kernel::sweep swept segment by segment: a line translates the profile along it, an arc turns it. At a sharp corner the next line runs within the plane the profile has reached, so the profile slid sideways in its own plane and the walls folded into an inside-out, zero-volume shell. Paths whose direction changes at a joint (more than 1e-3 rad) are now swept as a polyline by sweep_path, which mitres the corners; arcs become points within 0.01 mm. Smooth paths (lines and tangent arcs) are swept as before. Pipe uses the same path.
Tests: sweep_round_a_sharp_corner_is_mitred sweeps a 2 x 2 square along an L of two 10 mm lines and checks the volume, 80 mm3 (fails on main).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The loft's top lies in the pocket floor. truck's boolean cannot take coincident faces, and the exact polygon fallback (planar_boolean) only takes bodies whose faces are all planar, so the hole's cylinder ruled out the only path that handles this cut. planar_boolean now first tries a sliced form: when an axis-aligned slab holds the tool and none of the body's curved faces, the body is split on the slab's planes (planes clear of the body's own faces), the planar middle piece takes the exact boolean, and the pieces outside are glued back on along the cut faces (the middle piece is healed first so its edges match theirs). Pieces in several parts are refused rather than glued into one shell. split_by_plane's half-space code is now ops::half_space, which this uses.
Tests: loft_cut_beside_a_round_hole replays examples/regressions/loft_cut_beside_round_hole.json and checks the exact volume (slab - pocket - prismatoid - hole; fails on main). With the hole made after the loft, the body is the same (volume, area, 19 faces, 42 edges).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Moving a body transforms its intersection-curve edges, and truck-geometry 0.5.0 scales each curve's tolerance by the operator norm of the matrix. matext4cgmath 0.1.0 computes it from eigenvalues through a quartic whose Newton refinement loops while the residual is above a tolerance and never gives up. For a translation matrix all four eigenvalues are 1, and near that quadruple root the iteration can cycle forever (about one translation in eight of the boolean retries' shifts, by a scan of sizes 0.1-2000 mm). matext4cgmath is vendored with both Newton loops stopped after 100 steps (vendor/README.md, ATTRIBUTION.md).
truck master no longer computes the norm in transform_by, so the vendored copy can go once a truck release after 0.5.0 is adopted.
Tests: moving_a_body_with_an_intersection_curve_edge_ends (kernel) cuts a cylinder with a tilted box and moves the result by one of the shifts that hung (hangs on main). examples/regressions/oblique_cut_of_cylinder_hang.json is the variant of the oblique cut that hung the app; it hangs on main, and since the previous commit it passes without reaching the shifted retries, so it is replayed by the engine test as a scenario only.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The boolean checks its result by meshing both operands and the result, then sampling membership densely where the operands' boxes overlap. The mesh step came from the larger body's size (0.46 mm here), which facets the bar's 3 mm arc so coarsely that two meshes of the same arc disagree over a few percent of the thin overlap box, and a correct cut failed the 0.4 % mismatch limit. The step now also follows the smallest side of the overlap box (2e-3 of it), never finer than 1e-5 of the size, so it only changes where the bodies meet in a region much thinner than they are. truck's first result for the cut is then accepted, and it matches a split of the bar by the tilted plane.
Tests: oblique_cut_through_a_small_arc_matches_a_split replays examples/regressions/oblique_cut_of_cylinder.json and compares with solid.split_body by the same plane (fails on main). The full suite runs no slower.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
heal_keep evaluated the result's intersection curves outside any guard. truck panics where such a curve no longer evaluates, and the panic escaped boolean_whole, so the retries and fallbacks that would have succeeded never ran. heal_keep now runs under guard and returns the solid unhealed when healing panics (as its doc already promised for failures), and the boolean builds its result body under guard too, so a result that cannot be built counts as a failed attempt.
Tests: shallow_loft_cut_from_the_top_face replays examples/regressions/shallow_loft_cut_from_top_face.json and checks the exact prismatoid volume (fails on main with the panic).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The check is geometric, not a tuned threshold: on the reported basin (18 mm rim, sloped bowl wall) the rim strip reaches zero width at 15.17 mm and the shell is refused from there on. Tolerances only absorb rounding (scaled to the body size), and at most 2000 straight edges per face are compared pairwise. Arc edges and curved faces are not checked.
