paths instead of throwing.
- Report every input file as sources; inputs are merged into one model
before actions run.
* CLI --inspect-paint: reject a run without input
Without an input file or --load-assemble-list there is nothing to
inspect, and the run printed nothing and exited 0. Reject it up front
with CLI_INVALID_PARAMS, next to the other invalid-parameter checks.
The deps superbuild passes the Visual Studio generator and platform to
every sub-build but not the toolset, so build_win.bat -d -l without -x
compiled every dependency with cl even though the superbuild had been
configured with -T ClangCL; CMake replaces the forwarded
CMAKE_<LANG>_COMPILER with whatever the toolset ran. The recipes that
adapt to clang-cl then disagreed with what had been built, and
wxInspector told FindwxWidgets to look in lib/clang_x64_lib while the
cl-built wxWidgets had installed into lib/vc_x64_lib:
Could NOT find wxWidgets (missing: wxWidgets_LIBRARIES
wxWidgets_INCLUDE_DIRS core base aui propgrid)
Forward CMAKE_GENERATOR_TOOLSET as well, so the dependencies compile
with clang-cl under MSBuild the way they already do under Ninja. Four
of them need more than that:
- OpenSSL always builds with cl, and MSBuild runs its nmake steps in
the project's toolset environment, where ClangCL puts clang's include
directory first and cl trips over clang's stdint.h. The project gets
the default toolset.
- Boost.Container's dlmalloc needs -Wno-incompatible-pointer-types
under clang. boost_container links as C++, and the Visual Studio
generator writes only the link language's flags into the project, so
its C file never saw CMAKE_C_FLAGS. Under that generator the option
goes through the C++ flags as well, with the defaults kept.
- Draco's tools and NLopt's testopt compile sources their own static
library also contains. MSBuild lists libraries before objects and
lld-link resolves archive members as each input arrives, so the
library's copy is pulled in before the executable's own object and
the link fails on duplicate symbols; link.exe defers the search and
Ninja lists the objects first. Nothing uses those executables, so
they get /FORCE:MULTIPLE there.
The Ninja path is unchanged: the generated configure commands of all
29 dependencies are identical before and after. OCCT's arm64 override
to cl still applies under Ninja but not under the Visual Studio
generator, where the toolset wins; that combination never built and is
left for a follow-up.
* Toolchange Cyclic Order
* Apply cyclic order to first layer
* Unit test
* Copilot fixes
---------
Co-authored-by: Rodrigo Faselli <162915171+RF47@users.noreply.github.com>
* ci: build Windows with build_win.bat and drop the old scripts
The deps and slicer jobs called build_release_vs.bat; they now call
build_win.bat. --deps-dir and --build-dir name the build/build-arm64
directories the cache keys and later steps already use, and the
script's own VsDevCmd call replaces the Enter-VsDevShell blocks.
With both stages configured each way into the same directory, the deps
superbuild is byte-identical and the slicer build files are
byte-identical apart from CMakeCache.txt recording how DEP_BUILD_DIR
was set.
Two changes beyond the script swap:
- The compiler is the clang-cl bundled with Visual Studio, the script's
default. The old script's bare "clang-cl" resolved to the LLVM on the
runner image's PATH, 20.1.8 on x64 and 22.1.8 on arm64; both arches
now build with the 22.1.3 VS 18.9 ships. Cached dependencies are only
rebuilt when deps/ changes, so they stay on the LLVM they were built
with; the arm64 leg already links deps built with Clang 19 into a
Clang 22 slicer.
- The deps job no longer zips the dependencies afterwards. The zip was
never uploaded and was not in the cached path.
A failed cmake --build now fails the job. The old script returned 0, so
the arm64 failure fixed in #15719 was reported as success and the
half-built dependencies were saved to the cache.
build_release.bat, build_release_vs.bat and build_release_vs2022.bat are
removed; nothing referenced them any more.
