Commit Graph

2668 Commits

Author SHA1 Message Date
Dahak Builds 74d1db902e Turkish locale: complete tr.json catalog (#6240) (#7668)
* claim: #6240

* i18n(tr): complete the Turkish catalog with 1280 missing messages

---------

Co-authored-by: Dahak Builds <dahakbuilds@users.noreply.github.com>
2026-10-27 19:22:49 +00:00
github-actions[bot] 882737ff94 Regenerate generated files and fold pending files [skip ci] 2026-10-27 19:52:35 +00:00
Nathan Dane 295d04932f Render workspace: stage bar, aligned floating docks and the Shots strip host (#6281) (#6296)
* claim: #6281

* Render workspace: stage bar, aligned floating docks and the Shots strip host
2026-10-27 19:21:09 +00:00
semmelrinde 0d18757ed7 File dialogs open in the folder used last (#7756) (#7807)
* claim: #7756

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* File dialogs open in the folder used last (#7756)

The desktop dialogs never said where to open, and the Linux ones then
start in the home folder every time. Open, Insert, attach, Save As and
Export now open in the folder of the file last chosen in a dialog, or of
the open design when another design becomes the open one, whichever came
last. Also fixes #7735.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 19:19:10 +00:00
Hainlein4711 9030d2ff19 Ribbon: wrap Japanese button labels across two lines (#7586) (#7669)
* claim: #7586

* UI: wrap Japanese ribbon button labels across two lines at CJK script boundaries (Fixes #7586)
2026-10-27 19:06:07 +00:00
raykholo 38820658fa Body appearance replaces imported STEP face colours (#5769) (#5775)
* claim: #5769

* Let body appearances replace imported STEP face colours (#5769)

A body appearance assigned in the design now replaces the face colours the
body was imported with; face appearances set in the design still win over it.
2026-10-27 19:03:10 +00:00
Erlis Vidal c8c5c882de Sketch: name = value in a drawing tool's value box defines the parameter (#7567) (#7660)
* claim: #7567

* Sketch: a drawing tool's value box takes `name = value` (#7567)

While drawing, a value box (circle diameter, line length, rectangle width and
height, slot, arc, polygon) now accepts `name = expression` and
`#name = expression`, as the Dimension tool already does: the shape is drawn at
the expression's size and its dimension defines and uses the user parameter
`name`, in one undo step.

- sketch.draw.dimensioned checks a definition's right-hand side, read as the
  new parameter will hold it (a bare number in the design's unit), and passes
  the text on to sketch.dimension; the reply lists `defined`.
- The value boxes size the preview and validate by the right-hand side, refuse
  a definition in a box that is not a dimension (polygon sides, rounded
  rectangle corner radius, conic rho), and the status line names the
  parameters defined.
- Guides: parameters and sketching.

Fixes #7567

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 19:02:09 +00:00
semmelrinde 5dc95ef7f6 Viewport: a frame the GPU has no memory for is drawn smaller, not a crash (#7296) (#7398)
* claim: #7296

* Viewport: a frame the GPU has no memory for is drawn smaller, not a crash (#7296)

The viewport's offscreen frame, its see-through targets and the cached
still picture were created with no error scope, so a texture the card had
no video memory for went to wgpu's default handler and closed the app
("wgpu error: Out of Memory" on a GT 740). They are now created under error
scopes; a refused frame lowers the renderer's frame budget and is asked for
again smaller (less supersampling, or the view's own frame drawn smaller and
filtered up), refused OIT targets fall back to sorted see-through surfaces,
and the status line says once that the view is drawn at a lower quality.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Locales: Arabic and Hebrew for the new strings

ar and he are complete catalogs on main now (#3458): every new message
needs them. The pending-messages test compares right-to-left messages in
display order once #7532 lands (#7527).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* GPU memory test: targets take the OIT flag main added (#7296)

The test's two GpuTarget initializers predate the oit field and no longer
compiled once the branch met main.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Locales: Vietnamese for the new strings

Vietnamese became a complete catalog on main (#5823), so every new English
message needs it too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 18:58:31 +00:00
sascha2101 e2d409e6ed Save: a design changed while it is written stays unsaved, and scripts wait for it (#6003) (#7619)
* claim: #6003

