Commit Graph
4576 Commits
Author SHA1 Message Date
Mateusz SychandMateuszSychXpl e81a201ff8 Drawing W06a: parallel hidden lines and shared work limits (#2455) (#3013)
* geom/hlr: per-curve work on worker threads, shared work limits

Silhouette tracing, screen sampling, shadow cuts, classification, coverage and
the crossing search (by grid rows) with crossing refinement run on worker
threads through one ordered map: results go back in item order, so drawings
are the same bit for bit on any number of threads. Options::threads picks the
workers (0 = one per core, small views stay on the calling thread); the
browser build stays serial. Placements sharing a body and a turn share one
silhouette tracing.

Cancellation is polled before every item and inside long ones from every
worker; the first error stops the others; a panicking worker becomes
HlrError::Internal instead of escaping. Ray, curve, crossing, shadow and
coverage budgets are totals over all threads (shared counters; crossing pairs
are counted before the search).

Limits: placements and the default max_tracks are hlr::MAX_ITEMS, which is
section::MAX_ITEMS (also the brep record limit); trim_samples must be 1..=4096.

* docs/drawing-hlr: link the parallel-work PR

* geom/hlr: clearer limits and parallel-work docs, faster thread tests

docs/drawing-hlr.md now separates input limits (with the error each returns:
body records fail with Invalid, not Budget) from the work budgets, names every
budget's error, and states how errors and panics behave on worker threads.
Code comments say the same in plainer terms. The budget and cancellation
tests drop the oblique torus, and the freeform pass of the identity test
compares one parallel run, which keeps the HLR suite under 4 s.

* geom/hlr: poll cancellation inside long items, find spline spans by bisection

Review found two items a worker could run for seconds without seeing a
cancellation: tracing a body with many curved surfaces, and sampling one long
spline (up to 4096 control points). Silhouette tracing now asks before each
surface, screen sampling at every refinement pass, the cuts at left-out faces
every 4096 chords and coverage every 65536 checks.

The long spline was slow because every point and every part hull scanned all
of its knot spans. de Boor evaluation now finds its span by bisection over the
ordered knots, and a part's hull starts at its first piece and stops after
its last: the same spans, so the same bits (checked on the fixture scenes),
and a 4096-point spline edge prepares in 0.9 s instead of 2.2 s and samples
in 0.23 s instead of 1.47 s.

* geom/hlr: poll cancellation while forming and trimming silhouette sets

Review found that forming a body's silhouette sets scanned each face's whole
surface group before the first poll, quadratic in the faces on one surface
(about 4e8 steps for 20 000 faces). The group scan now stops at its first
supported face (groups list faces in order, so that is the same face), the
loop polls every 256 faces, and trimming a traced curve asks before each
face. A cancellation seen while trimming drops the partial curves, even if
the callback answers false later. Output unchanged on the fixture scenes.

* geom/hlr: poll cancellation per boxes met when cutting at left-out faces

Review found that one chord crossing the screen boxes of very many left-out
faces bisects every side without a poll (polls came only every 4096 chords).
It now also polls every 1024 boxes met. The docs state the cancellation
contract precisely: polls are where listed, and the work between two polls is
bounded by the input limits rather than by a time.

* docs/drawing-hlr: say where coverage polls cancellation

* geom/hlr: body records have brep's own, larger limit now

Main raised brep::MAX_ITEMS to 1 000 000 (#3093, which stores whole evaluated
models), so it no longer equals section::MAX_ITEMS. hlr::MAX_ITEMS stays
section's 100 000 for placements and the default curve budget; the docs and
the limits test now say that a body's records are bounded separately, by
brep's limit.

* geom/hlr: a panicking worker stops the others; the ray budget is HLR's own

A worker that panics now sets the shared stop flag while it unwinds, so the
other workers end the phase early instead of finishing work the call throws
away. The docs no longer compare the ray budget with the occlusion work
budget of file.export_view_dxf: the two count different work.

* geom/hlr: an ignored timing test, prepare and drawing timed apart

hlr_timing prepares four bodies, then draws 100 differently turned placements
on one thread, two workers and one worker per core, printing the total, the
phase timings and the ray count. SOLIDCRAFT_BENCH_HLR sets the placement count.

* geom/hlr: timing scenes for sharing and crossings; placements keep their lines when permuted

hlr_timing now starts its prepare clock after the fixtures are built, adds the
same grid at one turn (shared definitions against one per placement) and
sixteen thin bars crossing 8x8 from above, rotates the thread order over three
rounds and checks the drawings match across thread counts outside the timed
part. A new test permutes three placements of one prepared box and removes one:
each line keeps its own placement index, and the boxes' lines stay exact.