* ci: run Build all when the Windows build script changes; tests doc builds the deps
The push filter of build_all.yml never listed a build script, and CI
now depends on build_win.bat, so it and its test suite join the list
the pull_request filter already has.
tests/AGENTS.md told Windows to run build_win.bat --run-tests, which
only implies -s and stops at the dependency check on a clean checkout.
The old build_release_vs.bat tests built the dependencies first, so
the line now says -ds --run-tests.
* Run every profile maintenance job from one tool
orca_id_tool.py becomes orca_profile_tool.py, and orca_extra_profile_check.py
and orca_filament_lib.py fold into it as subcommands: check, generate-id, fix,
trim, update-index and update-snapshot. The three scripts already overlapped --
the checker imported half of its rules from the id tool, which in turn kept a
copy-pasted set of output helpers to avoid the resulting import cycle -- while
disagreeing on how a vendor is enumerated, how a JSON file is read and what the
exit code means. One file settles all three.
check, normalize, trim and update-index reproduce their predecessors exactly; normalize and
update-index were diffed byte-for-byte against the old scripts over a copy of
the whole tree. Deliberate changes: the compatible-printers check no longer
switches itself off when --check-materials is passed, an error exits 1 rather
than -1, update-index honours --profile-type and reports a profile it cannot
place instead of dropping it from the index, fix and update-index gained
--dry-run, trim keeps an unindexed file some surviving profile still inherits
from, and vendors are enumerated as directories with an index -- which is why
blacklist.json, a data file that an unscoped index rebuild once wrote four empty
list sections into, loses them here and will not collect them again. The dead
rename_filament_system() helper is gone.
The suite under scripts/tests now covers the maintenance commands too, and CI
runs it; nothing there ran in CI before. No shipped profile data changes apart
from those four keys.
* update vendor index files with "python3 ./scripts/orca_profile_tool.py update-index" and "python3 ./scripts/orca_profile_tool.py normalize"
* CLI: --ground-face-* / --lay-flat / --center-on-bed orientation primitives
Adds the CLI counterparts to the GUI's lay-flat / face-pick gizmos.
Scripted / CI / AI pipelines can now set orientation without rendering
a wxWidgets frame; today the only way is a GUI round-trip.
New CLI actions (all operate in the mesh-local frame so they compose
with prior --rotate-* / --orient flags):
--ground-largest-face 1 Auto-detect the largest planar-face
or --lay-flat 1 cluster (area-weighted), rotate so its
normal points -Z. Covers "this part has
one obvious flat side" cases.
--ground-face-normal NX,NY,NZ Pick the face whose mesh-local
normal best matches the given
vector; ground it. e.g.
`--ground-face-normal 1,0,0`
stands a part on its +X side.
--ground-face-point X,Y,Z Find the triangle containing the
given mesh-local point; ground its
face. Disambiguates when several
faces share a normal (largest
containing triangle wins).
--center-on-bed 1 Translate so the XY bounding-box
centroid lands at the bed center
(derived from printable_area).
New file `src/slic3r/Utils/MeshOrient.{hpp,cpp}`:
- collect_triangles_object / compute_face_clusters — quantize
per-triangle normals (0.001, ~0.06°) and area-weighted-average
within clusters. Same clustering logic used by lay-flat.
- apply_ground_rotation — same math as Selection::flattening_rotate
in the GUI (Selection.cpp:1432): world-space quaternion from the
transformed normal to -Z, applied as offset * new_rot * old_no_offset
on every instance of every object, then a per-instance Z-lift so the
grounded face lands at exactly 0 (avoids "No layers were detected"
from FP-error z≈-1e-9).
- ground_face_point uses a top-N cluster search + point-in-triangle
test in local space; largest-area triangle wins on ambiguity.
Rationale: without these, any CLI pipeline that needs a specific
face on the bed must either encode custom rotation math per part or
break out of the pipeline into the GUI. Both are bad for
reproducibility. The --ground-face-* triple + the largest-face
auto-mode cover essentially every orientation intent expressible
in a slicing wizard.
