Commit Graph
83 Commits
Author SHA1 Message Date
brett19andClaude Opus 5.5 083b525adf Merge origin/main into smlib
Conflicts kept both sides:
- eval.rs: main's thread-search change (#10227) moved into develop's eval_thread.rs; construction axis/point source bodies with the branch's fallible construction_state; main's identity shortcut in the branch's fallible placed().
- eval_more.rs: main's box helpers and beyond_cylinder_reach kept (main's cut and tap-marking code uses them), body_box meshing through Body::mesh.
- project3d/project.rs: SMLib meeting_edges kept; a face pick now takes its whole surface patch (#9009) through meeting_edges over a set of faces.
- frames.rs: main's split_move (#9660) with the branch's fallible to_local.
- ao.rs: main's draw_opaque with the branch's open-body triangles.
- dialogs.rs, move_preview.rs, viewport.rs: both sides.
- Generated files from main. Tests calling world_state() unwrap its Result.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SdcgEu4Lr3moCCisdZShKz
2026-11-02 01:33:06 -07:00
brett19andClaude Opus 5.5 7f2af7f729 Merge origin/main into smlib
Conflicts, both sides kept:
- Cargo.lock, crates/engine/Cargo.toml: smlib's solidcraft-profile/solidcraft-step beside
  main's solidcraft-testkit and the image comment.
- crates/doc/src/{eval,eval_more,name_lookup,project}.rs: smlib's "a stored name finds its
  face or edge or asks to reselect" logic kept; main's long-name quoting (#8673) applied to
  its messages. Main's split-name hole re-lookup (#10699) dropped: names here never report
  split.
- crates/doc/src/lib.rs: both exports, with main's revolve_extent_for_handle.
- crates/engine/src/cmd/silhouette.rs: main's outline()/Silhouette split, on the fallible
  world_state().
- crates/engine/src/cmd/sketch_project_tests.rs: both tests.
- crates/ui-egui/src/{canvas,hole_refs}.rs: main's anchor_item/model_text on
  open_meshing::session_world.
- crates/ui-egui/src/preview.rs: main's cut_tint switch around smlib's kept-surface ghost
  filter; both helpers and both tests.

Seam fixes: FeatureKind::Repair in main's kind_key/type_name_to_kind_key (#10748); main's
new world_state() callers in ui-egui use session_world or unwrap in tests; a sketch test's
Spline gets nurbs: None.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SdcgEu4Lr3moCCisdZShKz
2026-11-01 18:44:29 -07:00
CamWightandClaude Opus 5.5 325917f2ae Ambient occlusion: no artifacts on the preview or on hollow bodies (#9327) (#9335)
* Ambient occlusion: the radius follows what the frame draws (#9327)

In an empty design the radius came from the model's box, which is empty, so a
preview got the smallest radius and the depth map's precision showed as specks.
It now comes from the box the depth range uses, and with nothing to draw there
is no occlusion.

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

* Ambient occlusion: the term drawn once and blurred by depth (#9327)

The lit pass worked the term out for every fragment it shaded, under a pattern
that changed from pixel to pixel, which read as grain. The term now has its own
texture, drawn once for each pixel of the prepass, with a fixed 4 x 4 pattern, and
a 4 x 4 blur that keeps to each pixel's plane takes the pattern out without
crossing an edge. The lit pass reads the result. The prepass and the term share
their pipelines and draw walk.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-02 00:12:57 +00:00
semmelrindeandClaude Opus 5.5 862807ffc9 Lost GPU: the per-frame probe meets a D3D12 device reset before the first upload does (#6528) (#7767)
* claim: #6528

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

* Lost GPU: the per-frame probe opens a command list, as the upload would (#6528)

On a reset D3D12 device the empty submission before a frame's uploads
went through (it only signals a fence), and the first queue write then
panicked inside wgpu-core (PendingWrites::activate unwrapping Lost).
The probe now submits a command buffer with a debug marker: finishing it
makes the same begin_encoding call, through a path where wgpu loses the
device and runs its lost callback instead of panicking. The picture
capturer's probe does the same.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-01 18:20:40 +00:00
brett19andClaude Opus 5.5 7d760a923a Merge origin/main into smlib
Main's open flow keeps the session's check mode: a saved model under another mode is evaluated again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MNXLXkZLQ6T7TSi4wswcmk
2026-10-31 20:05:06 -07:00
Nathan Dane 7366fe0ecc Blur behind panels: render path, zero cost when off (#8876) (#9809)
* claim: #8876

* Add optional viewport-only Kawase blur and native UI composition

* Fix vendored panel hook imports and deprecated float constant import

* Drain texture updates in the full GPU blur fixture

* Hold the blur evidence fixture camera fixed across radii

* Address blur compositor lint checks

* Share animated Render preview background with blur source

* Use resolved Design background for blur source

* Exercise blur rendering independently of software adapter policy in CI

* Use stock egui hosts instead of vendoring the full crate

* Preserve feature-window open handling with stock egui hosts

* Target the native Window in close-handling regression fixture
2026-11-01 01:37:44 +00:00
brett19andClaude Opus 5.5 1b17dd7f54 Merge fixes: every crate and test target compiles on the merged tree, but for crates/geom's hlr
world_state, view::bounds, home_camera and scene return Result on develop; new
main code takes that with ? or session_world. New fixtures carry the spline's
nurbs. ForeignCopy::of returns Result. occurrence_placement_with walks the path
with its moves, since the placement index holds pending_moves only.

The hlr port is separate and still to come: this commit does not compile as it
stands in crates/geom/src/hlr.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MNXLXkZLQ6T7TSi4wswcmk
2026-10-31 17:01:30 -07:00
brett19andClaude Opus 5.5 5a161e542b Merge main into smlib (develop): 76 conflicting files resolved
main beb35029a (Release 0.29.0), develop 05f5e8601. Committed as resolved and
not yet compiling; fixes follow by layer.