This is a guard, not the full fix. Building such a shell needs the offset to change topology: drop the strip that collapses and let its neighbours meet, as offset_body already does for rounds thinner than the move, extended to planar faces. That is substantially more work in the most delicate part of the kernel (rims, ledges, steps, arc-bounded strips) and is left for a separate change; until then the workaround is to model the hollow directly (an offset loft and a box cut).
kernel::fillet_variable_points(body, edge, start, [(position, radius)...], smooth): the radius law is a B-spline of the edge parameter (straight pieces, or C1 cubic pieces with Catmull-Rom slopes), and since every section's arc is affine in its radius the blend is exactly a rational B-spline (degree 2 across, the law's degree along) with B-spline contact curves on the faces. The largest radius along the edge decides whether the blend fits. Fillet type variable takes mid: [[position, radius expr], ...] and smooth; middle radii are parameters of the feature (Mid radius) and follow user parameters. Tests: straight pieces against (1 - pi/4) L (a^2 + ab + b^2) / 3 per piece, smooth against the integrated law, either end as start, refusals; engine test with a parameter-driven middle radius.
Refs #9
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RjGw4gA5ZvKrU8MQP6sBgT
`cargo xtask step-corpus --iges` (and `solvecraft-cli step-corpus --iges`) opens every oracle part.igs as File → Open does and compares body count, volume, area and face count with Fusion's measurements, writing docs/iges-import.md. The run found faces closing at an apex (a drill point's cone, a sphere's pole: the seam runs out to the apex and straight back) kept whole and meshed wrongly; they are now split along the iso-line opposite the seam like bands are (18-hole-features and 40-box-sphere-torus pass). The one failure left, 50-sweep-tight-bend, is written by Fusion as five loose, partly degenerate surfaces that Fusion cannot read back as a solid either; documented in the report. Test: a drilled blind hole round-trips through IGES.
Fixes#14
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RjGw4gA5ZvKrU8MQP6sBgT
- solvecraft-cli fuzz / fuzz-one: seeded random modelling sessions (boxes, extruded profiles with arcs, revolves; joins and cuts of boxes, cylinders and spheres, fillets, chamfers, shells, holes, patterns) on a grid that makes coincident faces and tangencies common, every step checked (command ok, every body a sound solid via the new Body::validity, volume moving the right way), plus the boolean volume identity on random primitives; one process per seed with a time limit; failures minimised to scripts
- shell/offset: corners where a curved face meets another at an angle are refined onto the moved surfaces (Newton); edges that are neither lines nor arcs are rebuilt through their moved points; one shell per piece; faces that would cross say so (not supported yet) instead of leaving a cavity inside out; flat revolved faces count as planes; the unhealed body is tried when the healed one fails
- cylinders start their seams at an odd angle (grid-aligned seams broke booleans along them)
- partial revolves of profiles touching the axis build no zero-area face on the axis
- booleans on bodies in several pieces retry a piece at a time
- regression tests for each, and a fixed set of seeds in cargo test
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
The merged face count no longer joins faces across a Split Face line, and Offset Face's same-surface grouping stops there. A piece whose move the body's topology cannot follow (its neighbour stays put) is swept along its normal and joined on or cut in, giving walls along the split line as in Fusion. Engine test: a split box top counts 7 faces; one half up 5 (27000 mm³), the other down 3 (25200 mm³).
Fixes#12
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RjGw4gA5ZvKrU8MQP6sBgT
Split Face records the new edges as polylines on the body (Body::split_lines); the document carries them through later features. STEP export no longer joins pieces of one surface across such a line, and healing (fillets, chamfers, booleans, shell, draft) no longer merges coplanar pieces there, while incidental pieces (boolean leftovers, seam halves) still merge as before. Tests: plane and cylinder splits keep their face count through a STEP round trip (and join without the record), a second split keeps both lines, and an engine test exports a split box after a fillet with 8 faces.