Scope:
- `src/slic3r/Utils/MeshOrient.{hpp,cpp}` — new, ~420 lines
- `src/slic3r/CMakeLists.txt` — 2-line registration
- `src/libslic3r/PrintConfig.cpp` — 5 new CLIMiscConfigDef entries
- `src/OrcaSlicer.cpp` — 58-line handler block + 1 include
No behaviour change when the flags are absent.
(cherry picked from commit c45a9795e1)
* CLI grounding: choose among the Lay on Face planes, per object
Addresses review:
- Move the geometry of GLGizmoFlatten::update_planes() into
libslic3r/LayOnFace and use it from the gizmo and the CLI, so the
--ground-* options pick convex-hull faces per object and instance,
with part transformations (--rotate-x/y) applied.
- Drop --center-on-bed, the --lay-flat alias and MeshOrient; make
--ground-largest-face a coBool.
- Parse --ground-face-normal and --ground-face-point strictly. A point
that only some objects contain grounds those and leaves the others.
- Fold in --inspect-mesh from #14603, reporting the same planes.
- Tests in tests/libslic3r/test_lay_on_face.cpp: bounding boxes before
and after, rotate then ground, two objects, and a ribbed part whose
parallel inner faces outsum its base.
* CLI --inspect-mesh, --ground-face-*: reject missing input and empty values
- Without an input file or --load-assemble-list, --inspect-mesh printed
nothing and exited 0. Reject it up front with CLI_INVALID_PARAMS.
- An explicit empty --ground-face-normal or --ground-face-point was
silently ignored. Only options given on the command line reach the
transforms loop, so an empty value now fails the strict parse like any
other malformed value.
# Description
Add `--strict` for CI and scripted pipelines, and a structured
`warnings`
array in `result.json`.
## `--strict`
A NON_CRITICAL slicing warning is logged and the slice succeeds: return
code
`0`, G-code written. That suits interactive use, but a pipeline then
ships a
slice with a warning nobody saw. With `--strict`, such a warning fails
the run
with `CLI_SLICING_ERROR` before the G-code is exported. Without the
flag,
nothing changes.
In FFF the warning that reaches this path is "support needed but
disabled"
(`PrintObject::generate_support_material`). `--no-check` skips that
check, so
`--strict --no-check` is rejected with `CLI_INVALID_PARAMS`.
`--strict` is read before any work, so it doesn't depend on argument
order and
`result.json` reports it for early failures as well.
## `result.json`
Two new top-level fields:
- `warnings`: `[{"class", ...details}]`. One class is wired:
`slicing_warning_non_critical` with `plate_id` and `text`, recorded
whenever
such a warning fires, with or without `--strict`. The array also fills
on
runs that succeed, so `return_code` stays the verdict.
- `strict_mode`: whether `--strict` was on.
`record_exit_reson` writes `result.json` on Linux only, so both fields
exist
only there. The non-zero exit works on every platform.
## Tests
- `tests/fff_print/test_support_material.cpp` (all platforms): an
overhang
sliced with support off raises the NON_CRITICAL support-needed status,
and
the no-check flag suppresses it.
- `tests/cli/test_cli_strict.sh` (Linux only): runs `orca-slicer`
without
flags, with `--strict`, and with `--strict --no-check`, and checks the
shell
status and `result.json` of each. It runs the built binary, so it
carries the
`RequiresApp` label, which `scripts/run_unit_tests.sh` excludes because
the
unit-test job only receives `build/tests`. Run it with
`ctest --test-dir build/tests -C Release -L RequiresApp`.
- CI: `unit_tests.yml` now passes `Release` on Linux too.
`build_linux.sh`
configures Ninja Multi-Config, and without a config ctest drops the
labels of
plain `add_test()` tests, so this test ran as "Not Run" instead of being
excluded. The docs that assumed Linux was single-config are corrected
too.
Built and run locally on Linux (GCC 14) on current `main`: both tests
pass,
and the touched files compile clean under Clang with `-Werror`.
* ci(flatpak): run the unit suite in a separate job, mirroring the other arches
Alternative to the in-job step: split build and test like the Linux/Windows/
macOS legs. The flatpak build now builds the test binaries in-sandbox (the
action's run-tests fires the module's build-only test-commands), prunes the
kept build tree to the test binaries + CTest metadata + data, and uploads it
with /app as a test asset (size reported to the run summary).