Where the two sides could not both stay as written:
- Section caps: develop's stencil crossing count (with hatch by part) kept;
  main's WINDING/NOT_CAP marking and its fs_solid_touch removed. main's #9407
  test is kept and has to hold on the count.
- crates/kernel/src/prism.rs: develop's, which has no 3D prism boolean; main's
  #6904 route has nothing to change there. Its test still has to be ported.
- Continuity legend: main's details card, with develop's G3/G4 classes.
- inspect.analysis one at a time: while main's "several analyses" preference is
  set, turning one on leaves the others on.
- Model cache: FORMAT 8, since neither side's reader reads the other's records.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MNXLXkZLQ6T7TSi4wswcmk
2026-10-31 16:40:37 -07:00
leValdemar19andClaude Opus 5.5 3385d1c570 Viewport: dark parts keep their shape: linear-light shading and a sky fill (#7974) (#9028)
* claim: #7974

* Viewport: dark parts keep their shape: linear-light shading and a sky fill (#7974)

Preferences > Graphics "Keep the shape of dark parts" (off by default)
shades the viewport in linear light: the key light and eye fill over a
sky fill that grows as a face turns up, plus the 4% of that sky every
dielectric reflects (a little more toward grazing), encoded to sRGB
once. Near-black faces keep a gradient; light and mid-grey faces stay
within a few levels of the sRGB shading. The world's up direction and
the switch ride in a new `sky` vec4 at the end of the viewport uniform,
and the frame key, so turning it on redraws. Off, the shader takes the
old path unchanged.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-31 13:30:07 +00:00
PaddleStrokeandClaude Opus 5.5 526cdfd247 Zebra: flat faces keep one stripe colour while the view turns (#2153) (#9761)
* claim: #2153

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

* Zebra: flat faces keep one stripe colour while the view turns (#2153)

A face whose normal lies in the upright plane through the view (every
horizontal face) reflects straight ahead, where the zebra's sideways
coordinate is zero: exactly on the edge between a black and a white
stripe. Rounding then picked one or the other from frame to frame, so
the face flickered while the camera orbited. Straight ahead is now white.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-31 00:03:59 +00:00
PaddleStrokeandClaude Opus 5.5 2cf8df2ef6 Section caps: a body touching another along a curved face keeps its own cap (#9407) (#9665)
* claim: #9407

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

* Section caps: a body touching another along a curved face keeps its own cap (#9407)

Seen through a cut, a far wall and another body's face lying on it (a shaft
in its bore, a ring in its housing) are the same surface faceted two ways, so
the marking pass let the other body's face win in streaks. The section's
passes now first count the faces behind the cut in the stencil (back +1,
front -1): where the count says the cut runs through a body but a front face
won, the back faces are drawn again pulled toward the eye by a few pixels and
take the cap back.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 23:01:28 +00:00
Mario GeigerandClaude Opus 5.5 332f96c26a Point clouds: colour and filter by any property the file carries (#8816) (#8819)
* Point clouds: colour and filter by any property of a PLY file, on range tracks (#5151)

A PLY file's other vertex properties are kept as named numbers on the
points (whole-numbered for integer types; a comment `property: 0 name,
1 name` names a label's values). A cloud is coloured by any of them, a
ramp or a palette for whole numbers, and filtered by up to eight, each a
range or a set of named values. The GPU colours and filters the points
itself, so editing a range uploads nothing.

In the Edit Point Cloud card each range is a track, as in an audio
editor: a histogram of the property over what the other filters keep,
stacked by colour, under an overview strip. Ranges are drawn, moved,
snapped to colour changes, typed (Start, End, Length, relative values)
and played (Space or Play). The automatic colour range is exactly the
values the property takes over what the filters keep.

pointcloud.insert and pointcloud.edit take color_by: <property>,
color_range and filters; pointcloud.list reports them with the
properties. docs/point-clouds.md describes it all.

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

* Point clouds: note the saved properties, colouring and filters as a pending format change (#8816)

A design now saves a cloud's per-point columns and their value names, colouring by a property and
filters; an older release must refuse it instead of dropping them.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 14:02:29 +00:00
FrankandClaude Opus 5.5 44a9f9ee42 GPU path tracer: the path shader compiles on NVIDIA's Vulkan driver (#6705) (#6850)
Native compilers inline every call, and `main` inlined 33 BVH walks: three
sample_path calls, and a shadow walk per light source family. NVIDIA's
compiler (driver 610.88) crashed in nvgpucomp64.dll or ran out of memory.

- main settles the adaptive budget first, then samples at one call site.
- direct_light takes over direct_studio, visible and the fixed-light loop:
  one loop over the sources with one shadow-ray call site, drawing random
  numbers in the same order.
- next_vertex (a select) replaces (a+1u)%3u: after a BVH walk that index
  crashed NVIDIA's compiler within seconds.

main now inlines 5 BVH walks; main_inlines_few_bvh_walks counts them with
naga, so the limit holds without an NVIDIA GPU.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 13:28:57 +00:00
kaltenbrunnermarkusandClaude Opus 5.5 46a2d6b53c ui-egui offscreen tests pick their GPU by SOLIDCRAFT_GPU (#9052)
offscreen_device() asked wgpu for its default adapter and ignored the
SOLIDCRAFT_GPU policy that headless pictures follow, so with
SOLIDCRAFT_GPU=software the ui-egui offscreen tests still opened the
hardware GPU. Under hardware or software the adapter now comes from
solidcraft_gpu::select (new pick_among, which pick uses too); under cpu
there is none; unset (auto) keeps wgpu's default adapter as before.

Part of #9040.


Claude-Session: https://claude.ai/code/session_013WUFq6QFBqmZjxwaKy2J1M

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-30 03:55:09 +00:00
JK3D 8a6a064d37 Headless pictures: draw counts come back with the picture they count (#7521) (#7522)
* claim: #7521

* fix: headless pictures return their draw counts with the picture

`the_renderer_counts_what_it_draws` drew with `Capturer::rgba` and then
read `Capturer::last_stats()`. The capturer is shared with the binary's
other tests, which draw on it from parallel threads, so another picture
could land between the two calls and the test read its counts (e.g. 10224
triangles for the empty design).

`Capturer::rgba_with_stats` returns the picture and its counts, read while
the call still holds the device. `last_stats()` is removed; this test was
its only caller (the performance overlay reads the viewport renderer's own
stats handle).

New test `pictures_drawn_at_once_each_count_their_own` draws empty and box
designs from eight threads on the shared capturer. Against the old pair it
failed 25 of 40 full runs of the `capture` binary; it passes with the new
call.

* fix: a headless picture waits for its read-back's answer

A picture maps its read-back buffer, waits for its submission with
`device.poll(PollType::Wait)`, and then checked once with `try_recv`
whether the map callback had run. wgpu calls a finished mapping's
callback from whichever thread next maintains the device, in a poll or
in any `queue.submit`, after it lets go of the device. When another
thread got there first, the wait returned with the mapping done but its
callback not yet called, and the picture failed with "not supported yet:
reading a picture back without waiting for the browser", the error meant
for a browser.