* Save: a design changed while it is written stays unsaved, and scripts wait for it (#6003)

Two of the gaps @raykholo found in #6003 were still open on main:

- `engine.script` ran while the window's Save was being written: it calls
  the session directly, past `SolveApp::run`'s file-operation guard, and
  `drain_control` did not hold design-changing requests for a save. They
  now wait for the save like they wait for a file being loaded or picked.
- `adopt_save` marked whatever the session held when the save finished as
  saved. It now takes up the state the save wrote (`done.saved`, and its
  visibility), so a change made meanwhile stays unsaved instead of being
  lost on close without a prompt.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FN5eEMrqWWXqi65FcK8GXq

---------

Co-authored-by: sascha2101 <221609842+sascha2101@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 18:52:36 +00:00
selune 3179cd2ebd Sketch text: cover line pitch and format-12 glyph zero (#6570) (#6574)
* claim: #6570

* test(text): cover nonpositive pitch and format-12 glyph zero (#6570)

Original synthetic font regressions. Focused native font tests, strict scoped Clippy and wasm pass. Full repository gate status is recorded in PR #6574; no CI infrastructure changes.
2026-10-27 18:51:44 +00:00
phanoob d139e415e4 Sketch: switch a circle or arc dimension between radius and diameter (#7499) (#7729)
* Sketch: switch a circle or arc dimension between radius and diameter (#7499)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Sketch: Toggle Radius / Diameter for a circle's or an arc's dimension (#7499)

New command sketch.toggle_diameter, with input and output schemas, and a
context-menu item on a single selected radius or diameter dimension. The
circle keeps its size: the dimension's parameter is halved or doubled (a
plain value is rewritten, an expression is wrapped). Refused, with nothing
changed, when anything else uses that parameter.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Phanoob <phanoob@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 18:49:30 +00:00
semmelrinde d14dbe5e39 Move: a copying Move can be edited from the History (#7741) (#7761)
* claim: #7741

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Move: a copying Move can be edited from the History (#7741)

Editing a feature runs its command on a scratch session with the marker
right after it. For the last feature that marker sits at the end of the
timeline, and place_feature pushed the new feature after it, rolled back,
so a copying Move found no copies and refused the edit. A marker right
after the last feature now moves on past a feature added there.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 18:47:01 +00:00
Igor Bronovskyi 143145751c i18n(uk): add pending translations for missing strings (#7556) 2026-10-27 18:46:33 +00:00
semmelrinde 98658142c0 Delete: deleting one instance of a component takes its joints with it (#7749) (#7782)
* claim: #7749

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Delete: deleting one instance of a component takes its joints with it (#7749)

Deleting an occurrence while its component had other instances dropped
only the occurrence: the joints on it, their motion links, joint origins,
contact set places and exploded moves stayed, so a joint stayed in the
Browser naming an occurrence that was gone. The clean-up that
remove_component_tree did is now Document::remove_occurrences, used by
both.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 18:44:44 +00:00
github-actions[bot] ee6ec09b0a Regenerate generated files and fold pending files [skip ci] 2026-10-27 19:07:17 +00:00
github-actions[bot] 0accea759d Regenerate generated files and fold pending files [skip ci] 2026-10-27 18:59:13 +00:00
HiroKenEz 8962ae10e7 Box dialog: place the box by its corner or centre (#5638) (#6248)
* claim: #5638

* Box dialog: place the box by its corner or centre (#5638)

The Box dialog gets a Placement choice (Corner or Center) with X, Y and Z
fields that take expressions and units, and a Point input: a vertex, a
point on an edge or face, or the origin picked in the view fills them.
The Cylinder dialog gets X, Y and Z for the centre of its base. Left
alone, a new dialog sends no placement, so the box stays at the origin;
Edit Feature shows the stored point and OK without a change keeps it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Box dialog test: pass the active feature to Input::accepts