* docs/drawing-hlr: the review's limits and fixtures in the parallel tables

* geom/hlr: follow the B-rep record into solidcore-record

The record, its version and its item limit now come from solidcore-record
(#11496): the new tests build their bodies at `solidcore_record::brep::VERSION`
and make a spline's curve with `C3::bounded`, as the other fixtures do, and
the limits table gives the record's own limit, 4 000 000 items, checked by
`brep::check::validate`.

---------

Co-authored-by: MateuszSychXpl <151755914+MateuszSychXpl@users.noreply.github.com>
2026-11-04 16:56:07 -05:00
CamWightandClaude Opus 5.5 96bb41c480 Drawings: notes on a drawing, the record and commands (#4896) (#11898)
* Issue a revision: drawing.revise (#3874 R1)

drawing.revise appends a copy of a row to the issued history: an existing row as
it reads, or a new revision with the texts given, in one undo step. A row is
issued once. The copy keeps what it read through an edit, a hide or a delete, and
nothing is filled in: no date, checker or approval.

Issuing used to make a drawing unreadable, as a copy has its row's id and the
drawing's check refused any id seen twice. Each id now counts once, and issued
ids are unique among themselves.

model_version becomes {path, change}, on issued copies only. One function on the
session answers it, and answers nothing until version control exists.

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

* Drawing revisions: numbered 1, 2, 3 (#3874)

A drawing's new revisions are numbered 1, 2, 3, never lettered
(@CamWight, 3 Nov). drawing.set's revision_numbering is {start: 0 or 1},
the first number. A revision typed by hand is kept as typed; the next one
follows the highest number, and drawing.inspect says where numbering
starts. A record in letters is refused (drawings haven't shipped).

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

* Revision tags on a sheet: add, edit and delete (#3874 R3)

A tag gains the sheet it sits on and its place there. drawing.revision.tag.add,
.edit and .delete keep them, inspect and every revision reply list them with what
each shows, and a sheet's delete counts and takes the tags on it. A deleted row
still leaves its tags showing the revision they had. Tags draw on paper once scene
text lands.

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

* Saved tables on a sheet: drawing.table.add, .edit and .delete (#3874)

A drawing saves its parts lists and revision tables, each on a sheet with its
definition and place. Rows are never saved. A parts list starts at the top right
listing what its sheet documents, else the design; a revision table at the bottom
left; headings on top, QTY, rows of 7 mm.

Placing a list saves the numbers its view shows, once per view, so what is printed
doesn't move. Numbers are the design's: undoing the placing puts back only the
drawing, as every drawing step does, and the numbers stay. That one command changes
both, and the proof that nothing needs evaluating now allows it.

The tables are a ById list, as the balloons and images are, and keep its rules: at
most 1000, listed by id. Revision tags take their own id type.

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

* A balloon reads its row in a parts list (#3874 L1)

A balloon can name the parts list it reads; else it reads its sheet's only list,
else the drawing's only one, else none, with the choices listed. Its row's item is
worked out on each read and never saved: drawing.balloon.inspect gives item and
row_status (ok, no_list, reference, no_row or not_found). A lost row never removes
the balloon. A list's delete reports the balloons that named it and clears their
links.

Bom::item_of cuts a path from the top of the design to start below the assembly
listed, and row_of is no longer test-only.

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

* Notes on a drawing: the record and commands (W19e, #4896)

A note sits on a sheet, a view, or an edge or corner of a view's model, saved by
name as a dimension's is, with the title block's text and fields, its place and
its text height. drawing.note.add, .edit and .inspect, one undo step each, no
model evaluated. Its one leader is worked out, never saved: a lost reference keeps
the note and its text, with no leader. Anchor is one type so other symbols placed
on a sheet can use it.

A view's notes go with its copy and delete; a sheet's with the sheet. The plan left
delete for later, but selection.delete now handles a drawing's objects, so notes
join it. Balloons and notes share Lookup::middle. Drawing them waits for text in
the paper scene (slice 12).