On the shared capturer, the other thread is another picture's device
check (`probe_lost` submits nothing before it takes the capturer's lock),
so the `capture` tests failed now and then: `bad_picture_sizes_are_errors`
(its last 8 x 8 picture), `pictures_show_a_plain_cut_without_the_hatch`,
`pictures_blend_a_line_that_is_not_opaque`. In the app, the window's
frames are submitted on the device its thumbnails and `ui.render`
pictures are drawn on.

A native picture now waits for the answer, up to the picture timeout; a
browser still answers later, on its own.

New test `pictures_read_back_while_another_thread_uses_the_device`: 2,000
small pictures while another thread polls their device. Without the fix
it failed 19 of 20 runs (1 to 7 pictures each); with it, 150 of 150 runs
pass. `bad_picture_sizes_are_errors` now prints the error it gets.
2026-10-30 03:19:33 +00:00
Colin Faulkingham d11d268597 Render: reject excessive pixel work during preflight (#5088) (#5728)
* claim: #5088

* fix: reject excessive render pixel work during preflight (#5088)
2026-10-30 01:35:12 +00:00
jyvikoandClaude Opus 5.5 b71e5ae6f4 Tests: edges are read on their NURBS forms, a chamfer's chain is its caller's, the cap test's hatch
- An SMLib circle edge's `interval` is on its NURBS form (`Edge::parameters`), and an analytic
  face's `domain` on the form's (u, v). edges_are_on_smlibs_curves evaluated `curve` at the
  form's interval and took the domain for angles; it checks the side's whole turn and height
  on the form and a rim's turn by `curve_interval`. The tangent hole test's edge middles were
  read the same way (the copy seemed to join faces on opposite sides of the line).
- kernel::chamfer blends the edges it is given; a feature's pick is extended by
  kernel::tangent_chain first (doc eval). The chamfer tests do the same:
  one_chamfer_pick_follows_the_tangent_chain_at_any_part_size removes its 54.66 mm³ again.
- Since the section hatch per part (#8529) a cut is tinted with the hatch colour its body's
  name picks (Body1: steel); the GPU cap test turns the part hatch off, as the CPU one does
  (2885f101b on devgreen-engine-builds).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 12:20:38 -04:00
Andrzej Błasik cb05e1e0f7 GPU path tracer: run on Metal on macOS (#7183) (#7185)
* claim: #7183

* GPU path tracer runs on Metal on macOS, not only Vulkan (#3217)
2026-10-29 13:58:49 +00:00
raykholo 884c47496b Render preview: geometry budget from the device, readable refusals (#4127) (#4154)
* claim: #4127

* Render preview: geometry budget from the device, readable refusals (#4127)

The Render preview refused every scene over a fixed 128 MiB of geometry. On
desktop the budget now follows the device's buffer limit (eight of its
largest buffers, at most 2 GiB; 2 GiB with the usual 256 MiB buffers). The
web keeps 128 MiB.

The product upload runs under its own GPU error scopes: an allocation the
device refuses is reported in the preview instead of reaching the device's
uncaptured-error handler, and the half-uploaded batches are dropped. Buffer
uploads map the buffer themselves so a refused buffer fails the upload
rather than panicking in create_buffer_init.

Refusals name the geometry size and the limit and say how to get under it,
on a theme panel card that stays readable on any preview background.

* Render preview: keep the previous geometry budget behind a preference (#4127)

* Render preview: adapt the device budget to main's GPU upload errors (#4127)
2026-10-29 13:33:39 +00:00
Nathan Dane 3cc10acafa Render lights: uniform light list, solo and mute, and the Light editor (#6281) (#7265)
* Render lights: uniform light list, solo and mute, and the Light editor

* Render lights: Lighting and Light editors to the approved design

* Render lights: the scale-test frame carries the product light list

* Restore the stable Render light mute target on the shared eye control

* fix: validate the viewport BRDF probe with current shader dependencies

* fix: compile legacy lighting controls only for tests

* fix: scope the legacy lighting preset import to tests

* fix: reuse the merged Add Light label and translations

* test: enable the implemented Add Light radial action

* fix: restrict the legacy CPU vector editor to tests

* test: reveal Output before checking Lights browser focus

* test: use the published Browser card rectangle for Output scrolling

* test: reveal Output children before exercising editor navigation
2026-10-29 08:40:51 +00:00
MarkusReandClaude Opus 5.5 0baf49b25f Windows: never load a dxcompiler.dll found on PATH; FXC unless DXC ships beside the app (#6756) (#7923)
* claim: #6756

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

* Windows: never load a dxcompiler.dll found on PATH; FXC unless DXC ships beside the app (#6756)

wgpu's Dx12Compiler::Auto loads dxcompiler.dll by name, so Windows takes the
first one on PATH. SolidCraft ships none, and DXC 1.8.2505 from other programs'
folders (Obsidian, NVIDIA SDK Manager) crashed every start with
EXCEPTION_ILLEGAL_INSTRUCTION during the viewport shader compile.

solidcraft_gpu::dx12_compiler() now chooses: WGPU_DX12_COMPILER as wgpu reads
it when it names a compiler; otherwise DXC from the dxcompiler.dll beside the
program, by its full path, else FXC. The window, the headless/CLI instance and
the test helpers start with it, and dx12_compiles_with_fxc follows it.
WGPU_DX12_COMPILER=dxc keeps the old lookup.

Fixes #6756

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 04:14:57 +00:00
jyvikoandClaude Opus 5.5 bc6d31e603 GPU tests build their designs with SMLib, and the GPU jobs fetch it
The GPU path tests (and the Windows FXC job, cargo test -p solidcraft-gpu) build real
designs through solidcraft-engine, a dev-dependency without its smlib feature. Run on
their own (-p solidcraft-gpu), nothing in the build turns SMLib on, so every geometric
step answered "this build has no SMLib" (10 path::* tests on develop). The
dev-dependency now asks for engine/smlib, as the apps do, and both jobs fetch SMLib
before building.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 22:12:31 -04:00
jyvikoandClaude Opus 5.5 175b4f2acd GPU tessellation: the shader's basis slots follow the CPU evaluator's highest degree
tess.wgsl sized its basis arrays for degree 7 by hand. It now holds 17 slots, degree 16 as the
CPU evaluator, and a test fails when the two drift apart again.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 22:11:39 -04:00
3b983dc550 Viewport: vertex markers where edges are drawn (#7834) (#8119)
* claim: #7834

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

* Viewport: vertex markers where edges are drawn (#7834)

Each body's vertices (the ends of its pickable edges, once each) show as
small screen-space squares in the edge colour in the styles that draw
edges: shaded with edges, wireframe, and shaded with hidden edges, where
hidden vertices show faintly like hidden edges. Plain shaded draws none.
A marker is a lifted line from the vertex to itself in the body's block,
so it is instanced with the edges, cached with them, and depth-tested
like a highlight band. The existing vertex hover ring stays on top.