A new unit_tests_flatpak matrix job downloads that asset on a native runner,
restores the module-build symlink, and runs the suite via flatpak-builder
--run (which bind-mounts /run/build so TEST_DATA_DIR resolves) against the
GNOME SDK's bounds-checked STL. Results feed publish_test_results.
Costs a per-arch asset upload/download + a runtime install on the test
runner; the trade-off vs the in-job step is a genuine separate graph box.
* ci(flatpak): run tests via `flatpak build` to avoid rofiles-fuse
`flatpak-builder --run` sets up a rofiles-fuse overlay that this CI container
rejects (Failure spawning rofiles-fuse, exit_status: 256), even in a fresh job
with a machine-id and the runtime installed, and --disable-rofiles-fuse is not
accepted in --run mode. `flatpak build` enters the sandbox via bwrap directly,
so it sidesteps rofiles-fuse; bind-mounting the build tree at /run/build gives
the same path the compiled-in TEST_DATA_DIR expects.
* ci(flatpak): slim the test asset (strip binaries, drop source tree)
The first cut shipped ~1 GB: the test exes carried debug info (the SDK builds
with -g and only the app gets stripped) and the packaged module dir included
the whole copied source tree the tests never read at runtime. Strip the test
binaries and keep only build_flatpak/tests, tests/ (TEST_DATA_DIR) and scripts/.
The irreducible remainder is /app, which the exes link against.
* ci(flatpak): extract the test run into a reusable unit_tests_flatpak workflow
Move the flatpak test job out of build_all.yml into a reusable
unit_tests_flatpak.yml, called once per arch (Flatpak x86_64 / aarch64) the
same way the other arches call unit_tests.yml. build_all.yml keeps only the
build + asset packaging; the reusable workflow downloads the asset, runs the
suite via `flatpak build`, and uploads results as test-results-<artifact> for
publish_test_results. Drops the now-unused manifest checkout (flatpak build
does not need it).
* ci(flatpak): trim comments to the non-obvious
No behavior change.
* ci(flatpak): drop redundant caller comment
* ci(flatpak): drop redundant trim comment
* ci(flatpak): drop size-report scaffolding and redundant if-guards
* ci(flatpak): force the app module to rebuild so the test asset always exists
flatpak-builder caches modules by content hash and skips a hit, producing no
build tree and no test asset, so a re-run of the same commit would leave the
separate test job with nothing to download. Inject a per-run cache-buster into
the OrcaSlicer module's build-options (part of its cache key) so it always
rebuilds, mirroring how the other arches cache only deps and always rebuild the
app and tests. The deps modules stay cached.
* ci(flatpak): trim cache-buster comment, fix stale step name
* fix: guard H2C per-filament array reads against short config arrays
The H2C tool-ordering, wipe-tower, and g-code export paths index per-filament
config arrays by filament/tool id. A config with fewer entries than the filament
count (partial or legacy projects, minimal test configs) makes these reads run
past the end of the vector: silent under a normal STL, but UB that aborts under
the flatpak build's bounds-checked STL (_GLIBCXX_ASSERTIONS).
Route the reads through the existing clamping accessors (get_at,
get_filament_category, is_in_same_extruder) and add a small clamp helper for
filament_change_length. The guards are no-ops when the arrays are sized to the
filament count, so correctly specified configs are unaffected.
* ci(flatpak): build filament_group_tests too
The suite landed on main after this branch was cut and arrived via a later merge,
so it was missing from the target list and ctest failed the leg with
filament_group_tests_NOT_BUILT.
Not tests/all, which build_linux.sh uses: that is a Ninja subdirectory target and
this build configures with the default Makefile generator, where it does not exist.