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Box dialog: Base Center in the complete Arabic, Hebrew and Vietnamese catalogs (#5638)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-10-27 18:29:05 +00:00
semmelrinde 28b27aac8f Saved model: kept only when every body decodes to itself (#6760) (#6761)
The native codec gave one body back with other topology: a pattern copy
cut by an automatic extrude, 10 faces, 18 edges and 10 vertices, came
back with 12 vertices, so the design opened with another body than it
was saved with. Encoded::of now decodes each body it encodes and keeps
the model only when every body comes back with the same face, edge,
vertex and shell counts; otherwise the design is saved without its
model, as when the encoder refuses a body, and opening it evaluates it.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 18:28:41 +00:00
Lars Jansen 5a477b7774 Dialogs: open in the left column in place of the browser and History (#6707) (#6712)
* claim: #6707

* feat(dialogs): open in the left column in place of the browser and History (#6707)

* fix(i18n): Vietnamese messages for the dialog placement option (#6707)
2026-10-27 17:59:00 +00:00
raykholo 236b5c9ebf Add the built-in 1-bit and 1-bit Dark themes (#7550) (#7557) 2026-10-27 17:46:31 +00:00
github-actions[bot] 95ff76cb97 Regenerate generated files and fold pending files [skip ci] 2026-10-27 18:22:08 +00:00
takuto-NA d245541075 Reuse placement-independent model results (#6525) (#6527)
* claim: #6525

* perf(eval): reuse local geometry across placement changes

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* fix(engine): preserve recovery journal through shared refresh

The shared evaluator retains the document Arc, so refresh and body-name
pinning can copy it after command_done recorded its address. The next read
then mistakes that copy for an outside edit and discards the replay.

Record completion after refresh, name pinning or failed-command restoration.
Add a minimal journaled box/inspect regression in the existing crash suite.

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 16:51:09 +00:00
Mateusz Sych 9a7b657fdc Assistant: connect the last-used agent on startup (#7009) (#7124)
* claim: #7009

* Assistant: connect the last-used agent and open the panel at start (#7009)

Two Settings > AI options, off by default: when SolidCraft starts, connect
the agent started last (once the installed agents are known; a missing one
is not swapped for another) and open the Assistant panel. Not for runs
started with --control.

A connection that never gets ready (no agent, start error, the agent ends
while starting) is a new Failed state: the header reads Connection failed
and Connect reads Retry. The control channel reports it as "failed" and
takes connect_at_start / open_at_start.

* Assistant: keep the start-up agent when a design replaces the blank one (#7009)

Opening a design over the blank design SolidCraft starts with reset its
chat and stopped the agent connected at start. A chat not asked anything
yet now starts over in place and keeps its agent. Turning AI features
off also drops a start-up connection still waiting for discovery.

* i18n: Arabic, Hebrew and Vietnamese for the start-up options (#7009)

ar, he and vi are complete catalogs on main now, so every new English
message needs them.

---------

Co-authored-by: MateuszSychXpl <151755914+MateuszSychXpl@users.noreply.github.com>
2026-10-27 16:38:26 +00:00
tr1pp4 d0c594d311 History: a face picked in the view lights the feature that made it (#6910)
Hovering or picking a face in the view lights the feature that made that face (its
creator, as Edit Feature on a face finds it: the hole, not the plate it was cut into)
in the bottom timeline, the History on the left and the Browser's Features folder;
edges and vertices light the feature that made their body, as before. A new face or
edge pick scrolls the history shown to that feature once, through the Find in
Timeline reveal.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 16:25:28 +00:00
Parman Mohammadalizadeh 4515af3eb9 Update Links reads again only the linked designs that changed (#5606) (#5674)
* Update Links reads again only the linked designs that changed (#5606)

Update Links read, evaluated and STEP round-tripped every linked design on every run. Remember a digest of each linked file, of the STEP it gave and of its face names, and skip a link whose file, bodies and face names are all unchanged. Read a linked design through its saved model when the file has one, as opening it does.