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 16:53:53 -05:00
bherbruck ca6a095795 Merge pull request #12087 from tangency-labs/sync/main-into-drawings-1104
Sync main into drawings (Nov 4)
2026-11-04 15:51:54 -05:00
bherbruckandClaude Opus 5.5 0273441236 Sync: the three dimension functions that read the display mesh, on drawings only
Picker::placed, face_corners and BodyEdges::new decide with the display mesh.
They are listed on this branch only and must be cleared before drawings lands
on main, where the list takes no new lines.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 15:24:44 -05:00
bherbruckandClaude Opus 5.5 58a00f41be Sync: the repository checks know the drawings' commands and modules
The kernel rig's coverage check exempts the drawing and bill of materials
commands, each naming where it is tested. The mesh decisions check follows the
view line modules to mesh_lines/, and skips the two that only draw and count a
drawing's view lines.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 15:23:15 -05:00
bherbruckandClaude Opus 5.5 b841eb939f Merge main into drawings (Nov 4)
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 11:39:05 -05:00
PeppaWutz 8f67c7fb95 Sketch: constraint icons stay while the sketch fills the view (#10518) (#11399)
* Sketch: constraint icons stay while the sketch fills the view (#10518)

Constraint glyphs hid as soon as a millimetre of the sketch plane drew smaller than two pixels
(#8980), whatever the sketch's size, so a 500 mm sketch zoomed to fit lost every glyph although
it still filled the view. Below that scale the glyphs now hide only while their anchors span
less than twenty glyphs (300 px) on screen: a small sketch far away still loses them, a large
one that fills the view keeps them. The preference "Hide constraint glyphs when zoomed out" is
unchanged; off, glyphs stay at every zoom as before.

Test: constraint_glyphs_hide_when_zoomed_out adds a 500 mm sketch that keeps its glyphs zoomed
to fit and loses them far away; glyph_zoom_tests covers the span threshold.

* Sketch: glyph span measures only drawn glyphs and stops at the threshold (#10518)

The span check now walks the same constraints the glyph pass draws (one MidPoint glyph per shared midpoint, a pattern's constraints once) and returns as soon as the bounds reach MIN_GLYPH_SPAN_PX, so a large sketch zoomed out costs no more than before.

* Sketch: filter glyph constraints lazily, midpoints in a set (#10518)

* Sketch: measure the sketch, not its glyph anchors, before hiding glyphs (#10518)

* Sketch: collapse the midpoint check into one if (#10518)
2026-11-04 12:06:26 +00:00
CamWightandClaude Opus 5.5 49eedf122c Drawings: a reference to a suppressed or failed feature says so (#4892) (#11823)
* Labelled measurements on a view's snapshot (#5866)

measure: linear, aligned, angle from model points, radius from an edge, with
units and decimals for the labels. No new maths: drawing_measure's rules, and
a dimension's radial rule through one Lookup::whole_arc. Labels go through one
function, reading_label, which dimensions' text now uses too.

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

* Pieces on one circle are that circle, and one piece is that edge (#4892)

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

* A dimension between two flat faces (#4892)

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

* A suppressed or failed feature's references say so, not lost (#4892)

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:40:17 -05:00
CamWightandClaude Opus 5.5 669aaa7744 Drawings: a dimension finds its circle again after the edge is split (#4892) (#11869)
* Labelled measurements on a view's snapshot (#5866)

measure: linear, aligned, angle from model points, radius from an edge, with
units and decimals for the labels. No new maths: drawing_measure's rules, and
a dimension's radial rule through one Lookup::whole_arc. Labels go through one
function, reading_label, which dimensions' text now uses too.

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

* Pieces on one circle are that circle, and one piece is that edge (#4892)

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:40:10 -05:00
CamWightandClaude Opus 5.5 a3375c58e3 Render: a triangle tests only the samples of each row it can touch (#5624, W17b) (#11042)
The raster walked a triangle's whole box, testing every sample, and the pixel budget counted that box.
A plain plate's few long triangles have boxes that are mostly misses, so a view at 2:1 counted 18 tests
a sample. Now each row of the box tests only the span the triangle can touch there, and the budget
counts those spans (an empty row costs one unit, as the walk does).

row_span works the span out from each edge's weight along the row, allowing for the test's own
tolerance and the rounding of its arithmetic, so it is never short; sample_weights is the raster's test
pulled out unchanged. Pictures are the same to the byte. A product image shares the raster and the
count, so its 128 M admits more; the tests built on its old count hold more triangles.

