Packs the scene's planned bodies into one GPU upload and dispatch, and evaluates intersection edges by seeded Newton refinement from a fitted B-spline through exactly solved leader points, falling back to the previous evaluation near tangency. Intersection nodes carry their faces' parameters. This changes speed, not what the app draws.
Part 4 of 4 of #303.
Co-authored-by: jyviko <2501648+jyviko@users.noreply.github.com>
Thicken now moves faces through the same mover as Offset Surface. Spheres, cones, tori and B-spline faces thicken as well as planes and cylinders, and a thickness that would fold a curved face over is refused. Where thickened faces meet at a sharp edge, a wall closes the gap and corners are closed with flat patches. Thickening into an inside corner is still refused with a clear message.
Closes#765.
Co-authored-by: jyviko <2501648+jyviko@users.noreply.github.com>
Offsets picked faces along their normals into a new surface body: planes and cylinders move exactly, other faces get a fitted B-spline within tolerance, and offsets that would fold over are refused. Adds the SurfaceOffset feature (noted under format 18), the surface.offset engine command and an Offset Surface dialog.
Closes#758.
Co-authored-by: jyviko <2501648+jyviko@users.noreply.github.com>
A section stayed on with nothing in the browser to switch it off; only the
viewport's context menu had Remove Section. While there is one, the browser
now has an Analysis folder listing it. Its eye (and the folder's) hides the
section and shows it again on the same plane; a double-click or Edit Section
Analysis opens the dialog, and Delete removes it. Browser rows with an eye
publish an `eye:<label>` handle for automation.
Fixes#822
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Closing a socket with unread input resets the connection, and the reset could reach the client before the refusal did (requests_run_only_after_the_token failed 17 times in 100). The server now ends its side and drains, briefly and up to 64 KiB, what the client had sent before closing: 0 in 100.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A through-all hole's tool is twice the largest body long, and the boolean's tolerances
followed that length: a 6 mm hole through a 920 mm part was traced at nearly 1 mm, wider than
the hole, and the cut came out empty or failed. When every other strategy has failed, the
boolean is now asked again (unshifted, then once shifted) at tolerances scaled to where the
bodies meet: the overlap of their boxes and the thinner operand's thickness. Its result is
checked as the others are and held to the retry guard. It runs last, so whatever an earlier
strategy answered is kept as it was, and only for single-shell operands (bodies in several
pieces are retried a piece at a time, each with this pass of its own).
The sampled shortcut also gave up on a slender tool lying across its box (a long hole drilled at
an angle fills little of it): it now samples past the first 2000 points until enough land
inside, so a tool that misses a body leaves that body as it is. Bodies that had enough samples
in the first 2000 are sampled exactly as before.
Baseline: the 90 actions that newly pass, rechecked one by one with the final build.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A design with components and joints exports as a URDF robot description (file.export to .urdf, the MCP export tool, solidcraft-cli run --out, File > Export URDF...). It writes one binary STL per component in metres, links with mass, centre of mass and inertia from material density, and maps rigid, revolute, slider, cylindrical, pin-slot, planar and ball joints to URDF joints or short chains. Loops, broken joints and loose parts are reported as warnings. Options: mesh_prefix, effort, velocity. Docs in docs/urdf.md.
Each wall of a surface built from sketch curves (Surface Extrude, Revolve, Sweep and Loft) is now named after the curve it comes from, exactly rather than from geometry, so references to those faces survive edits.
The browser build fetched its Japanese fallback font only for a Japanese UI,
so a Japanese design name, file name or typed text showed as tofu with the UI
in English. After drawing each frame the web shell now looks through the
painted text for kana, CJK ideographs and full-width forms while no CJK font
is loaded, and fetches the same font once on the first hit. A failed fetch
with an English UI leaves the language alone and is not retried in a loop.
Sessions that never show such text download nothing extra.
Fixes#720
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* Toolbar drop-downs show all their rows after a shorter list (#874)
A panel's drop-down kept its size by an id made from the panel name alone,
so PLASTIC's MODIFY reused the one-row size of MESH's MODIFY and showed two
rows with a scroll bar. The id now includes the tab, and the rows' scroll area
may grow past the size the area remembered (up to the room below it), which
also covers the same panel collapsed and expanded.
Fixes#874
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Keep the Sketch Create menu's own sizing
Its rows sit in egui menus (also inside the overflow menu), where a minimum
height moved the rows between frames. Only the panel drop-downs need it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Moving a window between monitors of different scale can briefly ask for a
surface larger than the device's largest texture (11512 x 6808 against 8192 in
the report), and Surface::configure failed validation, which wgpu turns into a
panic. The surface is now clamped to the device limit with its aspect ratio
kept, and egui draws the whole window at the matching smaller scale, so the
frame is only softer until the size settles. The device also asks for the
adapter's own texture limit (16384 on most desktop GPUs) instead of 8192, and
the resize path and an unsupported surface log instead of unwrapping.