* ci(flatpak): give the embedded-interpreter tests a valid Python home
python_test_support.hpp sets PyConfig.home to <testdir>/python when that
path resolves. WIN32/APPLE populate it with a copied bundled runtime; the
flatpak leg had no such branch, so home resolved to a directory with no
stdlib and all 21 embedded plugin tests failed at "failed to get the
Python codec of the filesystem encoding".
Symlink <testdir>/python to the bundled /app/libpython that already ships
in the flatpak (the test exe links libpython3.12.so from there via rpath),
so the interpreter initializes without duplicating the runtime.
* ci(flatpak): sync the ToolOrdering guard mirror with #14789
Match #14709's build_filament_group_context guard to the version on
#14789 (size filament_info to filament_nums, truncate filament_ids)
so the folded guard is a byte-identical mirror that drops cleanly when
#14789 merges, instead of leaving a stale hunk that conflicts on rebase.
* fix: guard WipeTower per-filament array reads against short config arrays
The BambuStudio WipeTower sync reintroduced raw per-filament array
indexing that reads out of bounds when a config leaves an array shorter
than the filament count: m_physical_extruder_map in format_line_M104/M109
(indexed even when empty), and m_filament_categories in get_wall_skip_points
and get_wall_filament_for_all_layer. Silent on a normal STL, a hard abort
under the bounds-checked STL the Flatpak build uses.
Bounds-check the physical extruder map before indexing (omitting the T
token, as the existing -1 path already does), and route the two raw
m_filament_categories reads through the clamping get_filament_category()
accessor the surrounding code already uses. No change for correctly-sized
configs.
* fix: default-initialize WallToolPathsParams fields
min_length_factor and is_top_or_bottom_layer had no default initializers, and the FillConcentric/FillConcentricInternal callers never set them, so WallToolPaths::removeSmallLines() thresholded on stack garbage. Which short extrusion lines it dropped then depended on memory layout, so concentric solid-infill output was nondeterministic between runs and across machines. Give every member a default, matching the adjacent FillParams. The perimeter path was already fine because it builds the struct via make_paths_params().
* 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
* ci: name the flatpak rebuild token after the cache it defeats
Since #15650 the Flatpak job also has a compiler cache, so a bare
"cache-buster" no longer says which cache is meant. Call it
flatpak_builder_cache_buster, and name the build-dir trim step after
the flatpak-builder cache save it keeps lean.
* ci: ship resources/profiles and resources/printers in the flatpak test asset
Two slic3rutils tests added in 4aa0e1d60b read
resources/printers/bambu_filament_ids.json through PROFILES_DIR/.., and
the asset dropped resources/ entirely, so both failed parsing an empty
stream on each Flatpak leg. Keep the two subtrees the tests reach;
test_gcodewriter's shipped-profile case stops skipping on this leg too.
* ci: restore the CRLF line endings of build_all.yml
The last merge from upstream/main rewrote the file with LF endings, which
turns the 60-line change into a whole-file diff on GitHub. Upstream has had
this file as CRLF since it was created, so put it back.
* ci: trigger Build all on changes to the unit-test workflows
The path filters only matched build_*.yml, so an edit to unit_tests.yml or
unit_tests_flatpak.yml could merge without ever running.
* ci: put a timeout on the flatpak unit-test step
Matches the 20 minutes of the regular unit-test workflow; without it a hung
test holds the runner for the six-hour job default.
VS 2026's ARM64 code generator needs about 27 GB for
_PyUnicode_ToNumeric, a switch with 1951 cases in
Objects/unicodetype_db.h; the same file takes under 1 GB on x64. The
16 GB CI runner has an 18.9 GB commit limit and only gets through when
Windows grows the pagefile on the temp disk in time, so cold arm64
dependency builds fail at random with C1002 "compiler is out of heap
space". build_release_vs.bat returns 0 on failure, so the job still
reports success and the incomplete dependencies are cached.
A property sheet compiles that one file with optimisation off on arm64;
the rest stays whole-program optimised and x64 is unchanged.
MSBuild reads it from PCbuild/msbuild.rsp, which is now written at
configure time and copied in, so a checkout path with spaces works too.