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

* Viewport tests: count edge segments apart from vertex markers (#7834)

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

* Vertex markers: an option, off by default (#7834)

Adds the Vertex markers preference (Preferences > Graphics, off by default) and draws the dots only when it is on.

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>
2026-10-29 01:53:01 +00:00
jyvikoandClaude Opus 5.5 c7373279d8 Viewport: hidden edges are drawn before the visible ones, so a visible edge is never covered by a light hidden copy
An edge along a face seen edge on lies within a pixel of the nearer face beside it, which hides
part of its width; its hidden copy, drawn after the visible edges, laid a light line over the
visible edge of that nearer face (some 100 px on a box at yaw 1.565). Drawn first, it is covered.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 18:42:39 -04:00
jyvikoandClaude Opus 5.5 30dc4c8601 Section hatch by part: touching cut parts never look alike
Each body a section cuts is hatched at 45 or 135 degrees in one of three
colours of one family (the cap's clay, a steel and a sage, theme tokens
`section_cap` and `section_hatch`). Bodies whose cut outlines touch or
overlap get different hatches, the angle changed first, then the colour
(solidcraft_render::hatch, a greedy map colouring of the cut); otherwise a
body keeps the hatch its name seeds, so it stays the same from frame to
frame and as the plane moves. Spacing and line width are unchanged.

The hatch rides on the palette (an Rg32Uint texel: colour, 1 + hatch), so
the viewport, headless GPU pictures and the CPU raster draw the same.

Preference Graphics > Section hatch: By part (default) or One style, the
look before this change.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5Lzv2n8fyKqD61NCHSF
2026-10-28 17:42:37 -04:00
e6dc105134 Point clouds: insert PLY, XYZ and PTS scans as a modelling reference (#5151) (#5381)
* Point clouds: insert PLY, XYZ and PTS scans as a modelling reference (#5151)

A point cloud is kept in the design as a reference, drawn as depth-tested
dots and never a body.

- io: a PLY reader (ASCII, binary little/big endian; x, y, z of any numeric
  type, optional colours, other elements skipped; counts checked against the
  file) and XYZ/PTS text, thinned evenly past max_points. PLY meshes with
  faces can also come in through file.insert_mesh.
- doc: Document.point_clouds (shared f32 points relative to an f64 origin,
  colours, units, rigid placement, size and colour), stored as a binary
  container part; format change in pending/point-clouds.rs.
- engine: pointcloud.insert / edit / delete / list; doc.open opens a cloud as
  a new design; selection.delete takes point_clouds; view bounds include them.
- gpu: an instanced point-sprite pipeline (16 bytes a point).
- ui: Browser Point Clouds folder with eyes, menu and an Edit Point Cloud
  card; drop, Open, Insert and file dialogs take .ply/.xyz/.pts; a CPU
  viewport fallback; messages in en, en-GB, ja, zh, de and ko.

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

* Point clouds: colour by the file, one colour or height, from the UI (#5151)

The Edit Point Cloud card gets Color By (File Colors, Single Color,
Height). Height is a blue-to-red ramp over design Z, between the cloud's
own two heights or, by default, a range worked out from the points (all
but the lowest and highest 2 %, from an even sample). pointcloud.insert
and pointcloud.edit take color_by and height_range; pointcloud.list
reports them. The colour of each point is worked out in one place
(PointCloud::shade) for the GPU and CPU viewports.

Also: the card's dropdowns compare the picked index (a dropdown's
response doesn't report the pick, so File Units did not change either).

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

* Point clouds: integrate format registration and viewport caches (#5151)

* Point clouds: satisfy current UI audits and document comparisons (#5151)

Route model RGB data through the theme conversion boundary, provide a compact ribbon label, and include point clouds in the new document comparison. Gate the Unix-only CLI-link helper and test so Windows all-target checks remain buildable after the rebase.

* Point clouds: isolate translations for current main (#5151)

Move the existing 74 messages unchanged into one pending locale file and attribute it through a pending row, following the current repository workflow. Repository checks, 50 tooling tests, final workspace Clippy, wasm and generated-catalog equality checks passed after the rebase.

* Point clouds: fit the rebased code to main's GpuTarget and Frame

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

* Compare point clouds when reusing placement-independent results

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

* Point clouds: complete command help and required translations (#5151)

* Point clouds: size the save cap from the raised limit after the rebase

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>
2026-10-28 21:17:02 +00:00
jyvikoandClaude Opus 5.5 c33bae8701 Section cap: only closed bodies take part in the crossing count
An open body (a surface body, a mesh that does not close) has no inside: its faces counted
with the rest took the cap from the closed body it lay in or behind. The body's closedness,
which the kernel knows (`ModelBody::is_closed`: no shell left open, a mesh body every side of
which two triangles share), goes with it to the renderer (`SceneMesh`, `PlannedBody`,
`InstancedMesh`, a block's closed runs, the loose open triangles), and the count draws closed
bodies only.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5Lzv2n8fyKqD61NCHSF
2026-10-28 15:43:35 -04:00
jyvikoandClaude Opus 5.5 33acb6dc39 Section cap by crossing count: a body touching the cut body no longer shows through
The cap went to the nearer surface by depth: a back face (the cut body's far wall) or a front
face. Where bodies touch, the cut body's far wall and the other body's face lie at one depth,
and rounding settled each pixel either way: the other body showed through the cap in specks,
stripes and triangles (an engine's sleeves in their bores, pins in their bosses). The stencil
now counts the faces behind the plane instead, from a base of 128: a back face adds one, a
front face takes one away; above the base, the plane is inside a body and the cap is drawn in
the nearest back face's cut colour, at the plane's depth. Elsewhere the faces draw as without
a section, an open body's inside lit rather than hatched. All four draw paths (built, planned,
instanced, instanced flat).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DtL5Lzv2n8fyKqD61NCHSF
2026-10-28 15:43:35 -04:00
Christian HammacherandClaude Opus 5.5 aae2c5e417 Assemblies: edit a linked component in place, with the assembly around it (#7672) (#7819)
* Links: update linked designs that changed while the assembly's tab was away

A design coming back to its tab compares the modification time and size of its linked
design files with what they were when it went to the background, and runs Update Links
on the worker thread for those that changed: saved in another tab, another window or by
a checkout. Preference update_changed_links (Design, Linked designs) turns it off.