* Drop an import the rebase left unused

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 16:01:57 +00:00
github-actions[bot] c5fbf6650c Regenerate generated files and fold pending files [skip ci] 2026-10-27 17:09:41 +00:00
github-actions[bot] 1aaa2cdc2a Regenerate generated files and fold pending files [skip ci] 2026-10-27 17:03:03 +00:00
semmelrinde 98b57a3588 Report a bug: Retake rebuilds the details instead of refusing, and keeps the screenshot (#7350) (#7352)
* claim: #7350

* Report a bug: Retake rebuilds the details instead of refusing, and keeps the screenshot (#7350)

Once anything changed with a report open, every Retake refused with "Settings or model
changed. Close this draft…", and each refused, waiting, timed-out or stale capture deleted the
screenshot the draft had. Retake now rebuilds the technical details for the current view
(keeping what was typed, edited details and a crash draft's body) and re-baselines before it
captures; the screenshot is replaced only when a new frame is accepted.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Locales: Arabic and Hebrew for the new strings

ar and he are complete catalogs on main now (#3458): every new message
needs them. The pending-messages test compares right-to-left messages in
display order once #7532 lands (#7527).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Locales: Vietnamese for the new strings

Vietnamese became a complete catalog on main (#5823), so every new English
message needs it too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 15:17:11 +00:00
bherbruck 35913ced46 Fix flaky dielectric_preflight_and_start test: render a view that does not hit a box edge (#7722)
The default view sends the single pixel's ray through a box edge, which the
raytracer rejects as an ambiguous interface normal. The job then ended Failed
whenever the render thread got ahead of render.cancel (about 7% of runs under
load), and the test only accepted Cancelled or Completed. Render off-axis so
the outcome no longer depends on that race, and print the phase and error if
it ever fails again.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 15:16:48 +00:00
semmelrinde 02981d6bcf Lost GPU: a dropped frame no longer crashes the window's close (#7148) (#7400)
* claim: #7148

* Lost GPU: a dropped frame no longer crashes the window's close (#7148)

On a lost device wgpu's present and discard return early and leave the
acquired frame's image in the surface, holding a swapchain acquire
semaphore. Dropping the surface later (the window closing for the
restart after the loss) then panicked in wgpu-hal. The painter now holds
the frame as AcquiredFrame, which destroys a second handle to the
frame's texture after the frame is presented or dropped; wgpu-core's
destroy of a surface texture takes the image back without checking the
device, and does nothing once a present or discard took it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 14:48:04 +00:00
amolinari-autotex 96b3ebba54 Physical Material: initialize from selected bodies (#7240) (#7241)
* Fix physical material dialog selection initialization

* Fix material selector translations for complete catalogs
2026-10-27 14:40:47 +00:00
semmelrinde 5bd8387cf7 Repeat Last Command in the command palette, with a key you can change (#7077) (#7353)
* claim: #7077

* Repeat Last Command in command search, with a key you can change (#7077)

Shift+R already started the last tool or dialog again (#3045), but it was a
key the app kept for itself: command search could not find it and Keyboard
Shortcuts could not change it. It is now the command edit.repeat, with
Shift+R as its default key. A call through the engine is refused, because
only the app has a tool to start.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Locales: Arabic and Hebrew for the new strings

ar and he are complete catalogs on main now (#3458): every new message
needs them. The pending-messages test compares right-to-left messages in
display order once #7532 lands (#7527).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Locales: Vietnamese for the new strings

Vietnamese became a complete catalog on main (#5823), so every new English
message needs it too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 14:40:36 +00:00
semmelrinde ec877315c3 Sketch: add points to an existing spline (#6533) (#7384)
* claim: #6533

* Sketch: add points to an existing spline (#6533)

A spline could not take more points after it was drawn. The new command
sketch.spline.insert_point inserts a fit point on the curve where it
passes nearest a spot, between its neighbours, or adds a new first or
last point. Existing points keep their indices, so constraints stay. A
right-click on a fit point spline offers Add Spline Point at the clicked
spot. Control point splines are refused for now: with uniform knots one
more control point would reshape the whole curve.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Locales: Arabic and Hebrew for Add Spline Point

ar and he are complete catalogs on main now (#3458). The pending-messages
test compares right-to-left messages in display order once #7532 lands
(#7527).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Locales: Vietnamese for the new strings

Vietnamese became a complete catalog on main (#5823), so every new English
message needs it too.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 14:24:13 +00:00
github-actions[bot] 98d021c523 Regenerate generated files and fold pending files [skip ci] 2026-10-27 16:34:22 +00:00
tr1pp4 04e4e72ea1 History: delete a folder with its features, and suppress a whole folder (#7219) (#7220)
* claim: #7219