Fixes#869
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Surface Loft lofts open or closed curve chains, one section per sketch in pick order, into surface bodies. Adds the SurfaceLoft feature (format 18 note), the surface.loft command, a dialog with Edit Feature support, and kernel and engine tests.
Part of #757.
* Fillet and chamfer: refuse a smooth seam as having no corner to round
Picking the seams where R10 rounds run into the faces beside them, and asking for an R1
fillet there, failed with "not supported yet: blending a smooth edge (... both faces next
to the edge must be planar)". That reads like a missing feature, but the faces meet
tangentially at those edges and there is nothing to round.
An edge whose two faces are tangent along it is now refused as an invalid pick that names
the edge and asks for the sharp edges.
Fixes#896.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* Robustness goldens: seed 72's smooth-seam fillet is now a rejected pick, not unsupported
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
#825 chose the faces to split by boxing each face, dividing its curved edges down to the contact tolerance; on the sensor enclosure's lid that division never finished (press/pull and extrude on 18 faces grew past 24 GB and timed out in the action matrix, on CI as here). The boxes now come from the edges' control points, with no division; a face whose extent can't be found counts as near. A test holds the lid's press/pull to a time budget.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Surface Sweep carries open curves and closed profile boundaries along a sketch path into surface bodies, with smooth walls along curved point paths. Adds the SurfaceSweep feature (format 18 note), the surface.sweep command, a dialog with Edit Feature support, and kernel and engine tests.
Part of #757.
Walls swept along a spline have flat B-spline tops and bottoms, so booleans between them went to the general intersection, which fails on coincident spline faces. Such faces are now put on true planes, and two prisms of the same height and direction are decided in 2D with their spline outlines fitted by arcs within a micron. The result is kept only when its volume is plausible; otherwise the general boolean takes over. The free-form flag is cleared by a drop guard so an unwinding boolean can't leave it set.
Fixes#844.
The exact surface search could return an arbitrary parameter at a collapsed NURBS boundary, which corrupted the trimmed triangulation domain and lost volume on fine meshes of a fillet tapering to zero. Search results in the boundary projection are now validated, preferring a bounded parameter at a pole and snapping to a finite endpoint only when its transverse tangent has collapsed and the residual improves. Adds four regressions for decreasing chord tolerances, pole parameters, unbounded planar support and provisional preview.
Fixes#796.
Body::with_planar_sources compared every planar face of a result with every unsourced planar face of its inputs, so a boolean on an import with a few thousand facets ran out of its comparison budget before it started. The lists are now sorted by normal x and each face only compares with entries inside a narrow normal window found by binary search; the exact same_support test and the work budget are unchanged. Adds an engine test joining a pin onto a 4000-sided imported prism.
Surface Revolve turns open curves and closed profile boundaries about a sketch axis into surface bodies, with cone apexes and sphere poles where a chain ends on the axis, and a closed shell made solid. Adds the SurfaceRevolve feature (format 18 note), the surface.revolve command, a dialog with Edit Feature support and analytic kernel and engine tests.
Part of #757.
A lip picked on a face that isn't a shelled body's rim, such as the Sample Plate's top
face with its two holes and the boss standing on it, ran its band round the largest inner
edge. That was the boss's foot, so the band stood against the boss's wall and the join
failed with a boolean error about curved faces.
The band now goes round the face's opening (an inner edge with no material just over the
face inside it). A rim face that has a smaller hole in it still works. A face whose inner
edges are only where other features stand on it, or whose openings are alike in size (a
plate's holes), is refused with a plain message that names the rim face it needs.
Fixes#892.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Generates N parts of four primitive bodies, each moved into a placed component. The test (12 parts) checks the open, the world placement and the saved round trip; the ignored bench_generated_assembly prints parse, evaluation inside and outside the features, mesh warm-up and Measure.
A boolean that cuts a pocket into the top of a cylinder (Shell does this) splits the
outer wall's straight seam edges at the pocket's floor. The wall's parameter grid has no
division along its rulings, so the triangulation fanned out from the two split vertices
to the far rims and its triangles cut through the part: the shelled Bolted Flange sample
drew and exported its rim cut off (mesh volume 29201 mm3 for a 40796 mm3 body).
Give the grid a line through each vertex that splits a boundary segment running along a
ruling, in a direction that needs no division of its own and within the grid budget.
Fixes#819.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
The app's control protocol gets engine.batch: the steps run in one request (so one frame), each as engine.execute runs it, stopping at the first failure and rolling back to the checkpoint by default. The MCP batch tool sends it to a running app (with the control token), is never resent after a lost reply, and falls back to step-by-step when the app refuses (agent cursor shown, or a step loads a file). A batch may take up to an hour to answer.
find_close_pair's budget now counts the grid's own work too (each cell a triangle is filed under or visits), not only the pair tests, so a pathological pair of meshes can't spend unbounded time building and walking the grid. Adds tests for find_close_pair and the budget.
face_vertices lists each vertex once per face (first-met order) and normal_near measures each distance once; seam flags, merged topology and cone signatures are unchanged because the dropped repeats were identical entries.
