Two user reports from the same session on the deployed build, 2026-08-23.
FIRST: "if I select a shape (es a circle draw in construction lines) and then I
try to toggle contruction to obtain a full line, does not work". Reproduced: Q
converts the selection and so does the offer's Reference > Construction row —
both run m_keys_sketch['Q'] — but the CHECKBOX, the one control actually
labelled Construction, only ever called set_sketch_construction(), which arms
the mode for the NEXT entity. So the obvious control was the one route that
could not convert existing geometry, and it failed silently while also flipping
the draw mode behind the user's back. It now carries Q's meaning.
Scoped to Select mode, and that scoping is not cosmetic: drawing AUTO-SELECTS
what was just drawn (draw-then-edit), so with a draw tool armed "there is a
selection" does not mean the user picked anything — it means they finished a
line. The first version converted there and turned the box into a trap: arm
construction, draw the axis, click the box to go back to real geometry, and
instead of disarming the mode it converted the axis just drawn. The gesture
ladder's C4 rung does exactly that and reported three construction entities
where it wanted one. In Select mode the intent is unambiguous.
SECOND, and this one destroyed work: "after creation of a circle, a round angled
rectangle and a slot, and mirror of those shapes on a vertical line inside a
outer rectangle, preview is ok but application creates errors: the circle is
mirrored, but rectangle and slot are redrawn as pieces of circles screwing both
the original shapes and the copies." A screenshot came with it, and it showed
more than the words did: the ORIGINALS were wrecked too — the rounded rectangle
was drawn as a four-lobed cloud, each corner fillet having gone the long way
round, and the slot had ballooned into two near-full circles.
Measured on the rig, a slot mirrored about a vertical line:
rails 62.873 / 62.873 -> 2.082 / 62.913
caps r=21.554 sweep=-180.00 -> r=32.214 sweep=-237.66
and all four sources moved, the axis line with them
Cause: confirm_op's Mirror branch bound an Arc copy to its source with a
Symmetric constraint on the CENTRE ALONE. An arc has five degrees of freedom;
pinning two of them leaves the endpoints and the sweep free while the shape's
own coincidences still pull on them, and the solver answers with a different,
internally consistent sketch — which is what a reflex cap and a 2 mm rail are.
A circle came through the same code untouched because a circle HAS no endpoints
to leave free, which is exactly why the failure reads as "circles fine, rounded
rectangles and slots destroyed".
Three parts, and each one is here because the measurement caught the previous
one being half a fix:
1. Arcs are bound by BOTH ENDPOINTS. Endpoints before centre in the ladder:
{p0, p1} is four equations against five DoF and pins the sweep, while
{centre, p0, p1} is six and is refused — the refusal is what silently
degraded the batch to a set that left the sweep free.
2. Every copy is reflected from the PRE-BATCH source, so a batch that disturbs
the sketch cannot hand the next copy already-moved geometry.
3. THE APPLIED RESULT IS THE PREVIEW — checked on the sources AND the copies,
and on violation the whole constraint web is dropped and both halves are
restored to the reflection the preview drew. try_add_constraints rolls back
only when a solve FAILS, and every failure here came from a solve that
succeeded at something else. Guarding only the sources fixed the slot and
left the rounded rectangle's copies at a 13.8 mm rail and a 308 degree cap:
the original was safe and the copy was still wrong, which is half a fix.
The parametric link is kept whenever it provably holds the geometry, and dropped
when it does not. A wrong shape is worse than an unlinked one.
WHY NOTHING CAUGHT THIS: the gesture ladder's mirror rung reflects three
straight LINES. It sat green through the whole defect. C4b now mirrors a slot,
so the reflection has arcs in it, and grades the property the user actually
stated: the copy is the source reflected, the source does not move, and no cap
comes back reflex.
VERIFIED against the user's own scene, rebuilt gesture by gesture on the rig —
outer rectangle, circle, rounded rectangle and slot, a vertical CONSTRUCTION
line as the axis, all 17 entities mirrored in one gesture:
ok the mirror axis is a construction line
ok picked the axis and all 17 entities
ok originals unchanged (moved: [])
ok every copy is the exact reflection (worst 0.000000000)
ok no source arc turned reflex — the 'cloud' failure
ok no copied arc turned reflex (6 arcs checked)
One grader correction worth recording, because it cost a round and would cost
the next one too: a reflection REVERSES ORIENTATION, so a copy legitimately
stores p0/p1 the other way round. Comparing p0 to p0 grades the storage order,
not the geometry, and reported a perfect mirror as an 8.98 mm error. Endpoints
are compared as an unordered pair.
It passed alone and failed inside scripts/ladder-all.sh, twice, on the same property: "right-click
with two picked -> None". The diagnostic it now prints said what a screenshot could not — editing=True
with an empty selection after two clicks that should have picked two entities. Three rig facts, all
about the DRIVER, none of them a product defect:
A LEFT-CLICK ON A LINE'S MIDDLE OPENS ITS LENGTH FIELD. The Select branch tests m_live_quotes before
it picks, with a ~24 px label tolerance, and a line's Length quote sits at its midpoint — so the
click that was meant to select it promoted a dimension and froze the canvas instead. Everything
after it landed on nothing. It bit only when the previous step had left that line selected, because
live quotes are drawn for the SELECTION: hence passing alone and failing in the gate. Lines are now
picked at 0.3 along, clear of the label.
A FIELD THAT HAS NOT OPENED YET READS LIKE ONE THAT NEVER WILL. The queue opens each field from a
CallAfter, so a driver that looks once, sees nothing and moves on gets frozen by the field that
arrives a moment later. keep_as_drawn() now waits for QUIET — two consecutive clear readings — and
draw_line_at drains stragglers before handing back.
A KEY CANNOT CANCEL A SESSION WHOSE CANVAS IS FROZEN. gui-ladder's enter_sketch dismisses the old
sketch with Escapes, which an open field swallows, so the session survived and the four calibration
probes landed in it on top of what was already there: "calibration expected 4 points, got 7". Every
rung now enters through fresh_sketch(), which cancels through the socket first — that cannot be
swallowed.
Measured after the fix, in the sequence that failed: gesture ladder 93/93 then offer ladder 108/108,
back to back on the same app.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
snaporca-ziam said gen_offer_table.py would silently delete the model-mode "Constrain sketch"
row, because that row lived in the generated header and not in tool_atlas.json. Running it found
more than that: FOUR rows existed only in the header — constrain, rename, and the three typed-
value rows sk_length / sk_radius / sk_angdist — and sk_delete's action had drifted, pointing the
sketch row at btn:delete, the FEATURE delete.
All five are now in the atlas, so the header regenerates byte-identically from it. Verbs may carry
a `note`, emitted as a C++ comment above the row: a rationale written into a generated file is
deleted by the next regeneration, which is how this started.
snaporca-z8rs (P1), found by making that true: after the atlas held all 92 verbs, the regenerated
header differed from the checked-in one by EXACTLY ONE LINE — kOfferVerbCount, 91 against 92.
Every consumer loops i < kOfferVerbCount, so the last row of the table was invisible: never listed
by show_offer_menu, never findable by mcp_run_verb. The verb that fell off the end is sk_angdist,
"Angle / distance…" — the typed-value row for a two-entity selection. On the one selection where
you would ask for the angle between two lines, the row that types it was not in the menu.