Fixes#11
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RjGw4gA5ZvKrU8MQP6sBgT
blend_corner.rs rebuilds the planar faces at the ends of the filleted edges (three planar faces and three straight edges at each such vertex) and keeps the rest. One round at a vertex ends in the third face's exact section of its cylinder; two rounds with the third edge sharp trim each other (rolling ball: the crease from where their contact lines cross to the third edge for equal radii, else to the smaller round's contact line on the face that ends the larger one, which gets the exact section between); three rounds meet a sphere for equal radii, else a rational degree-4 patch through plane sections of the cylinders, the sections turned so that the cylinders' curvatures agree where two meet on a face and the patch fitted tangent to all three (within 0.01 degrees) and to the face at its closed corner.
kernel::fillet falls back to it when the local blends refuse: an edge that ends on an earlier round (any radius) takes the round off (delete_faces) and blends both edges together, so r3 next to r2 and a third edge into a mitred pair work. delete_faces: a gap stretch along several deleted faces keeps the corners between them (two rounds meeting in a mitre come off back to the sharp corner). splitface::fit is shared for the crease.
Tests: volumes against the strips-minus-overlap reference (exact strips, overlap integrated; checked against an independent 1D integral), the sphere-octant formula for three equal rounds, tangency along the patch, closed meshes, refusals without panics.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New kernel file blend_variable.rs. The hooks into blend.rs are visibility
only: plane_normal, vtx, rebuild_face and unsupported are now pub(crate).
- fillet_variable(body, edge, start, r0, r1) works on the same edge
configuration as the constant fillet (straight edge, planar faces, end
faces square to it). The radius runs linearly from the end nearest
`start`, and every section square to the edge is a circular arc. The
blend is an exact degree (2, 1) rational B-spline: arc control points
move linearly.
- fillet_chord(body, edges, chord): the radius per edge comes from its
face angle (chord = 2 r cos(phi/2)).
Fillet feature: style constant | chord | variable {radius2, start};
FusionFilletEdgesCommand takes type, radius2 and start.
Tests:
- Variable 2 -> 5 on a box edge removes (1 - pi/4) L (r0^2 + r0 r1 + r1^2)/3.
Starting from the other end gives the same solid; equal radii give the
constant fillet; too large is refused.
- Chord 4 on a right angle equals radius 4/sqrt(2).
- Engine: both types by command.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
For solvecraft-params' Split Face feature:
- Sketch tools are chains of Seg2. A single line splits by its plane and a
single arc by its cylinder, exactly (a circular edge on a planar face);
longer chains and splines follow the curve.
- split_faces_with_map gives the source face of each face of the result.
- A listed face the tool misses is an error naming it (split_any keeps
the lenient form for Trim).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
Kernel extend(body, edges, distance) moves the picked straight edges of a
planar surface outward in its plane, with neighbours meeting them. Doc
feature SurfaceExtend, engine FusionExtendCommand. Tests: 10x10 extended
by 5 past one edge gives 150, past two gives 225; the engine patch, trim,
extend and thicken gives 240.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
Kernel (surfaces):
- Bodies with an open shell are surface bodies (Body::is_surface).
- patch_region: a planar patch of a sketch region. patch_edges: a planar
patch filling a closed loop of a body's edges.
- copy_faces: faces copied out as a surface.
- stitch: sew surfaces along coinciding edges, turning faces to agree; the
result is a solid when it closes.
- trim: cut a surface by a plane or face and keep one side.
- thicken: offset planar and circular-cylindrical faces along their
normals (one side or symmetric) into a solid. Plane copies slide;
cylinder copies scale about the axis, exact for NURBS circles. Free
edges get ruled walls. Faces meeting at a sharp edge are refused.
Doc features Patch, Stitch, Thicken (with operations) and SurfaceTrim;
engine commands FusionSurfacePatchCommand, FusionSurfaceStitchCommand,
FusionSurfaceThickenCommand and FusionSurfaceTrimCommand.
Tests (volumes):
- A 10x10 patch thickened by 2 gives 200, also symmetric.
- Half a cylinder side thickened outward gives half a ring; both halves
inward give the full ring.
- Six faces of a box stitched back give 6000; an open box with its top
patched from the edges, stitched, gives 6000.
- Trim keeps 30 or 70 of 100.