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

* Docs: the assembly guide says linked designs update when their tab returns

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

* Links: Vietnamese strings, now that vi is a complete catalog

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

* Links: remember what each design last read, so an update that fails is tried again

A design's linked files are now recorded as the design last read them: when Update Links reads a
link without an error, or when the link is first seen. Coming back to the tab reads again what
differs, one link per run, so a file that cannot be read holds up no other and is tried again
the next time, and a part changed while the assembly was shown is taken up on its next return.
Closing several tabs together reads nothing on the way, and the web build, which keeps no file
times, looks at nothing.

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

* Links: no read during a save-changes prompt, records for adopted designs and missing files

A tab shown to answer "save changes?" reads no link before its save, and Cancel there or on a
batch lets the tab left in front read what it skipped. A design opened in place of the blank one
records its links at once, a link whose file is missing is recorded as missing so it is read when
the file comes back, and a read queued for a design that is no longer shown by the time it starts
is dropped. More changed links than can wait at once are read on the next return.

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

* Assemblies: edit a linked component in place, with the assembly around it

Activating a linked component opens the design it comes from in a tab of its own, with the
rest of the assembly drawn faded around it where the component sits, so the part's own history,
browser and commands are at hand. Finish Editing in the browser saves the part, goes back to the
assembly and closes the part's tab when the edit opened it; coming back to the assembly updates
the changed link (or Finish Editing runs Update Links when that is turned off). Update Links in
the part also reads the assembly's links again, off the window thread, and redraws the assembly
around the part; a result for an assembly changed meanwhile is dropped with a message.

The part is an ordinary tab, so saving, autosave, recovery, closing and quitting treat it as any
other design. The model scene's ghost triangles are now drawn (only the preview's were).
Preference edit_linked_in_place turns it off. Builds on the update of changed links when an
assembly's tab returns.

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

* Edit in place: a double-click opens it, Activate only activates

A double-click on a linked component edits it in place, in the browser as on a face in the
view (around the placement clicked), and the component's menu gets Edit in Place. Activate and
the browser's radio button make it the active component as they always did, so a linked
component can still be activated without opening its design.

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

* Double-click on a component's face: edit a linked one in place, activate any other

A double-click on a face in the view activates the component the body belongs to, at that
placement, as a double-click on its row does; a linked one is edited in place. Bodies of the
design itself or of the active component are left alone, and so is a double-click with a
modifier. One preference, double_click_components, covers both gestures.

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

* Edit in place: fixes from review

An edit whose part never opens is forgotten instead of starting on a later open of the same
file, and the web build starts the edit after its synchronous open. Finish Editing goes on only
for a save into the linked file and in the part's own tab, waits for a "changed on disk"
answer, and says so when the part was saved elsewhere. Showing the assembly no longer drops a
refresh, which now records the links it read so coming back reads nothing twice. The faded
assembly is worked out once per picture, a bad mesh index skips its whole triangle, and the
last two status texts come from the catalogs.

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

* Edit in place: Finish Editing on the MODEL tab and the mouse's back button

Finish Editing is a command now, component.finish_edit, with its schemas: the app window
answers it, anywhere else nothing changes, as with view.fit. While a linked design is edited in
place the MODEL tab ends with a FINISH EDITING panel, pinned at the right as FINISH SKETCH is,
with a new left-arrow icon; the browser keeps only the line naming the assembly. The mouse's
back button, or a keyboard's Back key, finishes too unless a dialog, tool or sketch is open.
The assembly guide describes editing in place.

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

* Edit in place: fixes from the second review

Saving the assembly waits while it reads its links for the part
Finish Editing waits for dialogs, tools, sketches and prompts open in the part, also when run as a command
The assembly reads the edited part explicitly when it does not check its links on its own
The FINISH EDITING panel cannot be customized
A tab opened by an edit stays that edit's to close when the part is entered again
The mouse's back button finishes editing; the keyboard's only while no text field has focus
Refreshing the assembly keeps its selection and active component

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

* Edit in place: fixes from the third review

component.finish_edit run as a command opens nothing and answers with an error when it cannot finish; engine.batch and engine.script refuse it
A double-click or Edit in Place while a dialog, tool or sketch is open in the assembly keeps it open
What changes in the assembly while it reads its links without changing its document stays: picks, activation, components dragged but not captured
Showing the assembly while it reads its links starts no second read
A new or opened design in the part's tab ends the edit
Finish Editing keeps the part's tab open while its agent chat runs

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

* Edit in place: Update Links run as a command in the part refreshes the assembly too

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

* Edit in place: the ribbon button reads Finish Edit, as Finish Sketch does

A bare "Finish" took the catalogs' surface-finish translation (German "Oberfläche").

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

* Edit in place: fixes from the fourth review

Finish Edit waits while the part's agent is in the middle of its turn, instead of keeping the tab of any chat open
A part saved while its assembly refreshes is read again when the edit finishes: the refresh records the files as it found them
A double-click on a linked component while something is open in the assembly says so instead of activating it
Commands are recognised in any spelling the engine takes, so engine.batch and engine.script cannot slip Finish Edit through
Update Links run as a command refreshes the assembly from the control channel only, not for the app's own reads
Moves of components the refresh removed are not carried over, and a refusal is one message
The command description and the guide say that the command does not save the part

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

* Edit in place: the command is Finish Edit, as Finish Sketch is

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

* Edit in place: a command to start it, and fixes from the full review

Added component.edit_in_place {component, path?}: the double-click and Edit in Place in the menu run it, and as a command it asks nothing, so a network location not allowed yet or a design open elsewhere makes it fail instead
Finish Edit asked by the part's own agent finishes once that agent's turn ends; the button still waits for the turn
The parameters and design properties panels, with text typed and not applied yet, hold up entering and finishing an edit
A refresh is not taken up under an undo, a queued command or a feature still being computed in the assembly
Refusals say what holds them up (a refresh, a file being read or written) instead of nothing
The browser strip names the assembly that is refreshing, not any assembly
The schema description says the command does not save the part
Korean strings for every new message

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

* Edit in place: fixes from the second full review

The MCP batch tool refuses component.edit_in_place and component.finish_edit as steps instead of sending them one by one
Edit in Place asked while the assembly's agent is in its turn opens the part once the turn ends; a double-click then is refused rather than stopping the agent
Both commands answer queued: true when they wait for the agent's turn, and stopping the agent forgets what it queued
An undo, a queued command or a delete check running in the design holds up entering and finishing an edit
A refresh redraws every edit of the assembly, the parts in other tabs too
In an assembly never saved, a double-click activates the linked component as before
The control and MCP docs list the new batch and script refusals, and the commands list every way they fail