Hole centres from a sketch now drill into the material on whichever side of the sketch plane it lies: material the spot starts in wins, then the nearest body ahead, else against the sketch normal as before. A sketch on a part's underside or below it now drills up into the part. Covers counterbores, through-all, edge and corner centres, patterns and reopening.
Delete Face takes heal (default true, as before). With heal: false the picked faces are removed and the gap stays open, leaving the remaining faces as a surface body (kernel remove_faces); deleting every face is refused. Saved only when false, noted under format 18.
A body transfer works out the moved bodies first and puts them in place in the evaluator's own copy of the model, instead of copying every body of the model for each transfer.
Display meshes after opening, the face names read when a feature finds a face by name, and inspect.measure over all bodies are now worked out per body on plain scoped threads (solidcraft_kernel::for_each_on_threads / map_on_threads), in body order and with the same results; the browser build stays serial.
After each feature the bodies before it are indexed by name once (first match first, as the linear searches were), feature components come from a map built once per evaluation, and Body::same compares shared face sources, paint and split lines by pointer first.
Body::retained_bytes counts face and edge uses on the borrowed boundaries (the same numbers, checked by a test), and each model body keeps the footprint measured when a feature made or changed it, so the evaluation cache's budget no longer walks every body's topology after every feature.
NameCell::same now refers to its source body and takes the source's names on first request, instead of naming every face of every moved or placed body up front. Names are unchanged.
Where two boundaries of a projected face family touch at a single vertex, an even number of boundary edges meet there. The walk now stays on the face it came along (or takes the first edge in geometric order), so every loop closes; an odd count is still refused as a real branch. Adds a test with two squares meeting at a corner.
A sketch text gets an optional style: bold and italic faces (a slanted fallback for italic), left/center/right alignment, letter and line spacing, and a path curve of the same sketch for the baseline to follow (reverse takes the other side). Edit Text gets B/I toggles, alignment, spacing fields and a Path chip, all live. The style is saved only when not plain and is noted under format 18.
solidcraft --control PORT writes a random per-launch token to a user-only file (<data>/control/PORT.token; 0700/0600 on Unix, icacls on Windows; SOLIDCRAFT_CONTROL_TOKEN overrides) and refuses any connection whose first request lacks it. mcp --connect, the demo and packaging smoke clients send it. MCP execute/batch/open/save/export/screenshot are annotated destructive and open-world, new_design destructive. Docs: control-protocol.md, mcp.md, AGENTS.md, pending release note in releasing.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Only one run per concurrency group can wait. On a busy main each push replaced the waiting run, so the nightly and Run workflow runs on main were cancelled before they started. The group now includes the event.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Renaming a body used to replace every string in later features equal to its old name, including sketch text, plane and feature names and expressions. Now only values under keys that name bodies (body, bodies, target, targets, tools) at any depth are rewritten, with a test that a sketch text spelling the body's name stays while a move follows the rename.
Corrects the modeling skill: an edge of the face a shell removes is rounded after the shell, a slot along Y is dimensioned horizontally, long shallow arcs measure to about 0.1%, and a boolean retry warning can hide wrong geometry when the volume does not match.
A hole drilled square to a flange's panel, not only the base's, is now a flat cut-out of the sheet: it unfolds with its panel to where the bend allowance puts it, shows in the flat pattern and DXF as a circle, and survives unfold and refold. A pocket tool spans the cut-outs already inside it (a counterbore's hole), so the order they are made in no longer matters. Fixes#560.
Main has failed the full robustness goldens since #838, #839, #841, #855 and #856 landed together. Records the eleven seeds that now build differently (2, 6, 14, 22, 36, 43, 50, 63, 71, 73, 94), with a note. Shell now refusing seed 14, which used to build, is tracked in #897.
When an extrude cuts with profiles whose outlines stay more than 1 µm apart (the letters of a text), their tools are apart too, so they are joined and cut from the body in one boolean instead of one per profile. A ten-letter text cut drops from about 13 s to 1.5 s with the same result. Sheet metal targets and touching profiles keep cutting tool by tool. Tapered extrudes are unaffected.
Agent tool calls (MCP, and bridged control requests carrying allow_paths) reach files only in the open design's folder (strictly inside home or the temporary folder, not in a hidden folder, and never its hidden files), the app's recovery folder, the server's private scratch folder and each --allow-path; --allow-any-path lifts the limit. Paths are resolved first, so .. and symlinks that lead out (including a save through a link, #845) are refused. On Windows, UNC, device, NT, verbatim non-drive, drive-relative, alternate-stream, reserved-device and 8.3 short-name paths are refused before any file system call, and the resolved path must be on a local drive. doc.recovery_discard and doc.recover {discard} delete only real recovery entries. Docs: mcp.md, control-protocol.md, README.