It survived because nothing compared the Sk2Ent menu against the table: sk_angdist accepts Sk2Ent
and nothing else, so an off-by-one that dropped the LAST verb was invisible from every other
selection. The vocabulary rung now covers Sk2Ent too, and picking the pair taught it one more
rig fact — shift-clicking a circle at its +X point grabs the RADIUS GRIP, which replaces the
selection with that one entity, so the pair silently collapsed to one and the offer answered
SkLine. Correctly, for the selection that actually existed.
gen_offer_table.py --check proves header == atlas and changes nothing; it is now the first step of
scripts/ladder-all.sh, and the only one that needs no rig.
Offer ladder 107/107, gesture ladder 93/93, both on the rig.
snaporca-ziam snaporca-z8rs
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
The 2D vocabulary is 46 verbs: 22 have a shortcut and the gesture ladder drives them, 24 have
none and nothing had ever exercised those. They are reachable only from the right-click offer, so
a key-driven ladder could not have touched them whatever it did. Four new rungs drive all 24, and
the coverage claim itself is now arithmetic against DesignOffer.hpp (rung O8) rather than a
sentence in a comment that rots when a verb is added.
The assertions are CONSTRUCTION invariants wherever a click cannot be exact — a regular polygon's
sides are equal to 1e-9 and its vertices lie on one circle; a tangent arc's radius at the shared
endpoint is perpendicular to the line to 1e-9 (measured cos 5.97e-17); the three clicks of a
3-point circle all lie on it; a circumscribed pentagon's circumradius is the inscribed one's over
cos(pi/5), 1.236067977 against 1.236067977. Where a value field opens, the typed value is graded
exactly: a moved line travels +25.000000000 in X and 0 in Y, a rotation turns 30.000000000 deg
and leaves the length alone, a scale multiplies it by exactly 3, a linear array's pitch is
[20.0, 20.0, 20.0] and a polar one's spokes are 60 deg apart all the way round.
Three defects found doing it, all fixed here:
snaporca-ua9g (P1) — delete_selected left three things behind. The AUTO-EDIT QUEUE, so a queued
field opened on a deleted entity and its commit went nowhere: draw a rounded rectangle, delete
everything, draw a 2-point circle, type 30 — the field opens, the digits are accepted, and the
radius stays 32.992020763. reset_autoedit() exists for exactly this and its own comment says so;
it was simply never called from here. The FEATURE GROUPS, whose [begin,end) ranges all shift on a
delete, so feature_of() answered with a group the user never drew — survivors are now remapped
and any group that lost a member is dropped, the rule the placed quotes already followed. And the
SOLVER STATE: no re-solve, so sketch_describe reported dof=16 for a document holding one circle.
snaporca-ekt9 (P2) — the read-back could not see three of its seven entity types. Ellipse,
EllipseArc and BSpline serialised as a bare type name: no centre, no semi-axes, no rotation, no
sweep, no poles. gui-ladder's ellipse rung had to grade the faceted area of the loop at 2e-2 —
that tolerance IS the faceting error — and its spline rung could only count entities. Now they
carry their parameters, and the ellipse arc's ends are asserted to satisfy (x/a)^2+(y/b)^2 = 1 to
1e-9.
Also read-only, and the reason the other two were found at all: sketch_describe now reports the
armed TOOL, the count of PENDING anchors, and whether a value field is EDITING. A menu walk that
lands one row off arms a neighbouring tool and then draws something plausible — the first run of
the authoring rung drew a circle of area 45238.93 and graded it as a rectangle. Every menu pick
now asserts which tool it armed, and the polyline rung (a per-segment Length field freezes the
canvas after every click) could only be written once the driver could ask whether a field was open.
Offer ladder 102/102 -> 105/105 with coverage. Gesture ladder 93/93 and the kernel suite
188 cases / 2532 assertions, both unchanged.
snaporca-ua9g snaporca-ekt9
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
The gesture ladder proved the TARGET — a complex closed profile, exact in vertices, lengths, arcs
and symmetry, voids correctly attributed. It proved it by arming every tool with a letter key,
which leaves the goal's own MECHANISM untested: the design logic pivots on right-click, and the
verbs offered are supposed to adapt to the element under the cursor. 47 of 86 Design-tab verbs
have a GUI action and no shortcut, so a key-driven ladder cannot reach more than half of them.
scripts/offer-ladder.py drives the menu. It asserts nothing from pixels: show_offer_menu emits an
[OFFER] trace from the same loop that builds the rows (behind the existing SNAPORCA_KEYTRACE), so
what the ladder reads cannot drift from what the user is shown, and the expected row set is
predicted by parsing DesignOffer.hpp rather than transcribed by hand. 25 properties, four rungs:
what each element type offers, that the menu equals the table for four selections AND that the
four differ, a 120 x 80 profile authored entirely through the menu, and a tool with no keyboard
route at all driven from the only door it has.
Two real defects, both found by it, both fixed here:
snaporca-ghcz (P1) — right-click was a black hole while any draw tool was armed. Every draw case
ended with `if (evt.RightDown()) { m_points.clear(); return true; }` and returned true even with
nothing to abandon; on_mouse records that in m_right_consumed and DesignCanvas suppresses the
offer whenever it is set. Measured: with Line armed, two right-clicks in a row produced no menu
and no tool change; only Escape freed it. Same rule snaporca-xmh6 wrote for the selection —
clearing nothing is not a gesture terminator. One shared right_abandon() now consumes the click
only when an anchor was really down; 16 sites, plus Polyline/BSpline (which end a chain, correct
only when there IS one) and Point (which has no anchor at all).
snaporca-lnri (P2) — right-clicking a sketch point offered the empty vocabulary. select_at_screen
tests hit_test_point first and records the hit in m_point_sel, but the offer counts m_selection
only, so a Point entity could never reach the entity branch and SkPoint was unreachable by
construction. A Point IS its own handle, so it is selected as an entity; other entities keep the
handle pick, since a line's endpoint is a drag target, not a vocabulary.
Offer ladder 25/25, gesture ladder 93/93 (no regression), both on the rig. The offer ladder joins
scripts/ladder-all.sh as the fifth rung.
snaporca-ghcz snaporca-lnri
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
Draw-then-edit is armed from a jump in the entity count: render() sees n > m_autoedit_seen,
selects the last entity and schedules open_primary_autoedit. A scripted add made while a
creation tool was armed looked exactly like a drawn gesture, so it opened that tool's value
field — and an open field freezes the canvas (on_mouse_impl returns early on m_awaiting_length)
and swallows every letter (in_text includes inline_busy()). Measured on the rig: after
sketch_add, 'p' + click added nothing (4 entities before, 4 after); one Escape and the identical
sequence gave 5. It also explains the selection = [last index] that sketch_describe reported
although action_sketch_add never selects anything — the render pass wrote it.
Escape worked because it sequences two set_tool calls: the pending CallAfter fires between them,
so the second one commits the field it finds open. Arming a tool directly is one call, and the
CallAfter fires after it.
Fix: resync m_autoedit_seen at the end of add_entities_scripted, so a scripted add is not read as
something the user just drew. An already-open field is left alone. Covers sketch_add,
sketch_mirror and sketch_offset — the three callers.