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* History: delete a folder with its features, and suppress a whole folder (#7219)

`timeline.delete` and `timeline.suppress` take `groups: [timeline folder ids]`:
every feature in those folders and their nested folders, in one undo step.
A delete takes the folders themselves too; `timeline.ungroup` still keeps the
features. The History on the left offers Suppress/Unsuppress Selected Folders
and Delete Folders and Features… (asking first, like Delete Selected
Features…, when something outside needs them); a folder whose features are
all suppressed is drawn faint.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* History folder actions: Arabic, Hebrew and Vietnamese strings (complete catalogs)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 13:28:16 +00:00
Mateusz Sych 2151bc0497 Drive Joints: say why a rigid joint can't be driven and offer Edit Joint (#6981) (#7051)
* claim: #6981

* Drive Joints: say why a rigid joint can't be driven and offer Edit Joint (#6981)

Drive Joints opened on joint 0 when no joint moved, and its preview ran
joint.drive on that rigid joint, so the dialog showed a red "a rigid joint
does not move" with a Create Issue button.

The dialog now says a rigid joint holds its parts still and how to let them
move (Revolute to rotate, Slider to slide, Cylindrical for both), with an
Edit Joint button that turns it into Edit Joint on that joint. Its commands
refuse a rigid joint, so nothing is previewed and OK stays off. A rigid
joint picked in the browser opens Drive Joints on it to say the same; the
joint box lists only joints that move. joint.drive's own refusal names the
types that move and joint.edit.

* Drive Joints: the rigid joint message in Arabic, Hebrew and Vietnamese (#6981)

Main now delivers ar, he and vi complete, so every new message needs them.

---------

Co-authored-by: MateuszSychXpl <151755914+MateuszSychXpl@users.noreply.github.com>
2026-10-27 13:20:35 +00:00
DuBearte 1a9c4cd99d Sketch: a circumscribed polygon stays regular when dragged (#7395) (#7438)
* claim: #7395

* Sketch: a circumscribed polygon stays regular when dragged (#7395)

Circumscribed Polygon held its sides equal and tangent to the support
circle only. Equal sides round one circle leave every other corner free
(a rhombus round a circle), so a hexagon had 9 degrees of freedom instead
of 4, and dragging its centre or a corner pulled it out of shape (radius
10 became 8.02, angles 174/66 degrees).

Its corners now also lie on a construction circle concentric with the
support circle, which makes it regular as an inscribed polygon is, and
one side is tangent to the support circle, which sets its size: dof 4
(centre, size, turn). One tangent rather than all of them, so no
constraint is redundant.

Test in sketch_ux_tests: 4, 5 and 6 sides, dragged by the centre and by
a corner, stay regular with their sides on the support circle.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 12:47:35 +00:00
Colin Faulkingham b1cf66b628 Fix stored world-axis inputs for revolve (#7179) (#7406) 2026-10-27 12:45:40 +00:00
NOWN-DEV 285fa9433b Sketch: a distance along a line keeps its side (#7506) (#7510)
A point a distance from another point on the same line has two places on it, the same distance
either way. A solve after a value edit or a link move could land on the other one and report the
sketch solved, with no warning: the point jumped to the other side of its anchor, twice the
distance away. Such a distance is now an orientation the solves keep, like an unsigned offset's
side: which way along the line the one point lies from the other. One point must be on the line
by a constraint (an end, a point held on it, or coincident with either); the other may lie on it
by what else holds it (the mirror image of a point on it across a line square to it), read where
it is in the reference state.

Test: a point 3 mm along a linked rail from a point held on it, or mirrored onto it: the image
back past that point counts as a flip, and the rail walked far keeps the side.

Fixes #7506.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 12:43:05 +00:00
sascha2101 f5bfd6d1f2 Sheet metal: a base or contour flange hides its sketch (#6804) (#7151)
* claim: #6804

* Sheet metal: a base or contour flange hides its sketch (#6804)

Only an Extrude or a Revolve counted as using its sketch, so the sketch of a
base flange stayed shown. Its line on the plate's bottom edge lies exactly on
that edge and took the Flange dialog's click: a sketch line makes a contour
flange, a separate flat wall with no bend whose Flip moves it into the plate.