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

* Edit in place: fixes from the third full review

The MCP batch tool refuses the two commands however a step names them, an execute step too
A queued Edit in Place is checked fully when asked for, a second one is refused, and both queued requests wait for what runs in the design instead of being dropped
Disconnecting the agent forgets what it queued, as stopping it does

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

* Edit in place: help text for the Finish Edit button

main now requires help for every toolbar command (#7134), and Finish Edit
is one. Written in every language the branch's catalog carries.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-28 15:57:44 +00:00
P'tite TêteandP'tite Tête a4d1b5b6a9 Viewport: fewer ambient occlusion samples while the view moves (#7165) (#7214)
* feat: Fewer ambient occlusion samples while the view moves (#7165)

- Draw AO with 8 samples while the view moves, 32 once it has held still for 0.25 s
- Share the settle tracking with adaptive supersampling and repaint when it settles
- ui.screenshot waits for the full-quality frame (at most half a second)
- Test the sample choice, the uniform slot, and ghosting at both counts

* fix: Set the AO sample count in the scale test frame

- scale_tests.rs, new on main, builds a gpu::Frame without ao_samples
- Use AO_SAMPLES, as the other still-frame builders do

---------

Co-authored-by: P'tite Tête <asso@ptitetete.com>
2026-10-28 15:05:34 +00:00
Nathan Dane 110efa3598 Render: allow picking in the product view with a Body, Face or Component filter and outline the selection in the product pass only (#6426) 2026-10-28 03:20:43 +00:00
semmelrindeandClaude Opus 5.5 e0e2c566e7 Viewport: clear new resolve targets so AMD Direct3D 12 shows the scene (#6828) (#7399)
* claim: #6828

* Viewport: clear new resolve targets so AMD Direct3D 12 shows the scene (#6828)

With anti-aliasing above 1x the scene is drawn multisampled and resolved
into textures that nothing but the resolve writes. wgpu counts the resolve
as their first write and never clears them. Direct3D 12 requires a render
target placed in reused memory to be cleared, discarded or copied over
first, and some AMD Radeon drivers then show the uncleared texture as
transparent, a flat colour or speckles: no grid, origin planes or bodies
(#6828, #7157, #6613, #7052). Each new resolve texture (the frame's colour
and the see-through sums) is now cleared once when it is made.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 20:01:40 +00:00
semmelrindeandClaude Opus 5.5 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
yhamil04andClaude Opus 5.5 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
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
Jackson OurslandandClaude Opus 5.5 76879f8b39 Viewport: write hover edges as one record per line, colours interned once (#7416)
Hovering many bodies re-emits every edge in the highlight's colours. Each edge
interned its three colours into the palette again, each segment was written in
eight small appends, and segments without their faces' normals encoded the
same zero normals again. Lines are now one record, the edge colours are
interned once per highlight, the buffers are reserved per edge, and the shared
width word is worked out once.

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-10-27 07:46:19 +00:00
Jackson OurslandandClaude Opus 5.5 514b0c84f9 Viewport: hold the blocks the sorted glass stream was taken from (#6970)
The no-OIT glass cache compared the blocks' vertex and run buffers by address
without holding them, so a freed buffer's address could come back for other
contents. It now holds them and compares with Arc::ptr_eq.

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-10-27 06:35:15 +00:00
Jackson OurslandandClaude Opus 5.5 ddca1054ab Viewport: leave bodies out of the view out of their blocks' draws (#7436)
Instanced placements were culled per view, but a block drew every body in it
whatever the camera saw. Blocks now keep each body's vertex span and box; when
the view changes, each class's runs are drawn without the bodies whose boxes
are out of the view. A block whose bodies are all in view draws its runs as
they are.

Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-10-27 06:35:11 +00:00
jyvikoandClaude Opus 5.5 5fc160732e Main's new callers of the placing calls that now return their error
Code main added since the last rebase reads the placed model, a camera or a scene
as values; on this branch they return Result. Picks and the edit dialogs read the
session's placed model through open_meshing::session_world, as the others do; the
revolve and circular-pattern axes, the model axis of a command and the warming of a
computed edit pass the error on; tests unwrap. A test meshing bodies uses
Body::mesh, the one mesher.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HoZgkCdAVrscuw2kjQTWDs
2026-10-27 02:28:53 -04:00
jyvikoandClaude Opus 5.5 cfc393d68f Main's new callers of the placing calls that now return their error
The branch's placement change (placing the model returns its error) made
world_state, home_camera, bounds, open_sketch_plane, the scene builders,
to_local and auto_operation return a Result. Main added callers of them
since the branch left it; they now take the error as the branch's own
callers do:

- engine: commands pass it on (`?`); helpers answering yes/no or "found"
  (snap points, a face's normal, the DXF export check) answer no for a
  model that cannot be placed. Frames keep the session for main's lazy body
  frames (#6356) beside the placed model the picks are on; a library apply
  maps its targets on that placed model. Before an extract or a link is
  replaced, the placed model is taken or the command fails.
- ui: panels, picks, dialogs and previews read the placed model through
  open_meshing::session_world (none when it cannot be placed: the viewport
  says why); Home and Fit frame nothing then.
- cli: the usage table is main's, with kernel-cases in it (a resolution of
  the rig's commit had kept the branch's whole file).
- tests read the Results with unwrap, and main's tests of Body::tessellate
  mesh through Body::mesh. Main's test of saved boolean retries goes (the
  branch saves none), as does the thread half of its dialog rendering test
  (the branch's thread dialog has its own); the tapped hole half stays.
- What no caller uses after the merge goes: a mesh body's span, the metric
  thread refusal text, and the revolve's targets-as-shown model (main used
  it for the coplanar run-on the branch removed). Main's SolidWorks Y-up
  test reads the exporting system's name (the branch's STEP header has no
  preprocessor or description fields).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 02:28:53 -04:00
jyvikoandClaude Opus 5.5 9a79588575 doc, engine, ui: placing the model returns its error; placements are checked where they are set
doc: world_state, world_state_reusing, state_in_frame_at, state_in_frame, face_point_state,
face_point_names and construction_state return Result. A body that could not be placed, a
frame that could not be inverted, a sketch plane or thread axis that did not survive the
placement, or a failed placing worker used to leave things where they were without a word;
now the call fails with what could not be placed. is_placement is the one test of an
occurrence's placement (finite, a rotation that does not mirror, a translation within
1e9 mm). Opening a design names every occurrence that fails it; add_occurrence refuses
one; a joint solution that would place an occurrence outside one is not applied and the
joints placing it report it.