The scale rung's Escape workaround is deleted, which is the issue's acceptance criterion; it is
now the regression test. Gesture ladder 93/93 on the rig.
snaporca-j7gc
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrMzTpAf78U4NG2M8jfvHY
Over the whole 977-sheet corpus: 767 graded, 767 fully clean, 0 failures. The 210 skips are sheets whose part outline is not a closed stroked path at all (largest loop 5 to 132 mm2, measured), and one of them — MPD133 — is password-protected, which was being reported as an engine ERROR. Both now say what they are.
snaporca-j6sr
Two commits carried across (snaporca 579a9a9162, f68613cfc5).
Past about 480 entities a sketch had NO constraints at all and said nothing: libslvs
declares MAX_UNKNOWNS = 1024 and is handed every entity in the sketch at two params
per point, so the whole system came back TOO_MANY_UNKNOWNS and try_add_constraints
rolled the entire inferred batch back. From there no dimension could ever be applied.
Constraints only couple entities that share a point, so the solver now falls back —
only on TOO_MANY_UNKNOWNS — to solving connected components separately and committing
all-or-nothing. The auto-constraint pass batches its Horizontal/Vertical constraints
instead of one solve each, which is what kept the bulk path fast once solves started
succeeding: a 1204-entity load went 1585 ms -> 562 ms.
Plus the scale rungs (a thousand-entity plate drawn on by hand; the heaviest real
drawings graded and timed), the --step 1 fix that used to select nothing while
reporting a clean run, and scripts/ladder-all.sh as the one-command gate.
Parity 17 identical / 8 diverging as expected. Kernel suite here: 188 cases /
2532 assertions, including "a sketch past the solver's unknown limit still solves".
snaporca-yww4, snaporca-x6v7, snaporca-j6sr
and a ladder that draws with the mouse
Three commits carried across (snaporca 4ffd60eacb, 421055c2ec, b71216ce0b):
1. A design made only of sketches must survive being saved. CadDocument::recompute
returned false with "no solid-producing features" for a document that has no
solid, and two callers read that as "unusable": the GUI syncs the 3MF recipe
only after a successful recompute, so a sketch-only design was saved with no
recipe at all, and deserialize_recipe ends with `return recompute()`, so even a
project that carried one was refused on load. Having nothing to build is now a
success; a feature that MEANT to build a solid and produced none still fails.
DesignPanel::refresh_tree syncs the recipe too, for the paths that call
m_doc.recompute() directly.
2. Scripted geometry arrives exact. The Horizontal/Vertical inference window and
the endpoint weld window both close to zero for add_entities_scripted; void
attribution probes from a point strictly inside each loop instead of from its
first vertex. Corpus rung 39 graded / 39 fully clean, was 35 with 6 failures.
3. scripts/gui-ladder.py — 17 rungs, 84 properties, all driven by synthetic clicks
and typed values rather than through the socket.
Parity 17 identical / 8 diverging as expected. Kernel suite here: 188 cases /
2532 assertions.
snaporca-mtav, snaporca-8xg1, snaporca-5hvl, snaporca-730j
Carries snaporca 572f794c84, d0f9a0052a, 9f7e4e3627 and 3974f8a170. Parity holds: 17 files
identical, 8 diverging by their expected counts.
EXACT AREA. A loop's area is now integrated entity by entity in traversal order — Green's
theorem — instead of being shoelaced over the render polyline, which faceted every arc into 24
chords and lost 2.02 mm2 on a 3706.86 mm2 stadium. 0.054%, invisible on screen, and wrong in a
number reported as "the area".
OFFSET FOLLOWS THE TRAVERSAL. Offsetting a mirrored profile put one half on the wrong side and
split the loop in two, because the chainer only followed p1->p0 links and each entity's offset
side was taken from its stored direction. Chains are now orientation-aware, seeded at a free end,
offset by `reversed ? -d : d`, and normalised head-to-tail on the way out — so offset is correct
for any input ordering and its own output cannot reintroduce the problem.
Both are the same underlying lesson, which has now cost three separate defects: an entity's
STORED direction is not its direction of TRAVEL around the loop.
THE LADDER. scripts/sketch-ladder.py is a graded suite of 2D sketches judged the way a person
judges them — VERTEX, LENGTH, ARC, TANGENT, SYMMETRY, CLOSED — with area only as a cross-check,
because area is derived and nobody can confirm it by eye. Eight rungs from a rectangle up to
MPD5 from the StudyCadCam corpus, a dia 27 x 95 pin reproduced as its revolve half-profile with
the R5 fillet tangency solved exactly. Entirely 2D: no extrude or any solid feature.
Kernel here: all tests passed, 2681 assertions in 231 test cases, including the new
"profile: a mirrored half offsets as one loop, not two". GUI target builds and links.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Carries snaporca 39fac9b725. Parity re-verified: 17 files identical, 8 diverging by their
expected counts, DesignPanel.cpp still at 32 — the mirrored files were copied and the two
divergent ones patched hunk by hunk, so the counts returning to their expected values is the
proof each landed on the right side.
All 90 offer verbs are now firable by name over the socket, which matters because a deck key
can only send a keystroke and 49 of them have no shortcut at all. sketch_set_value calls the
same apply_dimension the in-canvas value field calls, so a typed dimension can be asserted with
no window manager in the way.
Three guards came with it, each confirmed against the source: on_mass_properties bounds-checks
m_sel_solid_body (it defaults to -1, and run_verb bypasses the menu grey-out that used to hide
that); sketch_set_value validates its value at the boundary because apply_dimension records a
driving constraint even for values it refused to apply; and run_verb refuses btn:/fly: verbs
that do not apply to the selection while leaving key: verbs alone, so the socket offers exactly
what the GUI offers. Dispatch is deferred through CallAfter so no modal verb can wedge the
socket thread.
GUI target builds and links against the rebuilt deps image.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Carries snaporca 971320e129, 6b049f0dc6, 4aae782029, 444d59f212, 74cf3d7e54 and the build
guards from 597557a6e4. Parity re-verified after every hunk: 17 files identical, 8 diverging by
their expected counts — DesignPanel.cpp still 32, DesignCanvas.cpp still 16, which is the proof
each hunk landed on the right side rather than being copied over a real divergence.
OFFSET OFFSETS THE CHAIN. Per-entity offsetting returned a closed rectangle as four parallel
segments that no longer touch, so entities_to_wires gave four OPEN wires and nothing could be
extruded. offset_entities now chains by shared endpoints and repairs each seam by mitering the
neighbours to their intersection. Second bug, invisible to any single-entity test: +d meant
"left of travel" for a line but "radius + d" for an arc regardless of sweep, so a slot outline
offset with its straights going one way and its caps the other. The convention is now written on
the declaration and pinned by a test.
tests/libslic3r/test_sketchprofile.cpp is new and asserts the LOOP rather than coordinates —
the property that decides whether a profile can be built, and the one the existing single-entity
[SketchEdit] cases cannot see. Its include is catch2/catch_all.hpp here: this fork ships Catch2
v3 while snaporca is on v2, which is why the test files are a tolerated divergence.