A base or contour flange now uses its sketch as an Extrude does: hidden unless
shown on its own, and shown while its feature is edited.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FN5eEMrqWWXqi65FcK8GXq

* Sheet metal: test that a click on the bottom edge makes an edge flange (#6804)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FN5eEMrqWWXqi65FcK8GXq

---------

Co-authored-by: sascha2101 <221609842+sascha2101@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 12:43:00 +00:00
bherbruck 48c8965853 CI: run the GPU path tests only for render changes, plus nightly (#7742)
The merge queue ran the four-shard GPU path tracer tests (about 56 runner-minutes)
on every group. Now they run when the diff touches render code (crates/render,
crates/gpu/src/path, the render workspace UI, assets/render), or a pull request
carries the new render-tests label (a pull request or a group holding one). The
nightly and manual runs still always run them; a push to main does not.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 11:58:30 -04:00
github-actions[bot] e7e8a1ef26 Regenerate generated files and fold pending files [skip ci] 2026-10-27 15:37:21 +00:00
Matthew Knutson ae530bff44 Typed contracts for the cylinder, sphere, torus, cone, pipe, bounding solid and scale (#2829) (#7145)
* Typed contracts for the cylinder, sphere, torus, pipe, bounding solid and scale

Contracts for solid.cylinder, solid.sphere, solid.torus, solid.pipe, solid.bounding_solid and
solid.scale, with `FeatureResult` output, and the shared types `AxisDir` (calling
`cmd::axis_dir_form`, split from `axis_dir_param`), `NumberExpr` and `ScaleFactors`. The
descriptions say what the evaluation does: the limits it enforces, the real default scope of
`targets`, and where `participants` overrides a key or has no effect. Scale makes no body, so it
declares `name` and `participants` itself. Box's descriptions get the same corrections. The
reply test runs every form in Warn and checks for no `param_warnings`; the guide lists the new
types and adds lessons from the conversion.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Typed contract for the cone

`solid.cone`, added on main beside the primitives this PR converts, gets its contract and
`FeatureResult` output: the cylinder's keys plus `top_radius` / `top_diameter`, with the
evaluation's limits in the descriptions (a radius above 0, a top radius of 0 or more, a height
that is not 0 and may be negative, sizes within 1e7 mm). A `null` top radius stays the handler's
to refuse, since the handler takes it beside a `top_diameter`.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 12:42:16 +00:00
yhamil04 3574956d29 Graphics: the window no longer waits for Direct3D 12 shader compiles (#4860) (#7288)
* claim: #4860

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Graphics: the window no longer waits for Direct3D 12 shader compiles (#4860)

On Direct3D 12, wgpu compiles every pipeline's shaders with FXC when the
pipeline is made, and nothing is kept between launches, so the window
waited seconds for the viewport.

- The GPU tessellation shader, most of that time, compiles on a thread of
  its own. Bodies are built on the CPU until it is ready, then drawn from
  their plans; where it never compiles they stay on the CPU, as before.
- The viewport's pipelines build in parallel, one job per fragment shader
  (solidcraft_gpu::concurrent). A job that errs, panics or cannot start
  runs again on the calling thread, so errors reach the caller's scopes.
  Threads only on Vulkan, Metal and D3D12 with FXC.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 12:41:55 +00:00
semmelrinde 27728b10e2 Trackpad: inertia stops when Shift changes mid-orbit (#7327) (#7361)
* claim: #7327

* Trackpad: inertia stops when Shift changes mid-orbit (#7327)

macOS sends the inertia after the fingers lift as its own scroll bracket,
with the modifiers held at the time. Each move was routed by its own
modifiers, so releasing Shift while an orbit coasted turned the rest of
the inertia into a pan (and pressing Shift turned a coasting pan into an
orbit). A bracket that starts as the fingers' one ends now keeps the
fingers' mode and ends when the modifiers change. Switching between pan
and orbit while the fingers are down works as before.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 12:41:54 +00:00
JK3D 1c109be836 Headless pictures: no highlight left over from the picture before (#7225) (#7326)
* claim: #7225

* fix: headless pictures no longer keep the previous picture's highlight

Capturer::rgba_highlighted left the highlight slot empty when a picture had
no highlight. The renderer keeps its last highlight scene while the slot is
empty, so a selected edge's band drawn for one picture showed in the next
one on the same capturer. The depth tests share their capturers across test
threads, so they failed at random: an edge "192 of 612 px drawn", a washer
showing through its plate, a cut showing what lies behind it. A picture now
always gets a highlight scene of its own, empty without one.