engine: Session::world_state returns Result and keeps a failure in its cache as it keeps a
placed model. bounds, home_camera, open_sketch_plane, the scene builders, the world
previews, frames::to_local and auto_operation pass the error on; helpers answering "not
found" take the placed model from their caller. Commands setting a placement check it
(valid_position, exploded view, fastener at a point, STEP insert matrix).

ui: panels, picks and hovers read the session's placed model through
open_meshing::session_world; a model that cannot be placed shows no placed bodies, and the
viewport says why across its bottom. Pictures and the control channel report the error.

cli, mcp: snapshots, renders, benches and the oracle report the error; the kernel-case rig
records it as the case's failure. Tests read the new Results with unwrap.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017itGh5fpMEousnZ91cYcVt
2026-10-27 02:28:51 -04:00
jyvikoandClaude Opus 5.5 42997a62ed WIP kernel: Body::mesh(&MeshTolerance) as the one mesher (does not build yet)
Body::mesh / mesh_plan / mesh_of and MeshTolerance (chord, display, display_of, export_of)
replace tessellate, tessellate_with, display_mesh*, display_plan*, reference_display_mesh_angular;
every caller in kernel, doc, engine, io, ui-egui, gpu and the CLI renamed mechanically; the
truck-mesher fallback, mesh_shells*, refined* removed; Body::new orients from the planner;
truck-meshalgo dropped from Cargo.toml with the tests of its hooks. Still to fix: trait imports
in tess/*, splitface/lumps/preview/step_in/boolean_record meshalgo uses, measure.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017itGh5fpMEousnZ91cYcVt
2026-10-27 02:28:47 -04:00
Jackson OurslandandClaude Opus 5.5 efa2cae0b4 Viewport: draw faded and see-through bodies from their plans with OIT (#7423)
* Viewport: palette-indexed scene and incremental appearance updates

- Move triangle, line and instance colours and opacity into a per-scene
  palette (an R32Uint texture, WebGL2-safe); appearance edits update
  palette entries instead of vertex bytes.
- Build CPU body triangles as one canonical vertex stream per block with
  opaque/glass/faded index lists; an opacity class change rewrites only
  the index list.
- Cache resolved body appearances by a per-body key, so an edit resolves
  only the bodies whose look changed.
- Split the model scene key from the session revision: selection and
  view-only commands no longer rebuild or upload the model scene.
- Replace the O(n^2) section-index lookup and per-placement instance scan
  with hash lookups.

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

* Viewport: skip unchanged palette work when rebuilding

- Reuse each body's appearance key while the document is unchanged.
- Leave a palette slot alone when its value did not change, and allocate
  every body's slots once before the build loop.

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

* Viewport: measure scene rebuild cost by edit kind

An ignored scale test times the scene build after selection, appearance,
theme and one-body geometry edits at 200, 1,000 and 5,000 bodies.

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

* Viewport: read faded grid alphas through the scene palette

The grid-fade tests now resolve line palette indices instead of reading the
line record's old colour byte.

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

* Viewport: measure rebuild cost on few large meshes

Adds the 20/100-torus case where a block rebuild used to re-emit a large
part of the model.

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

* Viewport: measure rebuild cost on an assembly

Unique detailed parts plus one component placed many times, timed for
selection, part and component colour, opacity, theme, hide and move.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* Viewport: draw block classes as vertex runs, not index lists

A body's vertices are three per triangle in mesh order, so each draw class
is a few runs of the block's vertex buffer: all-opaque blocks need no index
buffer, and an opacity class change uploads nothing. A fragmented class
(more than 64 runs) falls back to one index buffer.

Each triangle is written as one record into a block reserved up front,
which takes the cold scene build of a 7.1M-triangle assembly from 97 ms
to 55 ms (main: 83 ms).

Also fix reindexing a block whose bodies are drawn elsewhere: it walked
the mesh's triangles from that body's start, past its own vertices.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* Viewport: draw the highlight's own block, and light faded bodies without copies

The highlight scene keeps its feedback (a hovered face, a body copied on
the CPU) in a block of its own; since blocks became vertex runs, only the
model's blocks were drawn, so that feedback was uploaded and never shown.
Draw the highlight's blocks with its triangles.

Whole faded and see-through bodies (another component active, or an
appearance below full opacity) were lit by copying every triangle on the
CPU at each hover change: 85 ms for a design row of 812 bodies. They are
now drawn again from the model's own block vertices, in the highlight
palette with their slots set to the highlight colour, or as placements of
the model's shared triangles. The picture is the same as the copy's.

Adds a GPU frame benchmark for document-tree hover (ignored test).

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* Viewport: count scene uploads, and keep the sorted glass stream across colour edits

Renderer::uploads reports the bytes the model scene's last change uploaded
(geometry, placements, palette rows, planned colours, re-evaluated vertices).
Without order-independent transparency the global glass stream is kept while
the blocks' see-through triangles are the same, so a colour edit on a glass
body uploads palette rows only. Tests assert both on a GPU.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* Viewport: keep palette slots and looks without rehashing every body

A rebuild found every body's palette slots by name and wrote them back,
allocating and hashing a new key per body, and copied every shown name into a
set to check the arena still belongs to the design. Slots are written back only
when a body's faces grew, and the looks cache is pruned only when it holds more
bodies than are shown. Adds the STEP-file benchmark (SOLIDCRAFT_SCALE_STEP) and
the colour command's own time to the assembly benchmark.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* Viewport: round-pick construction lines through the scene palette

The axis and centre feedback for round picks (#4610) drew with a raw colour;
on the palette scene it is interned once like the other construction geometry.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

* Viewport: draw faded and see-through bodies from their plans with OIT

A body faded outside the active component, or see-through, left the GPU
tessellation path and was built triangle by triangle on the CPU, so working in
a component of a large assembly rebuilt every other body. Where the renderer
blends see-through surfaces order-independently (GpuTarget::oit, set by the
renderer), a body whose faces share one class is drawn from its plan: chunks
are split by class, opaque passes, AO and the section draw only opaque chunks,
and the OIT pass draws the rest. The sorted fallback keeps them on the CPU.
The model and its highlights decide through one rule (model_plan_class).