RIGHT-CLICK PICKS WHAT YOU POINTED AT, so a line's own verbs are offered instead of the
empty-selection vocabulary; sk_delete stops sharing btn:delete with the feature tree; and an
element's defining number (length / radius / diameter / angle / distance) can be typed, from the
menu or from V.
TWELVE MCP SKETCH VERBS. The socket had ~40 verbs and none touched a sketch, so the 2D layer
could only be exercised by driving a GUI with synthetic clicks. sketch_describe reports each
closed loop, the loops it encloses as voids, exact areas, and where a chain is still open;
sketch_validate/sketch_heal are FreeCAD's ValidateSketch — find vertices that overlap within a
tolerance but carry no coincidence, then weld them AND record the constraint, so a loop closed
by floating-point luck becomes one closed by construction. scripts/mcp-sketch-smoke.py is the
loop that asserts all of it.
Kernel suite on this fork: all tests passed, 2677 assertions in 230 test cases. The GUI target
links against the rebuilt deps image (the wxInspector blockage is gone) and the binary carries
the new verbs.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Building orcacad-deps from this Dockerfile removes the snaporca-deps base, which is the point
(Trap 1: the old image's baked tree was project(Snapmaker_Orca)) — but that base was also where
Xvfb, openbox, xdotool and scrot came from. Without them gui-session.sh reports display DOWN and
orca-slicer dies with a trace trap on no display. Verified: the session now comes up with
display up, wm up, and the Untitled - OrcaSlicer window present.
Raises the bound from -j8 to -j12: the incident dump measured 1.17 GB average per cc1plus,
so 12 in flight is ~14 GB typical, well inside the 40 GB cgroup ceiling that is the real
guarantee. Keeps the file byte-identical to snaporca's copy, which the header requires.
Dockerfile.deps builds the deps with -j 12 for the same reason, and symlinks
deps/build/destdir -> deps/build/OrcaSlicer_dep: this tree installs under the latter name
while the orcacad_buildcache volume has the former baked as absolute paths in its CMakeCache.
Same tree, two names — without the link, one directory rename costs a full cold rebuild.
Both forks ran this script at once on 2026-08-21, each with ninja -j$(nproc)=16.
36 cc1plus held 42 GB of a 62 GB box, the kernel OOM-killed for 2h28m, ssh went
unreachable, lightdm was destroyed (2946 session kill events), and neither build
produced a single object file.
Three bounds, weakest to strongest: a flock on a path SHARED by both forks so
they serialise instead of summing; -j8 so the box stays usable while it
compiles; and --memory on the container, which is the actual guarantee — a
runaway build now dies inside its own cgroup instead of taking the host with it.
--memory-swap is pinned equal to --memory because swap thrash is what made ssh
hang rather than fail.
Kept byte-identical to the copy in the snaporca fork, as the header requires.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* Add caching system for presets
* Removing user\bundle serialization and keeping it only for system presets
* Integrate caching into WebGuideDialog which speeds up time of SetupWizzard and PrinterSelection dialog
* Add CI\CD step to prepare cache file in ahead of time so user does not need to wait
* Add partial cache generation when only one of the vendros is changed to speed up recalculation time
* Handle corrupted files
* Add cache to GuideDialog as previos version didn't work as expected
* Add inspecting tool and fix CI cache generation
* Generate cache per vendor
* Simplify code by mergin it in PresetBundle
* Simplify code a bit more
* Add cereal serialize() to VendorProfile, PrinterModel, Preset, and Semver
* Remove CachedPrinterModel/VendorProfile/Preset mirror structs from VendorCache
* Fix use-after-free in CallAfter lambda; replace raw thread pointer with unique_ptr
* Use get_vendor_cache_key() to match cache keys written by the app
* Remove BOM added by VSC
* Skip invalid vendors
* Remove leftover cache file
* Fix build for windows arm64
* Revert json cache back
* Update check for stale cache
* Serealize all value fields for Preset class to minimize regression later
* Minimize field duplication by moving Cache thing into PresetBundle
* Add tests for Cache system
* Add a bit more tests
* Merge branch 'main' into feature/cache_profiles_and_optimize_loading_speed
* Rvert from per-verndor to single cache file
Replace N per-vendor .cache files with a single system_presets.cache
that holds all vendors and presets in one serialized blob.
Cache load is now all-or-nothing: on hit all vendors are applied from
the bundle (sub-second); on miss all vendors are parsed from JSON and
a fresh bundle is written to the user cache dir.
Invalidation is driven by bundle_key - a sorted concatenation of all
vendor JSON version strings. Any vendor update invalidates the whole
cache and triggers re-parse on next launch.
Guide wizard (WebGuideDialog) loads the bundled cache into a plain
PresetBundle instead of a separate VendorGuideData struct, removing
the duplicate data model.
generate_system_cache simplified from a per-vendor loop to a single
save_system_presets_cache() call producing one output file.
* Transfer all Preset fields from cache via move assignmet
apply_vendor_preset_group was copying fields manually and missed
bundle_id, user_id, base_id, sync_info, updated_time, key_values,
ini_str. Replace field-by-field copy with move assignment of the
fully-deserialized Preset, then restore the vendor pointer which
is excluded from serialization.
* Ignore cache for future
* Remove not used files
* Ship one preset cache per vendor in place of the profile JSONs
Each vendor's system presets serialize into a single <vendor>.opc built at
package time, and a shipped build carries that file alone — the profile JSON
and its sub-file tree are pruned. The vendor loader, the setup wizard's profile
list and the resource installer all read a vendor through its cache, falling
back to parsing whenever one is absent, stale or unreadable, so the cache stays
an optimization and never a source of truth. Caches hold presets in source form
and resolve inheritance at load, through the same code the JSON path uses.
* Make the preset cache self-describing and load each vendor from the system folder alone
The cached DynamicPrintConfig is keyed by name, through a per-file dictionary of the
distinct opt_keys, the type each was written as, and the distinct enum value names,
instead of by serialization_key_ordinal — a position assigned by declaration order at
static init, where inserting one option shifts every later ordinal and the lookup then
succeeds on the wrong option. Because a name-keyed payload drops the options this build
cannot place rather than being rejected wholesale, the schema fingerprint goes, and with
it the two fallbacks that existed only because an installed cache died on every app
upgrade: the second lookup tier into resources/profiles and the parse fallback to the
same place. A vendor is loaded from <data_dir>/system/ and nowhere else, as on main —
which is what makes the app write its .opc files there again.
* Simplify the preset cache internals after review
* Use the shared temp-dir helper in the preset bundle loading test
* Bound stamp string reads in the preset cache
* Speed up the setup wizard with a profile-data cache
The wizard's per-vendor fast path threw on vendors present only in
resources, falling back to a ~29 s raw JSON scan on every open. Each
vendor now loads from the directory it was found in, and the derived
model/machine/filament/process catalog is cached whole in
<data_dir>/cache/wizard_profile_data.json, stamped by each vendor's
name and version - a fresh cache makes an open one file read, with no
bundle built and no presets installed (~0.2 s vs ~2 s).