* 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
* fix: guard H2C per-filament array reads against short config arrays
The H2C tool-ordering, wipe-tower, and g-code export paths index per-filament
config arrays by filament/tool id. A config with fewer entries than the filament
count (partial or legacy projects, minimal test configs) makes these reads run
past the end of the vector: silent under a normal STL, but UB that aborts under
the flatpak build's bounds-checked STL (_GLIBCXX_ASSERTIONS).
Route the reads through the existing clamping accessors (get_at,
get_filament_category, is_in_same_extruder) and add a small clamp helper for
filament_change_length. The guards are no-ops when the arrays are sized to the
filament count, so correctly specified configs are unaffected.
* fix: size the grouping context's filament_info to the filament count
build_filament_group_context built model_info.filament_info by walking
filament_type, so a config whose filament_type is shorter than the filament
count produced a short vector. FilamentGroup indexes filament_info by filament
id, so clamping the individual reads only moved the out-of-bounds access
downstream. Loop to filament_nums and read all three fields through get_at,
and drop filament_ids entries past the filament count, since the grouping code
pairs filament_ids and filament_info by position.
Adds a regression test with four filaments and one-entry filament_type /
filament_is_support. Without the fix it throws bad_alloc from copying a garbage
std::string read past the end.
* fix: guard the carousel nozzle-change length reads too
The carousel branch added in b90ac13d86/b0dddb4648 reads
m_filaments_change_length by tool id without a bounds check, the same
pattern this branch already routed through filament_change_length_at
a few lines above in both plan_toolchange and plan_tower_new.
* fix: guard WipeTower per-filament array reads against short config arrays
The BambuStudio WipeTower sync reintroduced raw per-filament array
indexing that reads out of bounds when a config leaves an array shorter
than the filament count: m_physical_extruder_map in format_line_M104/M109
(indexed even when empty), and m_filament_categories in get_wall_skip_points
and get_wall_filament_for_all_layer. Silent on a normal STL, a hard abort
under the bounds-checked STL the Flatpak build uses.
Bounds-check the physical extruder map before indexing (omitting the T
token, as the existing -1 path already does), and route the two raw
m_filament_categories reads through the clamping get_filament_category()
accessor the surrounding code already uses. No change for correctly-sized
configs.
# 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?
-->
Adds a nightly workflow that runs the long parity checks from
[orca-test-repo](https://github.com/OrcaSlicer/orca-test-repo), which
are too slow for the per-build "Run external slicer regression tests"
step and are kept out of every PR and merge build.
## What it runs
`.github/workflows/parity_nightly.yml`, three jobs:
| Job | What it does |
|---|---|
| Find the build to test | Picks the latest successful `build_all.yml`
run for the branch (`main` by default) and records its commit. |
| Override sweep effect stage (shard 0 and 1) | Runs orca-test-repo's
override sweep with `--effect-full`: every config option that lands on
the CLI is re-sliced on its own to check that it actually changes the
G-code. Split into 2 shards, each with a 60-minute timeout. |
| GUI-vs-CLI parity harness | Slices a set of fixtures in the GUI
(headless under Xvfb) and on the CLI, compares the exports, and scores
divergences against a known-differences ledger. It reports only and
never fails on a divergence. |
## When it runs
- **Every night at 21:00 UTC,** after `build_all.yml`'s 17:00 UTC run
has finished.
- **By hand** through `workflow_dispatch`, with optional inputs:
- `build_branch`: the branch whose latest successful build to test;
- `test_repo_ref`: the orca-test-repo ref;
- `fixtures`: a subset of harness fixtures;
- `cli_presets`: `flat` or `raw`.
- **No `push` or `pull_request` trigger,** so nothing here runs on PRs
or merges. The per-build CI step is unchanged.
## How it tests a build
- **Binary:** the Linux x86_64 AppImage from the chosen `build_all` run.
- **Source:** OrcaSlicer checked out at that run's exact commit. The
AppImage only ships packed preset caches, so profiles and the CLI option
list come from this checkout, matched to the binary.