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

---------

Co-authored-by: Claude Code <noreply@anthropic.com>
2026-10-27 03:27:32 +00:00
Homero ThompsonandClaude Opus 5.5 4b327bab1f Canvas: the value boxes do not cover the command window (#6608) (#6661)
* claim: #6608

* Canvas: keep value boxes clear of the command window (#6608)

* Canvas: describe how value boxes clear the command window (#6608)

* Canvas: keep the value boxes clear of the gizmo handles (#6608)

* ui-egui: format the value box regression test (#6608)

* gpu: the startup MSAA check skips when there is no GPU adapter

The check needs a real adapter to draw the multisampled view and to
resolve it, so on a machine with no software Vulkan adapter it is
skipped, with the reason printed, instead of failing the build.
SOLIDCRAFT_REQUIRE_GPU=1 still makes a missing adapter a failure, as in
the CI job that has a software adapter.

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

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-27 02:29:53 +00:00
semmelrindeandClaude Opus 5.5 e29075376c Curvature Map: a scale that lightens steadily from flat to tight (#975) (#6136)
The blue, green, red scale's lightness rose and then fell (L* 44 up to
72 and back to 50), so a flat face and a tight radius looked alike with
hue lost, and its upper part merged for a red-green loss. The stops are
now dark blue, a middle grey-gold and light yellow in one straight
blend, as measured in the issue: L* rises the whole way and every
quarter step stays clear for normal, protan, deutan and tritan vision.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 12:25:11 +00:00
c6d361347f Section Analysis: the cap is marked by true depth, so far edges no longer take a cap band (#6328) (#6332)
With a section shown, the marking pass decided where the cap shows (the stencil) and wrote the
depth everything after it tests against. #1768 gave it the faces' FACE_BIAS for the second job,
so edges draw over their own faces (#1434). For the first job its slope part is unfair: next to an
edge where a back face seen through the cut meets a front face, the front face won a band about a
pixel wide although the back face is nearer. The cap skipped that band, and the edge line showed
through the cut.

#6363 marked those samples again from the back faces, pulled toward the eye as far as FACE_BIAS
pushes. That closes the gaps beside the edges, but the pull also lets back faces win wherever they
lie within the push of a front face, along every body's far silhouettes, and the cap hatches that
band. On bherbruck's block, the band away from the cut went from 58 px orthographic and 534 px
perspective before #6363 to 618 and 780 with it. In the app, his filleted hub showed it in every
frame of a 36-step orbit in both projections.

This replaces that pass. While a section is shown the opaque model now draws in four steps (`Cut`):
1. mark: the faces by their true depth (no push): colour, the stencil and depth;
2. a triangle over the viewport sets the depth back to the far plane, colour and stencil kept;
3. push: the faces again with FACE_BIAS, depth only, so the depth is what unsectioned frames have
   and edges draw over their faces as before;
4. cap: depth only, as #6728 made it; the mark pass has drawn the cap's colour and hatch, as fs_solid
   draws a back face seen through the cut.
The window pass's depth reset and this one are one helper, `depth_reset_pipeline`, for either
depth direction, and its passes are built by #6728's `solid_writes`. Unsectioned frames are unchanged. The CPU rasterizer already chose the cap by
true depth.

Tests (crates/gpu/tests/depth.rs), with the device's float depth and the 24-bit fallback:
- a_cut_hides_the_edges_behind_it (new): bherbruck's block cut at y = -16, home view, orthographic
  and perspective. No edge shows through the cap, and away from the cut the picture is a box of
  just the kept part's: 0 px orthographic and 1 perspective (main: 618 and 780).
- bherbruck's pocket and hub tests from #6363 are kept and pass (pocket 0 px; hub 1 and 2 px over
  its twelve views, as with #6363's pass).
- an_edge_on_a_face_draws_whole_at_any_angle and
  an_edge_just_behind_a_face_stays_hidden_on_a_large_design also run with a section shown.

In the app (safe mode, WARP), with 625 patterned cylinders cut by a section, the frame time is
that of main (#6363): 33.5 against 34.3 ms at the home view, 34.1 against 35.2 ms with the model
filling the view.

Co-authored-by: bherbruck <brennen.herbruck@gmail.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 11:16:44 +00:00
raykholo 30e9b87b48 Render workspace uses the window's graphics backend on Windows (#6592) (#6629)
* claim: #6592

* Keep Windows Render setup on the viewport backend
2026-10-26 11:10:32 +00:00
Homero Thompson 4ced1d7bcf Viewport: the grid fades with the distance instead of cutting off (#4214) (#4629)
* claim: #4214

* Viewport: the grid fades with the distance instead of cutting off (#4214)

The ground grid stopped at a fixed distance from the view, so its edge showed
as a cut across the picture. It now fades out with the distance from the
middle of the view, and it is gone before either of the two lines that used to
cut it: its own edge, and the picture's far plane.

Each grid line is drawn in pieces, and each piece is as faint as the fade
makes it at the distance where it lies. The 'under' batch is drawn by the
overlay pipeline with alpha blending, so the colour's alpha reaches the
frame.

* The headless grid fades with the distance too (#4214)

The headless path (solidcraft-cli snapshot, the home-page thumbnail, the MCP
screenshot) drew the grid from the engine's palette, where a line had no
alpha: the grid still ended in a cut, and at a low camera angle the cut was
in the picture.

A SceneLine now carries an alpha, the CPU rasterizer blends with it and the
GPU capture passes it on, so the engine's grid can fade. The fade itself
moves to the engine, next to the grid's other geometry, and both paths use
it: the headless grid now reaches as far on the ground as the view is wide
(the ground stretches as it is seen at a slant) and dissolves into the
distance instead of ending in a line.

* Fix the headless/viewport grid fade divergence (#4214)

The engine stretched the grid's size with the slant of the plane it is seen at, the viewport stretched it only on the ground, so on a sketch's plane the picture's grid reached further than the window's (153.42 mm against 120.00 mm). Both now share one grid_ext, which the picture and the window draw the same grid from.

* The grid fades out before the picture's far plane cuts the ground (#4214)

The window's grid faded against a far plane of its own: the scene's radius
clipped by the projection's camera (260 mm away at 400 mm of distance), while
the picture itself cuts the ground at the far plane of its own projection
(223.7 mm at the same distance, because the projection frames the scene's
bounds and not its radius). The grid was still a third of its strength where
the picture cut the ground, so the cut showed as a straight line across the
view.

The fade now ends at the far plane the picture's projection uses, and the
grid is gone before that plane cuts the ground.

The test that watched the fade now measures the ink of each row of the
picture, and checks that the grid is already gone in the ten rows above the
row where the far plane cuts the ground: the point it compared before lay
beyond that plane, where both renders draw nothing at all, so the comparison
was always true.
2026-10-26 07:59:56 +00:00
raykholo 36a60a70d9 Graphics: 1x anti-aliasing where the multisampled view draws nothing (#5819, #5870) (#5832) 2026-10-25 22:10:04 +00:00