With the misses gone nearly every counted test paints, and a painted surface costs about 240 ns a
test on one thread, so a picture's limit falls from 400 M to 250 M to stay near a minute at worst.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:12:20 -05:00
CamWightandClaude Opus 5.5 a301359652 Dimensions on the sheet: lines and arrowheads in paper millimetres (#4896) (#11867)
Each dimension of a view on the sheet is one group, ObjectKind::Dimension,
after the views: extension and dimension lines and filled arrowheads
(Terminator::place, #9466), laid out by dim_layout_sized in paper millimetres, so it is the same size at any
view scale. All seven kinds draw; a chain is one line of frames at one offset
with each extension line once, an angle sits where its lines meet on the view,
and a cylinder's line runs across its silhouettes at the end on its offset's
side.

drawing.scene and drawing.export list each dimension: its status, whether it
is drawn, and why not. A lost one, or one on a view that isn't drawn, has no
lines and leaves the sheet and the file written. The provider keeps each
dimension's group by its record, world state, view drawing and units: a moved
view moves them unread, and an edit reads that dimension alone, in one short
step after the views. Text joins with B1.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 04:53:24 -05:00
github-actions[bot] 6bca7d2854 Regenerate generated files and fold pending files [skip ci] 2026-11-04 08:13:26 +00:00
eeb686f164 Measure: a circle and a line now show 'Line to centre' instead of 'Center to center' (#11524)
* Measure: a circle and a line show 'Line to centre' instead of 'Centre to centre'

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* Measure: US spelling for Line to center, UK in en-GB

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

---------

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 07:38:21 +00:00
9e3e403b3f Browser and History: keep row tools working when either panel is hidden (#11303) (#11416)
* claim: #11303

* Keep Browser and History row interactions independent

* Browser tests: use solidcore_record::Vec3 after the layer split

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>
2026-11-04 07:15:40 +00:00
dudedagger42andClaude Opus 5.5 0231eec4c6 Nix flake: zenity on the app's PATH, for file dialogs without a portal (#2177) (#5183)
rfd opens file dialogs through the XDG desktop portal and falls back to zenity
when there is none. The flake's package didn't bring zenity, so on a system
without a portal or zenity, Open and Save showed nothing. The wrapper now adds
nixpkgs' zenity at the end of PATH, after the system's own.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:53:37 +00:00
erman 69f202f1c8 Measure: show complete point coordinates (#11069) (#11085)
* Measure: show complete point coordinates (#11069)

Render measurement item details as wrapping text instead of fixed-height cells. Keep the existing values, units and precision. Add a regression test in the existing measurement UI test module.

Local validation: the regression test compiled, but the Windows GNU test binary crashes with 0xc0000005 even for --list, before tests run. Runtime verification and visual evidence remain pending; this contribution is a draft.

* Keep repository LF line endings in the coordinate fix
2026-11-04 06:53:04 +00:00
1cbfb889d5 Agents and scripts: Compare Bodies describes its exact parameters and reply (#10994) (#10995)
* claim: #10994

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

* Compare Bodies: input and output schemas (#10994)

inspect.compare gets a typed contract and reply types in
schema/compare.rs and leaves unconverted.txt. Each key reads as the
handler reads it, with its message; `color` checks each colour with the
handler's own color_param, now pub(crate) and re-exported from cmd. The
reply is built from Comparison and its parts, with the same JSON.

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

* Compare: drop the json import the typed reply no longer uses

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

* Compare schema: colour strings carry the hex pattern

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

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 06:45:10 +00:00
ee3cc32525 Settings profiles: export and import stay inside an agent's folders (#11362) (#11390)
* claim: #11362

* Settings profiles: export and import go through vfs, inside an agent's folders (#11362)

settings.profile.export and settings.profile.import read and wrote their
path with std::fs, so the folder scope an MCP request's allow_paths sets
up never saw it: an agent could write a profile anywhere, or probe any
path. Both now go through vfs (len, read_to_string, write), as file.export
and ui.screenshot do. requests_with_allow_paths_stay_inside_them covers
export and import inside and outside the agent's folders.

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

---------

Co-authored-by: thocsmith <193096300+thocsmith@users.noreply.github.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:28:11 +00:00
Colin Faulkingham 954232cef1 AI client setup: preserve configuration files when a write fails (#11363) (#11410)
* claim: #11363