- **Output:** each job writes a summary to the run page and uploads its
report (`override-report-shard*`, `parity-scorecard`) for 30 days.
- **Failures:** a failing effect shard fails the run, and GitHub's usual
failure notification for scheduled workflows applies.
# 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.
-->
- Dispatched on this branch against orca-test-repo `main` and #15693's
build ([run
34933239912](https://github.com/OrcaSlicer/OrcaSlicer/actions/runs/34933239912)).
Every job passed:
- **effect shard 0:** 19m31s; 225 options sliced, 151 effective, none
crashed or hung, every fixture sliced;
- **effect shard 1:** 19m33s; 349 options sliced, 254 effective, same;
- **harness:** 4m32s; all 13 fixtures, 0 new divergences, 0 errors.
- An earlier dispatch on this branch, testing #15693's build against
orca-test-repo's parity branch ([run
34836468900](https://github.com/OrcaSlicer/OrcaSlicer/actions/runs/34836468900)),
passed: effect shards in 20m28s and 23m30s, and the harness reported 0
new divergences.
- The workflow only runs from the default branch on its schedule, so the
nightly trigger itself takes effect once this is merged.
<!--
> 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?
-->
Since #15438, every CLI run that loads a system preset spends about a
second per preset file re-parsing that vendor's entire profile tree. A
slice with a machine, process and filament preset got roughly 2.5 s
slower, and a four-filament slice roughly 4 s slower. This PR loads each
vendor tree once per run instead. Resolved presets and G-code are
unchanged.
The GUI never takes this path, and no release contains #15438, so the
regression only affects CLI runs on current dev and nightly builds. That
includes print farms, slicing services and plugins that call
`orca-slicer --slice`, and CI suites.
## Changes
### Why it was slow
`PresetBundle::resolve_preset_config` resolves a system preset through
its vendor manifest by loading the whole OrcaFilamentLibrary bundle and
the whole vendor tree from JSON, then picking the one preset out. The
CLI did that separately for every `--load-settings` and
`--load-filaments` file, on a fresh `PresetBundle` each time. With BBL
presets, a machine + process + filament run opened `BBL.json` three
times and read BBL's 2,879 profile files and the library's 512 three
times over.
### Load each vendor tree once
- `PresetBundle` keeps every vendor bundle its manifest path loads,
keyed by source root, vendor and substitution rule, and reuses them for
later resolutions on the same bundle.
- OrcaFilamentLibrary is cached the same way, so vendors under one root
share a single library load and the library's own presets resolve from
that same instance. A vendor bundle only reads from its base while
loading, so sharing it is safe.
- A failed or throwing load is not kept, so error reporting is
unchanged.
- The key includes the source root, so presets from two different
profile roots still resolve separately.
- The CLI resolves every system preset through one `PresetBundle` for
the whole run, instead of creating one per file.
The resolved configurations still come from the same canonical vendor
loader, so what a preset resolves to does not change. Only the CLI calls
`resolve_preset_config`, so a long-lived GUI bundle cannot end up
holding profile trees that later change on disk.
# Screenshots/Recordings/Graphs
<!--
> Please attach relevant screenshots to showcase the UI changes.
> Please attach images that can help explain the changes.
-->
CLI slice of a 20 mm cube with X1 Carbon system presets. Both builds get
the same datadir, best of 3, Linux. "Before" is this PR's base from CI.
| System presets loaded | Before | After |
|---|---|---|
| machine | 0.95 s | 0.87 s |
| machine + process | 1.67 s | 0.92 s |
| machine + process + 1 filament | 2.51 s | 0.97 s |
| machine + process + 4 filaments | 5.00 s | 0.99 s |
Files opened during the machine + process + filament run (`strace -e
openat`):
| | Before | After |
|---|---|---|
| `BBL.json` | 3 | 1 |
| `OrcaFilamentLibrary.json` | 3 | 1 |
| `system/BBL/**/*.json` | 8,634 | 2,880 |
| `system/OrcaFilamentLibrary/**/*.json` | 1,536 | 512 |
Peak memory did not rise: max RSS 306 MB → 286 MB for the three-preset
run, and 305 MB → 285 MB for four filaments. The "before" figure is an
AppImage, so part of that gap is probably packaging.