Clearing the leak showed that a_selected_edge_on_a_grazing_face_draws_whole
passed at 88 degrees in a design of radius 30 only by counting the leaked
band: its reference line lies 143 mm behind the sheet's middle, past that
design's far plane, and drew nothing. The small design now holds the edge.

New test: a_picture_without_a_highlight_shows_none_of_the_one_before.
Full `--test depth` runs on Apple silicon: 10 of 10 failed before, 0 of 20
after.
2026-10-27 12:41:48 +00:00
JK3D 388266ca37 Rename: an imported body with numbered siblings is renamed without evaluating (#7274) (#7277)
* claim: #7274

* Renaming an imported body with numbered siblings evaluates nothing again

Since #2467, Model::rename_body evaluates the design whenever another body's
name is worked out from the renamed one's (`X (2)`). Every STEP import whose
solids share a label has such names (`Part`, `Part (2)`, ...), so renaming
one of its bodies re-read the import and meshed every body again.

body.rename keeps the names of a body's siblings and of later bodies named
after it, so for a body no later feature changes (the same mesh from the
feature that made it to the end), renaming it in the model gives the names
evaluating the renamed design gives. That case is followed again, as before
#2467; bodies later features change still take #2467's checks.
2026-10-27 12:41:39 +00:00
semmelrinde 2346c1ae81 Web app: Report a bug saves its screenshot as a download (#7192) (#7197)
* claim: #7192

* Web app: Report a bug saves its screenshot as a download (#7192)

Save screenshot… wrote the PNG with image's save, that is std::fs, which
the browser doesn't have. The screenshot is now encoded in memory and
written through vfs, as every other save is, so the web app hands it out
as a download. Natively the same file is written and ui.screenshot
replies as before. A failed save says why in plain text instead of a
Rust debug value.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Web app: a screenshot without a path no longer panics (#7192)

`ui.screenshot` without a `path` saved to `std::env::temp_dir()`, which panics on
wasm32-unknown-unknown (std has no file system there), and the web app's control
requests come through `window.solidcraft.request`. `vfs::temp_dir()` now returns
`vfs::BROWSER_TEMP`, a folder of the in-memory files, in the browser, and
`save_screenshot` and `vfs::design_folder_refusal` (reached when a request names
`allow_paths` while a design is open) use it. Natively nothing changes.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Web app: keymap and theme exports download (#7192)

Both exports wrote with `std::fs::write`, which always fails in the browser. They
now write through `vfs::write` like the other saves, so the web app hands the file
out as a download. Natively the file is the same; an agent's folder scope now
applies to them as to every other write.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

* Tests: say what the folder-scope checks of the saves show (#7192)

The screenshot test wraps the write in a folder scope by hand: that proves the write goes through vfs, not that ui.screenshot's write runs inside the request's scope. ui.screenshot checks the agent's folders when the request arrives; the PNG is written in a later frame, outside that scope, as on main.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 12:41:21 +00:00
raykholo 8372c476ab Add the built-in CGA theme (#7593) (#7594) 2026-10-27 12:41:08 +00:00
Liu XingJian 4ad285570c Control channel: retire a subscription when the events thread ends (#7226) (#7336)
The seat of a subscribed connection on the DesignWatch lived in the
connection's serve thread. It left only when that thread saw the end of
the stream, so one unbounded blocking read stood between a closed client
and the count. Under load (GPU backend probing and crash-dump children
on Windows) that read can sit for seconds; the test's 10 s deadline then
expired and `!WATCH.subscribed()` failed (#7226).

Move the seat onto the events thread. The thread ends on `stop`, or on
the first write a closed client refuses, so the next publish retires the
seat. The test now publishes while it waits, and its wait no longer
depends on the serve thread.
2026-10-27 12:40:41 +00:00