* AI client setup: replace configuration files atomically (#11363)
2026-11-04 06:27:37 +00:00
github-actions[bot] 9092df9b97 Regenerate generated files and fold pending files [skip ci] 2026-11-04 07:25:51 +00:00
semmelrindeandClaude Opus 5.5 c1a23f8f97 Section Analysis: the drag arrow is back when the dialog is reopened (#11423) (#11432)
Edit reopens the dialog from the section's own geometry (Show viewport
section status, on by default): the plane is kept as numbers and nothing
is picked. The arrow was placed from the first pick only, so it was gone.
With nothing picked, it now stands on the kept plane, its tip at the cut.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:11:46 +00:00
austin62jakeandClaude Opus 5.5 0fc4cd2747 Agents and scripts: the render stage commands (Ground, Backdrop, Clear) describe their exact parameters and reply (#11455) (#11456)
* claim: #11455

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

* Render stage: input and output schemas for Ground, Backdrop and Clear (#11455)

The three handlers' strict parameter types move to schema/render_stage.rs as
GroundParams, BackdropParams and ClearParams, with descriptions and
JsonSchema, and become the contracts; the handlers read the same types. The
handlers build the shared StageSaved reply instead of json! literals, and the
saved setup is serialized with an error instead of json!'s panic path.

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

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 06:11:09 +00:00
leValdemar19andClaude Opus 5.5 6d83f95fba Isolate a body: other components are hidden as components, in one step (#7896) (#7951)
* claim: #7896

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

* Isolate a body: other components are hidden as components, in one step (#7896)

Isolating a body hid every other body one by one: browser.visibility with `items` writes only
hidden_bodies, so every other component's Browser row kept an open eye and normal text, and
its eye could not bring it back (the first click hid an already invisible component).

browser.visibility takes `isolate: true` with `items` (bodies), as it does with `paths`. The
picked bodies' own placements stay open, their other bodies (and the root's) are hidden one by
one, and every other placement is hidden as a placement. Isolate in the Browser and the view
runs that one command, so it is also one undo step instead of two.

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

* Isolate a body: a hidden Bodies folder or group it is in shows again (#7896)

Isolating a body inside a hidden Bodies folder or group left that folder or
group hidden, so the picked body stayed hidden. Isolate now shows the
folders and groups (with the groups they are in) that hold the picked
bodies; their other bodies are hidden one by one as before.

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

* Isolate a body: look for hidden folders only when some are hidden, and find body owners once (#7896)

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:06:23 +00:00
bf625dff05 Box, Sphere and Torus: clicking a point of a sketch places them there (#9403) (#9405)
* Box, Sphere and Torus: a sketch point picks where they go (#9403)

The Point input of the Box, Sphere and Torus windows took a vertex, a point on an edge or face
and the origin, but not a sketch's points, so a sphere could not be centred on a point drawn in a
sketch. It now takes the points of every shown sketch (POINTS | EXTERNAL_POINTS, as Measure
does): a click fills X, Y and Z with where the sketch puts the point in the world, and a point
selected before the window opens places the primitive there, as a vertex does. The input's chip
names the point and its sketch as Select Other does ("Point 2 · Marks").

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

* Box, Sphere and Torus: the sketch point chip label for every complete locale

Main's i18n check wants each message in the complete catalogs (ar, he, ja,
vi, zh); the label has no words, so each carries the same "{point} · {sketch}".
With main's new selection kinds and this one, every Sel variant has its arm,
so the catch-all arm is unreachable (clippy) and goes.

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

* Primitive at a sketch point: a point picked in the sketch being finished still places it

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

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 06:03:39 +00:00
semmelrindeandClaude Opus 5.5 ed3510a91f Typed schemas for the design properties commands (part of #2829) (#10920)
document.metadata, document.metadata.set and document.metadata.remove
get input and output schemas: contracts in schema/metadata.rs, checked
in front of the unchanged handlers, and reply types the handlers now
build their replies from. Their lines leave unconverted.txt.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:03:21 +00:00
github-actions[bot] 8f0130a61b Regenerate generated files and fold pending files [skip ci] 2026-11-04 06:39:31 +00:00
github-actions[bot] 72df682af7 Regenerate generated files and fold pending files [skip ci] 2026-11-04 06:27:43 +00:00
Davit MatchakhelidzeandClaude Opus 5.5 41a450be7c i18n: add Georgian (ka) interface language (#11501)
A Georgian catalog (assets/locales/ka.json) carrying every English message,
registered in crates/ui-egui/src/i18n.rs, with a row in the new-language test,
a pending attribution file and a terminology note in docs/localization.md.
The catalog is Mkhedruli only; egui's built-in Hack already draws it.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:00:45 +00:00
HiroKenEzandClaude Opus 5.5 b227924be9 Render: the Cancel button keeps its full label during a render (#11222) (#11241)
* claim: #11222

* Render: the progress row keeps Cancel whole (#11222)