* Remove debug SVG dump from a geometry test
* Move the per-vendor cache file format into PresetCacheFormat
* Move the vendor install helpers from PresetBundle into Utils
* rename
* fix flatpak
* change cache version to 1
---------
Co-authored-by: SoftFever <softfeverever@gmail.com>
This script builds only libslic3r_tests, which links libslic3r and no GUI code —
but cmake still processed the whole if(SLIC3R_GUI) block and every find_package
inside it, so the kernel suite silently depended on the GUI's dependency set.
That came due the moment upstream added wxInspector as a REQUIRED find_package:
the orcacad-deps image predates it, so configure died pointing at
src/CMakeLists.txt:92 with nothing about the kernel having changed. Turning the
block off is not a workaround for that one dependency — it is the suite finally
declaring what it actually needs, so the next GUI-side dependency added upstream
cannot break it either.
Surfaced by taking SoftFever's merge of main into the PR branch.
snaporca-7ih's remaining half. Both flyout factories registered their verbs INSIDE the
widget-building loop, so the ~40 retired tool buttons had to be constructed and then
Hide()n: skipping construction would have deleted 42 offer verbs (26 fly:<family>#<row>
+ 16 Shift+keys) while their rows still rendered and did nothing when picked.
Register first, build second. The addresses are pure data; the widget is one door onto
them, not their owner. A family absent from kBarKeep now returns before any wxWindow is
made. The keep-list stays a one-line data decision, not a structural one.
And close the class of bug for good: the constructor now verifies, once, that every verb
the atlas marks wired resolves to a real registration, logging each break and asserting in
debug. Rows that render and do nothing have shipped three times (edit_feature and sk_move
with action:null, then this) and are invisible from either side alone.
Verified on the snaporca rig by walking the offer, not by reading the code — all four
at-risk address kinds run with no widget behind them: fly:design_rect#2 drew an OBLIQUE
rectangle (the third variant, not the family's first), key:S+E opened Extrude with its
10 mm gizmo, fly:material#4 opened Thicken. Hover hints, icons and nesting intact. This
fork is code-identical here bar the two permitted DropDown divergences; it still owes a
build of its own (snaporca-5pl).
Two hints were wrong and are fixed: Cut said "Split the body with a plane", colliding with
the Split verb one row away and pointing at a card for a value the canvas already offers as
a draggable arrow; Split never said its plane comes from a picked face.
Also, because it blocked the verification and will block the next one: gui-session.sh
killed by full path while its own app_pid() matched by basename, so a differently-pathed
instance survived, held the single-instance lock, and got reported as a healthy session —
a Jul-30 binary nearly passed as this build. It now kills by basename and prints which
binary is actually on screen. Traps 6 and 7 documented.
Mirror of snaporca ac85277bac..0e7cb3ec78 (six changes, applied as a patch to
DesignPanel.cpp rather than copied, so this fork's 30 permitted divergent lines survive
— parity re-checked afterwards: the five shared files are byte-identical, DesignCanvas.cpp
and DesignPanel.cpp differ by exactly 16 and 30 lines).
- The straight slot's inline field says Radius, which is what it sets. It stores the
half-width and passed the typed number through unchanged, so 30 produced a 60 mm slot.
- The offer opens from the keyboard (Menu, Shift+F10), anchored on the viewport rather
than wherever the pointer happens to be. The card hint names the new route.
- The sketch card's Plane combo is gone; the plane comes from the viewport. Also stops
build_candidate collapsing a face plane to a base plane while editing.
- Mass properties and the dead Edit row are wired into the offer; DesignOffer.hpp is
regenerated from tool_atlas.json, verified by re-running the generator and diffing.
- docs/rig_build_traps.md + scripts/rig-build.sh, which derives its fork identity from
project() so it cannot be pointed at the other fork's image or volume.
- docs/design_tab.md refreshed (44 commits stale) + a PR description, with this fork's
own merge-base and diff shape rather than snaporca's.
Built green in the deps container with the new script and verified on the rig: Menu and
Shift+F10 both open the offer at the viewport centre with the pointer parked off-canvas,
and the sketch card shows no Plane row.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LyRwbuq6fjn3VV9U9UvhBM
The relaunch sequence was an ad-hoc pile of docker exec one-liners, and it
had a real bug: it dismissed the first-run dialogs by computing the
titlebar close box from `xdotool getwindowgeometry --shell` and clicking
it. When the dialog had already closed, that eval left the geometry
variables stale or empty, the click landed at a garbage coordinate, and it
kept hitting the Sketch button in the toolbar underneath — so the app came
up in sketch mode with a stray Sketch feature that then had to be
cancelled by hand. Three times in one session.
The fix is not a different mechanism. `xdotool windowclose` looks cleaner
and KILLS THE APP: it destroys the GdkWindow out from under the dialog and
the process dies with "GdkWindow unexpectedly destroyed", three
GLib-GObject criticals and a segfault. Measured, not guessed — that is
what the first version of this script did. Escape does not close the Setup
Wizard either, which is why it needs handling at all. So the titlebar
click stays, and what changes is that it refuses to click geometry it has
not validated: the window id is re-resolved immediately before, all four
geometry variables are unset first and must come back numeric, and the
computed point must be inside the screen. Any of those failing logs why
and clicks nothing.
Two further honesty fixes in the status output, both caught by reading it
rather than by it failing: app_pid skipped nothing, so with a container
full of <defunct> instances it printed a dead pid as though the session
were healthy — it now walks /proc/<pid>/stat and ignores zombies. And a
container without x11vnc reported "vnc: DOWN" as if something had broken,
when nothing was ever installed; it now says so, and does not try to start
what is not there.
Also replaces the fixed post-launch sleep with a wait for the main window,
because cold starts under software GL vary by a lot, and adds --status for
diagnosis: "no windows but the desktop is up" means the app died, "cannot
connect at all" means the desktop did. That distinction cost real time to
work out by hand.
Verified on both containers: one run each, wizard closed on validated
geometry, no stray sketch mode, live pid reported, and the app still up.
snaporca-e1p adjacent (tooling, not the tab itself).
* Add wxInspector dep
* Initial intergration of wxInspector
* docs: add wxInspector plugins design spec
Design spec for two wxInspector plugins (DPIAware + CustomWidgets) that expose
OrcaSlicer's custom control properties in the inspector property grid.
Covers: DPIAware scale-factor properties, Button, CheckBox, TextInput,
SwitchButton, ProgressBar, Label, and LabeledStaticBox.
* docs: add wxInspector plugins implementation plan
6-task plan covering: source changes to existing widget headers,
DPIAwarePlugin, CustomWidgetsPlugin, registration helper,
MainFrame/CMake wiring, and build verification.
* feat: add getters/setters for wxInspector plugin access
Add minimal public accessors to DPIAware (set_scale_factor,
set_prev_scale_factor, set_em_unit, force_rescale), Button
(GetStyle, GetType, IsSelected), CheckBox (IsHalfChecked),
TextInput (GetCornerRadius), and LabeledStaticBox
(GetCornerRadius, GetBorderWidth, GetBorderColor, GetScale).
* feat: add wxInspector plugin registration helper
Add RegisterOrcaInspectorPlugins() inline function that creates
and registers the DPIAwarePlugin and CustomWidgetsPlugin as
static instances (matching wxInspector's built-in pattern).