## Tests
<!--
> Please describe the tests that you have conducted to verify the
changes made in this PR.
-->
- New test "Manifest-backed resolution reuses the vendor tree it already
loaded" in `tests/libslic3r/test_preset_bundle_loading.cpp`. It resolves
one preset, changes the parent profile on disk, then resolves a sibling.
The same bundle returns the value it already loaded, and a fresh bundle
picks up the change.
- New test "Manifest-backed resolution shares the library between
vendors under one root". It resolves through one vendor, changes a
library profile on disk, then resolves through a second vendor and a
library preset on the same bundle. Both return the value already loaded,
and a fresh bundle picks up the change.
- All `[Preset][Bundle]` tests pass (87 test cases, 1,069 assertions),
including the existing manifest-backed resolution cases for source-root
scoping, malformed vendor loads, missing parents and type mismatches.
- G-code of the three-preset slice is identical before and after, header
lines excluded.
- The external CLI regression suite passes. Two cases report as
unexpectedly passing because #15639 fixed their bug. They pass the same
way on this PR's base without the change.
- A GUI-vs-CLI parity run over 10 fixtures shows no new differences.
- Builds clean on Linux (Release, with tests).
<!--
> 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 manifest resolver loaded OrcaFilamentLibrary once per vendor it
resolved through, so a run that mixes vendors parsed the library tree
again for each of them. The library is now cached like any other vendor
tree, keyed on its root and substitution rule, and doubles as the base
every vendor under that root loads against. A vendor bundle only reads
from its base while loading, so sharing the instance is safe.
The cache key carries the substitution rule as its enum, and the lookup
lambdas take a const bundle since they only read.
`--export-settings` already writes the merged config as JSON at the
right point in the CLI flow. Passing `-` now writes that same document
to stdout, so scripts can inspect the effective config without a temp
file. This replaces #14605.
- `ConfigBase::save_to_json` gains a stream overload. The file overload
serializes through it before opening the file, so the output format is
unchanged, and a config that cannot be serialized (invalid UTF-8) now
leaves the existing file untouched instead of truncating it.
- On stdout, invalid UTF-8 in string values is written as U+FFFD instead
of ending the process with an uncaught `type_error`. Files keep the
strict behaviour.
- To keep stdout pure JSON, `-` is rejected up front (stderr message,
`CLI_INVALID_PARAMS`, shell status 254) when combined with an action or
transform that can write to stdout or does real work: `--info`,
`--help`, `--orient`, slicing and exporting. Options that do nothing
without a slice (`--uptodate`, `--min-save`, `--pipe`, ...) are still
accepted.
- The one unconditional stdout write on a success path, "skip locked
instance" during arrange, now goes to the log.
- Every other value, including the default `output.json`, behaves as
before.
Tests in `tests/libslic3r/test_config.cpp`: the stream output equals the
file output and keeps the tab-indented format; invalid UTF-8 throws on
the strict path and is replaced when asked; a failed save leaves the
previous file intact.
Opening a model with a relative path, for example `orca-slicer ./some.3mf`,
failed with "Loading of a model file failed." and "The file does not contain
any geometry data.", while the same file opened by an absolute path or by
drag and drop worked.
GUI_App::init_app_config() changes the working directory to <data_dir>/log,
and it runs from the GUI_App constructor because the app config is needed
early for instance checking. The input files are opened much later, in
post_init(), so a path still relative at that point resolved against the log
directory instead of the directory OrcaSlicer was started from, and the 3MF
reader failed to open it.
Resolve the input paths in CLI::setup(), which runs before GUI_App is
constructed and therefore before the working directory moves. Absolute paths
are returned unchanged, so the forms that open today are unaffected, and
custom open protocol URLs are passed through since post_init() hands those to
the downloader rather than the file loader.
The working directory change is left alone. It was added in #3248 so the TUTK
logs land in the data directory instead of the working directory (#3209).