The progress card is an egui area, laid out within last frame's size. When
the status grew (9 % to 10 %, a longer phase or name), the caption and the
Cancel button were truncated, down to "..." with a long shot name.

progress_action now takes the name and the status apart and gives each part
an explicit width from the room the caller offers (the canvas for the
floating card): Cancel and the status always show whole, a long name is cut
with an ellipsis (full name on hover), then the meter narrows.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 06:00:31 +00:00
Mateusz SychandMateuszSychXpl 28b47059bb Agent chat: a text selection dragged past the edge scrolls the conversation (#11527) (#11533)
* claim: #11527

* Agent chat: a text selection dragged past the edge scrolls the conversation (#11527)

Holding a text selection near or past the top or bottom of the conversation
scrolls it, faster the further out, so a long answer can be selected in one
drag. egui scrolls a label's selection only for the label under the cursor, and
not at all for a table, whose label sits in its own horizontal scroll area.

egui also leaves a label out of the text selection while it is scrolled out of
view: a selection whose start had scrolled away was dropped, and Copy missed
the hidden text. The conversation's labels now take part out of view too,
clipped to nothing, so they keep their end of the selection and give their
text to Copy without being drawn or moving the selection.

* Agent chat: only a drag towards the edge scrolls the conversation (#11527)

A press on a label creates an empty selection on the press frame, so a click,
a double click or a held press on the line by the edge scrolled the
conversation, and so did the start of a drag from there inwards. Scrolling now
waits until the pointer has moved more than 8 points towards that edge since
the press.

* Agent chat: approval titles and pending messages keep a selection scrolled away

They were plain selectable labels, so a selection that started in one was dropped
once its label scrolled out of view, as in the rest of #11527. They now use the
same off-screen-aware label as the conversation's text.

---------

Co-authored-by: MateuszSychXpl <151755914+MateuszSychXpl@users.noreply.github.com>
2026-11-04 06:00:28 +00:00
Mateusz SychandMateuszSychXpl a19a32ddd7 Interference: Compute checks in the background, so the window no longer freezes (#11483) (#11541)
* claim: #11483

* Interference: intersect and mesh each overlap on a kernel worker (#11483)

Compute in Inspect > Interference intersected each overlapping pair and
meshed the shared volume on the window thread. Its stack is 1 MiB on
Windows, too small for the kernel's recursion, so the app could close
without a crash message. Each pair now runs through
kernel::map_with_local_pools, on a worker with the kernel's 8 MiB stack,
as inspect.interference and inspect.clearance already do.

* Interference: Compute checks off the window thread (#11483)

On 0.30.0 the report's two parts took the window thread 95 s (x64) to
154 s (x86) in the overlap boolean, and the window stopped responding
until the app was closed. Compute now runs inspect.clearance and the
shared volumes on a copy of the session, on a kernel thread, when the
app runs commands off the window thread (as Find Limits does). The
dialog shows the elapsed seconds with Cancel meanwhile, and drops the
reply of a design changed since.

* Interference: Cancel on its own row while Compute checks (#11483)

---------

Co-authored-by: MateuszSychXpl <151755914+MateuszSychXpl@users.noreply.github.com>
2026-11-04 06:00:15 +00:00
Homero Thompson 606e264f6d UI: the Dependents button of the INSPECT HISTORY ribbon no longer fails without a feature (#11642) 2026-11-04 05:59:52 +00:00
semmelrindeandClaude Opus 5.5 d25f8198fa Pieces of a cut keep their names when the design is evaluated again (#11658)
A cut that leaves a body in pieces names them after their faces
(`order_lumps`): the piece whose least face name sorts first takes the
first name, and so on. It moved the names but left the pieces where the
kernel put them. The session then pinned the feature's body names in
that state order, and the next full evaluation (opening the design, or
editing a parameter upstream) handed them out again in a different
order: two pieces swapped names, and a later feature that named one met
the other. Found by a local model fuzzer.