* feat: add DPIAware wxInspector plugin
Exposes DPI scaling properties (scale_factor, prev_scale_factor,
em_unit, normal_font, force_rescale) on DPIFrame and DPIDialog
widgets. Uses dynamic_cast for detection and a template helper
to capture the correct static type for lambda accessors.
* feat: add OrcaCustomWidgets wxInspector plugin
Exposes Orca-specific properties on 7 widget types:
- Button: Style, Type, Selected
- CheckBox: Half Checked
- TextInput: Label, Text Value, Corner Radius
- SwitchButton: Value
- ProgressBar: Proportion, Show Number
- Label: Is Hyperlink, Font Point Size
- LabeledStaticBox: Corner Radius, Border Width, Border Color, Scale
Each widget type uses dynamic_cast for safe detection.
* feat: wire wxInspector plugins into MainFrame and build
Call RegisterOrcaInspectorPlugins() in MainFrame constructor after
SetupInspectorAccelerator(). Add all 5 plugin source files to
SLIC3R_GUI_SOURCES in CMakeLists.txt.
* fix: move plugin registration to GUI_App::on_init_inner
Register plugins once in app init rather than in MainFrame
constructor, which may be recreated during the application
lifetime.
* fix: include plugin headers in Registration.hpp for complete types
Static locals require complete type. Include DPIAwarePlugin.hpp and
CustomWidgetsPlugin.hpp instead of forward-declaring. Also remove
unused include from MainFrame.cpp (registration moved to GUI_App).
* fix: qualify DPIFrame/DPIDialog with Slic3r::GUI namespace
* Make DPIDialog inspectable. For other dialogs, we will add them if necessary later.
* docs: add spec for moving wxInspectable into DPIAware template
Move wxInspector::wxInspectable base class from DPIDialog and MainFrame
into the common DPIAware<P> template, making all DPIAware widgets
automatically visible in the inspector tree.
Co-Authored-By: Claude <noreply@anthropic.com>
* docs: add implementation plan for moving wxInspectable into DPIAware
Co-Authored-By: Claude <noreply@anthropic.com>
* docs: update spec/plan — move SetupInspectorAccelerator into DPIAware too
Co-Authored-By: Claude <noreply@anthropic.com>
* refactor: move wxInspectable and SetupInspectorAccelerator into DPIAware
DPIAware<P> now inherits wxInspector::wxInspectable and calls
SetupInspectorAccelerator in its constructor, making all DPIAware
widgets automatically appear in the inspector tree with the
Ctrl+Shift+I shortcut. DPIDialog now uses 'using' to inherit the
constructor. Remove redundant wxInspectable inheritance and
SetupInspectorAccelerator calls from DPIDialog and MainFrame.
Co-Authored-By: Claude <noreply@anthropic.com>
* fix: use LB_HYPERLINK constant instead of magic number 0x0020
Co-Authored-By: Claude <noreply@anthropic.com>
* Clean up
* Fix Linux build
* Don't build wxInspector sample
* Use shallow clone
* Try fix flatpak build
* Attempt to fix build again
* Fix build failure caused by https://github.com/wxWidgets/wxWidgets/commit/436c16135ec7ddf580f44624bf74c592aae43b66
* wxWidgets build only download required submodules
* This should fix build on Windows on ARM
* Enable PIC
* Disable layout inspector by default for public release
* Use wxInspector 1.0.0 release
---------
Co-authored-by: Claude <noreply@anthropic.com>
snaporca-tkz, the last quarantined test. Root cause read out of the vendored
source rather than guessed: slvs/constrainteq.cpp, Type::ARC_LINE_TANGENT does
ExprVector ap = SK.GetEntity(arc->point[other ? 2 : 1])->PointGetExprs();
so it dereferences the ARC'S ENDPOINTS. A full circle entity carries only
point[0], its centre. point[1] and point[2] are zero handles, FindById throws
"Cannot find handle", and the process ABORTS rather than failing the solve —
taking every later test in the binary with it. That is also the wrong equation
for a circle regardless: it only makes the line perpendicular to the radius at
an endpoint that does not exist.
CT::Tangent no longer hands a full circle to that constraint. For a circle it
emits PT_LINE_DISTANCE(centre, line) = radius, which is precisely what tangency
to a circle means. Arcs keep the ARC_LINE_TANGENT path they are built for.
One limitation, stated rather than buried: the slvs C API takes a constant
distance and offers no way to reference the circle's radius parameter, so the
radius is captured when the constraint is emitted. That is exact whenever the
radius is fixed or is simply not driven by another constraint in the same
solve, and re-solving restores tangency if something else moves it. Tying them
would need an auxiliary point constrained onto both the circle and the line.
With this and eeca6794e7, both quarantined tests are gone and the exclusion in
kernel-test.sh goes with them. A green run now means the whole CAD suite
passed, not "everything except the two we gave up on":
149 cases / 2043 assertions, no filters.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Follow-through from eeca6794e7. The header claimed two pre-existing failures
are excluded and named both; the internal-thread case now runs like any other,
so only the solver SIGABRT is left. A comment that lists a test which is no
longer excluded sends the next reader looking for something that is not there.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The Shellcheck workflow has failed on every push and every scheduled run since
2026-07-24, on exactly one finding: SC2029 in scripts/kernel-test.sh, the file
the CAD branch added. So the CAD work is what turned that job red, and a PR
arriving with a red job is a bad way to open a conversation with a maintainer.
Client-side expansion of $REMOTE is the intended behaviour — it is derived from
$VOL locally and the remote has no such variable, exactly as the rsync
destination two lines down relies on. So this is a disable with a reason, not a
silencing: the note says why the warning does not apply.
Verified by running the workflow's own command over all 24 matched scripts:
exit 0, no findings.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This fork's Unit Tests job failed on every single commit, because CI runs the
whole ctest suite including the two cases tagged [known-broken] that
scripts/kernel-test.sh has always excluded locally. A job that is red
unconditionally is worse than no job: it trains everyone to ignore it, so the
next genuine regression arrives invisible.
No workflow change was needed. scripts/run_unit_tests.sh already passes
-LE NotWorking, and tests/CMakeLists.txt registers Catch2 tags as ctest labels
via catch_discover_tests(ADD_TAGS_AS_LABELS) — so the exclusion upstream
already ships works as soon as the cases carry the tag. Verified against the
built test tree: 337 tests unfiltered, 335 with -LE NotWorking, i.e. exactly
these two dropped and nothing else.
The second cause recorded in the issue, the test-reporter step failing with
"Resource not accessible by integration: 403" on a fork, is already fixed
upstream: the Publish Test Results step now carries continue-on-error: true.
The comment these cases carried claimed CI kept the bugs visible by reporting
them forever. That is now false and was never a good mechanism anyway, so
visibility moves to the tracker: snaporca-tkz for the solver SIGABRT, and
snaporca-kzy, filed now, for the thread groove volume. Neither is fixed;
neither is forgotten.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
First successful link of orca-slicer in this fork's history (158 MB, 738/738
objects, plus OrcaSlicer_profile_validator). 1633005bba got the test binary
green; the GUI had still never linked.