The pieces now also take the family's places in name order, so the
names pinned are the order the next evaluation hands them out in. A
design saved before keeps the names a fresh evaluation gave it.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:59:42 +00:00
Hainlein4711 464f4ac748 Intel OpenGL: restore the grid and sketch axes (#8053) (#11573)
* claim: #8053

* Intel OpenGL: restore the grid and sketch axes
2026-11-04 05:49:15 +00:00
semmelrindeandClaude Opus 5.5 03b5248dc1 Sketch: Line and Construction Line refuse a second point next to the first (#11595)
A through-point 0.0000001 mm from the start made an infinite line whose
direction nothing held, and a Line that short had both ends connected to
the origin, a line of length zero. Both now refuse points closer than the
distance at which a new point is connected to an existing one.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:49:05 +00:00
Colin Faulkingham f9dd77e383 Configurations: reject invalid suppression values when saving rows (#9088) (#11571)
* claim: #9088

* Configurations: reject invalid suppression values before saving rows (#9088)
2026-11-04 05:39:04 +00:00
semmelrindeandClaude Opus 5.5 4afd24d29f Design tabs: the tooltip says when a saved design has unsaved changes (#11408) (#11590)
* Design tabs: the tooltip says when a saved design has unsaved changes (#11408)

A tab marks unsaved changes with a bullet after the design's name, and a
long name cut short by the tab hides it. The tab's tooltip, which says
where the design is saved, now adds "Unsaved changes" below for a saved
design with changes not saved yet. A design never saved already says
"Not saved yet". The new string is in a pending locale file, with ten
translations taken from the same words elsewhere in the catalogs.

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

* Design tabs: the unsaved line in every complete catalog (#11408)

Arabic, Hebrew and Vietnamese are complete catalogs, which must carry
every English message: they get the words they already use for
"Unsaved changes" elsewhere.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:29:03 +00:00
kaltenbrunnermarkusandClaude Opus 5.5 8a74336713 Tests: the autosave recovery test no longer fails at random on a busy machine (#11815)
* claim: Recovery: the save test waits for the autosave worker before it checks

* Recovery: the save test waits for the autosave worker before it checks

saving_drops_the_entry_and_a_clean_close_leaves_nothing marked the
session saved right after the tick that started an autosave. Under load
the worker was still writing, so the next tick found the write running,
took nothing up and dropped nothing; the file landed afterwards and the
check that it is gone failed (seen in CI on #9021).

The test now ticks, bounded, until the autosave has caught up
(Autosaver::caught_up), as the app's own frames would, before it marks
the session saved. A small helper in the test module does this. The
app's behaviour is unchanged.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:22:20 +00:00
Colin Faulkingham 5e4c52a7cb Sketch Transform: keep the applied rectangle at the previewed rotation and size (#11617) (#11620)
* claim: #11617

* Sketch Transform: preserve dimensioned geometry on apply
2026-11-04 05:21:27 +00:00
github-actions[bot] 77e0e11775 Regenerate generated files and fold pending files [skip ci] 2026-11-04 05:42:40 +00:00
7bd66b7c06 Sketch dimensions: enter the value in a small window with step arrows (#7246) (#7251)
* claim: #7246

* feat(sketch): enter a dimension's value in a small window with step arrows (#7246)

* Dimension window: fit the layer split, and keep the wheel off the camera over the window

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>
2026-11-04 05:12:31 +00:00
DanielandClaude Opus 5.5 86db05e538 Sketch: regression test for a Line started on a snap point (#11171) (#11322)
* claim: #11171

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

* Sketch: test that a line starts on a snap point of an existing shape (#11171)

The Line tool's first click on a line's midpoint, a circle's centre or edge,
a rectangle's edge or a point part-way along a line starts the line there and
ties its start to that shape. In 0.30.0 and 0.30.1 the start was dropped on the
next hover (fixed by #11184); this covers each shape the reports name.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:11:14 +00:00
ABDULLAH d912f3533b Arabic: better wording for Appearance, Graphics, Assistant, Layout, Sketch and Navigation preferences, thread drill notes and other messages (#11358)
* Arabic: viewport term in the Appearance section heading, desktop app badge, deeper-overlap note in Find Limits

* Arabic: drop the stray direction mark and match the Refresh wording in the old Node.js note

* Arabic: tanween in the thread fit hint, tap drill wording, crease term in continuity, spline label and setback wording in the fillet help

* Arabic: sweep and boolean terms in the body check note

* Arabic: badge term for the status chips note and summary wording for Fold items in Graphics preferences

* Arabic: tanween, catalog terms and clearer labels in the Assistant, Design and Tools preferences

* Arabic: tanween in the shortcut key prompt, Look At wording in the Assistant follow help

* Arabic: complete the Several choice for bodies folders and use a verb form for the dialogs Open row

* Arabic: orbit style term, Select Other and center-point arc names in Sketch preferences, place-from-origin wording, tanween in the radial menu help, tap makers in the tap drill notes

* Arabic: graphics device term in the blur notes, boss and turn wording in the fillet corner option
2026-11-04 05:10:29 +00:00
kaltenbrunnermarkusandClaude Opus 5.5 9d7bbb7fa9 Tests: two Properties tests no longer fail at random on a busy machine (#11750)
* claim: Properties: hold the measuring worker in tests until the card is checked

* Properties: hold the measuring worker in tests until the card is checked

open_object starts the measurement on a worker and looks for the reply
once without waiting. Under load the window side could be held up
between the two, the reply was already in, and two Properties tests saw
the card filled where it should still be measuring (about 0.3 % and 4 %
of runs with 18 tests at a time).

measure_beside now runs a hold before sending the reply. In tests it
waits on the app's existing apply_gate, which feature evaluation already
uses for the same purpose; outside tests it does nothing. Both tests set
the gate, check the card is measuring, then let the worker go.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:08:36 +00:00
leValdemar19andClaude Opus 5.5 6312c2760a MCP list_commands: ask for commands by id prefix or panel, ids and labels only, or a page at a time (#11726) (#11788)
* claim: #11726

* MCP list_commands: ask for commands by id prefix or panel, ids and labels only, or a page at a time (#11726)

list_commands answered with the whole catalog, every key of 575 commands
(about 350 KB), and could only narrow it by a substring. It now also takes
`prefix` (start of the id), `panel` (a whole panel or tab name), `compact`
(id and label), `fields` (the keys to keep) and `limit`/`offset` (pages, as
{commands, total, offset, next_offset}). With none of them the reply is the
same as before. A batch step naming list_commands narrows the same way, and
a bad argument is an error that names it.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 05:07:29 +00:00
3a23d21795 Lip: clearer guidance when the face is not a shelled rim (#10437) (#10871)
* claim: #10437

Co-authored-by: Cursor <cursoragent@cursor.com>

* Lip: say Shell first when the face is not a rim (#10437)

A solid box top left OK disabled with a vague shelled-body hint. Name Shell and the rim-around-an-opening requirement in the refusal, the dialog tip, and the command help.

Co-authored-by: Cursor <cursoragent@cursor.com>

* Lip: keep the dialog hint as it was, match the new refusal text in the test

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

---------

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-04 04:42:08 +00:00
e199f9aeff Agents and scripts: Match Surface describes its exact parameters and reply (#10981) (#10982)
* claim: #10981

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

* Match Surface: input and output schemas (#10981)

surface.match gets a typed contract in schema/surface_match.rs and the
shared FeatureResult as its output, and leaves unconverted.txt. Each key
reads as the handler reads it, with its message; like solid.scale it
declares `name` and `participants` itself, since a match makes no body.

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

* Follow main's kernel merge: world_state and the shared continuity reader

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

* Surface match schema: Vec3 comes from solidcore_record now

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

---------

Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
2026-11-04 04:14:11 +00:00
austin62jakeandClaude Opus 5.5 e06d6e36d8 Agents and scripts: the material card checks describe their exact parameters and reply (#11036) (#11037)
* claim: #11036

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

* Physical material property commands: input and output schemas (#11036)

material.properties.validate and material.properties.directional get
typed contracts and reply types in schema/material_properties.rs and
leave unconverted.txt. The card stays the handler's to check; the
direction has the handler's messages. The replies mirror the document's
Validation and DirectionalModulus, destructured in full.

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

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 04:08:15 +00:00
9d1ad9f0b9 Feature editor: OK accepts the joint placements a roll back re-solves (#11094) (#11095)
* Feature editor: a roll back's re-solved placements are not a changed design

Rolling the marker solves the joints again and stores where that puts each occurrence, so the
editor's pinned original differed from its source there and OK was refused with "the original
editor model does not match its source design". The check now takes the placements as the roll
leaves them.

* Feature editor: also accept the ratio and lift of coupled joints that rolling the marker works out

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>
2026-11-04 03:56:46 +00:00
austin62jakeandClaude Opus 5.5 d39a2d28db Agents and scripts: the clearance check describes its exact parameters and reply (#11051) (#11052)
* claim: #11051

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

* Clearance: input and output schemas (#11051)

inspect.clearance gets a typed contract and reply types in
schema/clearance.rs and leaves unconverted.txt. Each key reads as the
handler reads it, with its message; MAX_CLEARANCE and MAX_NEAREST move
to the schema module so both read one limit. The reply is built from
Clearance and its item types, with the same JSON as before.

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

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-11-04 03:44:59 +00:00