The blocker was in the deps image, not the code. That commit's message says
OCCT V7_6_0, Boost 1.84.0 and OpenCV 4.6.0 "are pinned identically in both
forks and were reused as-is". That is true of the VERSIONS and false of the
FLAGS: this fork's deps/OpenCV/OpenCV.cmake passes -DWITH_JPEG=OFF,
-DWITH_TIFF=OFF and -DBUILD_TIFF=OFF, and snaporca's does not. So the reused
build shipped lib/opencv4/3rdparty/liblibjpeg-turbo.a, which collides with the
deps' own lib/libjpeg.a — wx pulls that in via wxUSE_LIBJPEG=sys — on
jpeg_stdio_dest. Under -flto that duplicate is fatal, not a warning. snaporca
never sees it because its GUI does not link libopencv_world.a at all.
Fixed by rebuilding OpenCV 4.6.0 in orcacad-deps with this fork's own flags
(read out of the recipe rather than retyped), from the source already cached in
the image, into a clean build dir so no stale cache entry survived, and deleting
the orphaned bundled jpeg archive. The rebuilt libopencv_world.a has no jpeg or
tiff symbol references and its CMake config no longer names either library.
--allow-multiple-definition would have hidden this while leaving OpenCV carrying
codecs mainline deliberately turns off.
Lesson for the next dep: comparing deps recipes by version is not enough, diff
the CMAKE_ARGS.
Two more mounts, same root cause as the CMakeLists.txt mount this script
already documents — the baked tree is snaporca's:
- build_linux.sh, which builds `--target Snapmaker_Orca`; here the target is
OrcaSlicer and its output name is orca-slicer, so configure passed and ninja
then died on "unknown target". The binary check was looking for the wrong
name too.
- scripts/, because the packaging step needs scripts/appimage_lib_policy.sh;
without it a fully successful link still exited non-zero with "missing
AppImage helper" and the binary check never ran.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Both scripts defaulted IMAGE to snaporca-deps. That image is Snapmaker-based
and lacks Eigen 5.0.1, CGAL 5.6.3, wx 3.3.2 and Python 3.12 Development.Embed,
so running either script here without an explicit IMAGE= dies at CMake
configure — which is a large part of why this fork reached M8 having never once
compiled (1633005bba). The volume defaults were fixed in that commit; the image
default was missed in both files.
docker-iter-build.sh also never mounted deps_src, so it could not have
configured even with the right image: root CMakeLists.txt:947 FATAL_ERRORs
when deps_src/pybind11/include/pybind11/pybind11.h is absent, and
src/CMakeLists.txt pulls semver/hints/imgui/imguizmo/hidapi from the same tree.
kernel-test.sh got that mount in 1633005bba; this is the same fix for the GUI
build path, needed before the Phase B GUI work can be ported here.
CadDocument.hpp: mate_kind comment, mirrored verbatim from snaporca dbaa104f62
so the header stays byte-identical across forks.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
206/206 targets built, 139 [CadDocument] cases / 1960 assertions passing —
identical to snaporca's suite. Until now this fork had never compiled at all:
CMake died at configure, so the M1-M8 "suite green" figures were snaporca's
alone and the ports rested on patch-apply plus byte-identical sources.
Three fixes here; the deps work is in the orcacad-deps image (see below).
1. kernel-test.sh mounts deps_src. pybind11 is vendored in-tree and CMakeLists
requires its headers; without the mount the container fell back to the
image's baked tree, which predates it.
2. tests/libslic3r/test_3mf.cpp: repair the upstream-merge conflict resolution.
Resolving it as a union dropped the three closing braces of our SCENARIO, so
upstream's SCENARIO opened inside ours ("a function-definition is not allowed
here", plus 12 cascading catch2 registry errors). Restored from the pre-merge
file; whole-file brace balance is now 0 and the case count reconciles as
5 (ours) + 8 (upstream) - 3 (shared) = 10, with both CAD recipe tests intact.
3. tests/libslic3r/test_caddocument.cpp: REQUIRE_CONTAINS / CHECK_CONTAINS.
Catch2 v2 (snaporca) spells substring-match Matchers::Contains; v3 (here)
spells it ContainsSubstring and gives Contains an incompatible meaning,
range-contains-ELEMENT, which fails to COMPILE against std::string. Four
sites had been hand-adapted long ago, but M2-M8 kept porting in un-adapted
Contains calls — 16 of them — and nothing objected because nothing compiled.
Both forks now use the same find()-based macros, so the assertion lines are
byte-identical again and future format-patch ports carry across unchanged.
Five orphaned `using Catch::Matchers::Contains;` lines removed with them.
The deps gap that blocked configure needed five additions on top of
snaporca-deps, built into image orcacad-deps: Eigen 5.0.1, Python 3.12.13
(exact, with Development.Embed), wxWidgets 3.3.2 (was 3.1.5), CGAL 5.6.3
(was 5.4 — mainline's own MeshBoolean.cpp calls CGAL::parameters::default_values,
added in 5.5), plus the pybind11 mount above. OCCT V7_6_0, Boost 1.84.0 and
OpenCV 4.6.0 are pinned identically in both forks and were reused as-is.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
BUILD_VOL defaulted to snaporca_buildcache — snaporca's volume — so running
this fork's kernel test wrote into the other fork's build cache. Defaults to
orcacad_kerneltest now.
docker-iter-build.sh had the identical defect and was fixed to
orcacad_buildcache; this script was missed at the time.
Observed rather than theorised: an orca_cad run under a different deps image
overwrote snaporca_buildcache's CMakeCache.txt, after which snaporca's own
kernel test failed to configure ("Cannot find NLopt library 'nlopt_cxx' in
.../lib/cmake/nlopt/lib") because it inherited the foreign cached paths. No
foreign object files were written — the run died at configure — but the cache
was poisoned, and the volume had to be wiped and rebuilt clean.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Catches the fork up from 449a4cf9fc (2026-06-28) to d6cb667b89 (2026-07-24).
Upstream touched 2326 files; 17 of them overlap the 164 this branch touches.
16 of the 17 auto-merged, including all three CMakeLists.txt, the build_all.yml
CI workflow, and every GUI file. The CAD core never conflicts: CadDocument,
SketchEngine, SketchSolver, McpControl and test_caddocument are files this fork
adds, so upstream does not touch them.
The one conflict, tests/libslic3r/test_3mf.cpp, was purely additive in all three
hunks and is resolved as a union: our test pinning that store_bbs_3mf embeds the
CAD recipe as Metadata/SnapOrca_cad.bin, upstream's multi-nozzle plate-metadata
round-trip tests, and both sets of includes. All three were verified present
after resolution rather than assumed.
NOT BUILD-VERIFIED, for a reason that predates this merge and is not caused by
it: this fork cannot be configured on nativedev at all. Its CMakeLists has
required Eigen3 5.0.1 since before the merge (line 592 pre-merge), while the
only deps image on the machine is snaporca-deps, built for snaporca's
find_package(Eigen3 3.3). CMake fails at configure, so nothing compiles.
That means this fork's Catch2 suite has never run. Every "suite green" figure
recorded for M1-M8 was snaporca's suite; the ports were verified by patch-apply
plus the CAD sources being byte-identical to snaporca's. Building an orca_cad
deps image with Eigen 5.0.1 is what would finally close that gap.
Pre-merge state is preserved at branch cad-mainline-pre-upstream-2026-07-25
(30d54f0074).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Adds scripts/kernel-test.sh: build the libslic3r_tests target and run the CAD
kernel tags, exit code as the whole contract. Clean build 4m36s, incremental
18s, no display needed -- every [CadDocument] case builds a CadDocument,
recompute()s it and asserts on geometry.
Four things this had to get right, each found by it going wrong first:
- SLIC3R_GTK=3 and BUILD_TESTS=ON are mandatory. src/CMakeLists.txt turns
SLIC3R_GTK into 'wx-config --toolkit=gtk<N>', so omitting it asks for
toolkit "gtk", nothing matches, and configure dies with the thoroughly
misleading "Could NOT find wxWidgets" -- while wx-config sits right there
in the deps prefix, working. BUILD_TESTS=ON is what creates the target.
- Configure runs unconditionally. Guarding on "CMakeCache.txt exists" is
wrong because a FAILED configure writes that file too, after which the
guard skips reconfiguring forever and every later run silently reuses the
poisoned cache, ignoring corrected flags.
- A new build volume is always built clean. Cloning a warm cache from
another tree is a correctness trap: rsync preserves source mtimes, ninja
compares them against foreign object timestamps, concludes everything is
current and relinks stale objects. That produced a binary containing NO
[CadDocument] tests at all -- while exiting 0. A green run that tests
nothing is worse than a red one.
- --host builds where the deps image already lives, staged per volume so
parallel workers never share a tree.
The two [known-broken] tags: "entity constraints: tangent/midpoint/symmetric/
angle" aborts inside the vendored solver (slvs/dsc.h FindById, "Cannot find
handle"), and SIGABRT is fatal to the Catch2 process -- that single case took
the suite down at 12 of 31, so a green baseline was unreachable. The thread
groove case is a plain pre-existing assertion failure. Both are excluded from
the dev loop's default filter ONLY; ctest in CI still runs and reports them,
so neither bug is hidden. Baseline is now 603 assertions / 29 cases green,
which is what makes "I broke nothing" a meaningful statement.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Mirrors snaporca-cad a016dffb48 + 4b828c4798 so the two forks do not drift:
- Sidebar on Orca's widget set: ComboBox for all 25 pickers, StaticBox
frames for the tool cards / feature tree / bodies list, framed double
spins, and no stack of full-width action buttons (Prepare's left panel
is parameters only).
- Toolbar: document group + undo/redo + mode-gated tools, commit far
right, ordered via keyed slots, at Prepare's 40 px / 4 px geometry.
- Polygon options and a new Move/Rotate card (distance, axis, angle) in
the sidebar instead of inline in the toolbar row.
- DesignSketchTool: stop consuming the LeftDown over a solid. Orca starts
a rotate drag on the press, so swallowing it killed orbit/pan whenever a
body was on screen. The pick now resolves on LeftUp within 4 px.
DesignSketchTool.{cpp,hpp} were byte-identical across the forks and are
copied verbatim. DesignPanel.cpp needed one hand-merge: mainline's
DropDown is Item-based (DropDown::Item{text,tip,icon}, DropDown(items&))
where snaporca passes parallel vectors, so feat_dropdown/ToolFlyout keep
the mainline form. Checked field-by-field against this fork's
Widgets/DropDown.hpp.
scripts/docker-iter-build.sh: mount the root CMakeLists.txt and cmake/
rather than inheriting the baked copies, which silently drops the
SLIC3R_CAD gate, and stop defaulting to snaporca's build volume - sharing
it made the two forks overwrite each other's cache and binary.
NOT COMPILED. This fork needs Eigen 5.0.1 (find_package(Eigen3 5.0.1
REQUIRED)) while the available snaporca-deps image supplies 3.3, so it
needs its own deps build to verify. The behaviour above was verified only
on the snaporca side.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BVzKmX6Y1aEteit1HTXG4Q
Resolve five conflicts, all of which needed both sides rather than a pick:
- BackgroundSlicingProcess: ours was a pure tabs->spaces reformat of base, so
keep main's per-filament volume/nozzle map read-back (its only change here).
- GUI_App: main's #12506 else-if attached to an `if` this branch deleted;
re-expressed onto the same-agent early-return path (the agent factory caches
per id, so pointer equality is the same predicate).
- MainFrame: both sides relocated Sync Presets independently; keep main's
push_notification plus the branch's Plugins menu items.
- Tab: the "TODO: Orca: Support hybrid" blocks were unchanged base, not a branch
decision; take main's enabled Hybrid to match the already auto-merged siblings.
- test_config: union of both sides' cases (6 plugin + 9 multi-nozzle).
* ci: run unit tests on Windows and macOS
The Unit Tests CI job only ran on Linux, so platform-specific bugs
invisible to a Linux build could land undetected (e.g. the MSVC-only
NfpPlacer crash fixed in #14267). The full non-[NotWorking] suite already
builds and passes on every shipped arch, so this wires them into CI.
A reusable unit_tests.yml, called once per built arch, downloads that
arch's test artifact and runs ctest. Each build leg builds the test
executables and uploads them; a single publish_test_results job on Linux
aggregates the JUnit results into one check.
Coverage: Linux x86_64 + aarch64, Windows x64 + arm64, macOS arm64.
macOS x86_64 is deferred (cross-built on arm64, needs Rosetta).
- build_release_vs.bat: "tests" token enables BUILD_TESTS
- build_release_macos.sh: -T builds and runs tests; ORCA_TESTS_BUILD_ONLY
builds them without running (used by CI)
- scripts/run_unit_tests.sh: parameterized test dir and build config
Addresses #11273.
* ci: bump actions/checkout to v7 in reusable unit_tests workflow
Match the actions/checkout v6->v7 bump (#14517) that upstream applied to
the inline test job this reusable workflow replaces.
* profiles: enforce globally-unique, per-vendor-namespaced setting_id
Many non-Bambu vendors copied Bambu's generic setting_ids (GFSA04 alone
appeared in 1557 files), so setting_id was not globally unique. This
namespaces every vendor's ids and reserves Bambu/OrcaFilamentLibrary space.
- Reserve "G*" (Bambu) and "O*" (OrcaFilamentLibrary) id spaces.
- Assign each other vendor a 2-char prefix (first+last letter, collision
resolved) and renumber every instantiated preset to <PREFIX><NNNN>.
- Strip setting_id from base profiles (instantiation:false) per Bambu's
convention; assign one to instantiated presets that lacked it.
- Remove the pre-existing misspelled "settings_id" key (91 files).
- filament_id is left untouched (it is a per-material id).
- Add one-time migration script scripts/assign_vendor_setting_ids.py with a
persisted registry resources/profiles/vendor_prefixes.json. Re-runs freeze
existing ids; only new vendors/profiles get new ids.
- Bump version in each changed vendor index file.
- Extend scripts/orca_extra_profile_check.py with a CI guard: global
uniqueness, in-namespace, no base setting_id, no gaps, no settings_id typo.
7425 profile files changed across 61 vendors; 0 cross-vendor collisions;
validator clean; migration idempotent. BBL and OrcaFilamentLibrary id spaces
untouched.
* profiles: add setting_id authoring guide for new vendors / profiles
* profiles: drop in-repo README; setting_id guide now lives in the wiki
* profiles: derive setting_id deterministically from vendor/type/name
* bump profile version
SoftFever