- Engine: patch, trim, thicken gives 140.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
split_body(body, tool) divides every face where the tool meets it and
closes each side with the piece of the tool surface inside the body.
- Patches have outer loops and holes.
- Bands go round a cylinder or another surface of revolution.
- Each side falls apart into its connected parts.
- A body the tool misses comes back whole.
Faces that are B-spline sheets but really planes or circular cylinders
(extruded faces) split as the whole plane or cylinder: the tool is
extended, and the field is exact and cheap.
Tests (volumes): a box split by a plane (2400/3600); a holed plate split
through its hole into two equal halves; a box split by a cylinder's side
into the core (pi*16*30) and the rest; a plane that misses.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
split_faces(body, faces, tool) divides the chosen faces where the tool
meets them and leaves the solid unchanged. It handles:
- lines across a face;
- closed lines inside a face (a disc inside a ring);
- lines that run into a hole (the face is divided around it).
Edges the line ends on are cut in every face that uses them. The tool's
zero set is traced on a finely divided triangulation of the face, each
point is placed exactly on the surface by bisection in its parameters,
the ends are placed exactly on the boundary edges, and a cubic is fitted
through the points (a line when they are straight).
Tests: a box top split by a plane (areas 80/120, later booleans still
work); a cylinder side split by a sketch line; a cylinder top split in
ring and disc by a smaller cylinder's face; a plate top split through
its hole into two equal halves; a tool that misses is an error. Volume
and area are unchanged in every case.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
- IGES import (kernel iges_in, from the published specification). The file
is restated as STEP and read by the STEP reader, so an IGES file becomes
the same Import base feature.
- Manifold solid B-reps (186/514/510/508/504/502).
- Trimmed (144), bounded (143) and bare B-spline surfaces, sewn into a
solid by shared edges, with face orientation made consistent.
- Curves: lines, arcs, ellipses, composite and rational B-spline curves.
- Surfaces: planes, analytic surfaces 190-198, B-spline surfaces,
surfaces of revolution and tabulated cylinders.
- Transformations, units, colours, names (406 form 15 or labels).
Open, insert and drop .igs/.iges like STEP.
- IGES export (kernel iges_out): each body as a manifold solid B-rep on the
surfaces a STEP export uses, with name properties and colours.
ExportCommand to .igs.
- OBJ import as mesh bodies: objects/groups, polygons fanned, negative
indices, bad lines reported.
- STL: binary files whose header starts with "solid" and those with a zero
triangle count read as binary; non-finite vertices dropped; facet
normals ignored; ASCII with non-UTF-8 names. Mesh bodies get consistent
winding (stray inverted facets turned over).
- Tests: primitives and features round-trip through IGES (volume, area,
name, colour). Six trimmed patches sew into a closed box. Hostile IGES,
OBJ and STL inputs never panic. STL quirks. The engine exports a design
to IGES and opens it again. All 76 oracle parts round-trip through IGES
with matching volume (local check).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
The read-back check (fallback to unmerged pieces) imported every export.
Now it runs only when some body actually had faces merged. Oracle 41
(42 faces, nothing merged): export no longer re-imports, which saved 7 s.
Pipe flange (merges): 1.1 s for the read-back.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
- Bodies carry a Paint for export (opacity, per-face colours), set by
Document::painted from the appearance model. The paint comes from a
face appearance, a body/component appearance's opacity or a material
look (Glass), and falls back to the paint the body was imported with.
- STEP: faces with colours of their own get a STYLED_ITEM on their
ADVANCED_FACE. See-through looks add SURFACE_STYLE_TRANSPARENT.
Differently coloured pieces are not merged into one face.
- STEP import reads face styles and transparency. The reader tracks the
file face of each kernel face (bands split in two keep it), so imported
face colours land on the right faces; the design shows them through
face_colors.
- 3MF: per-triangle colours (pid/p1 into the base materials, carried
through welding) and the object's alpha (#RRGGBBAA).
- Tests: a blue half-transparent cube with a red top exports, re-imports
with its opacity and the red top face, and keeps it when opened and
exported again. 3MF writes the alpha and the face entry.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
Fusion round trip WARNs on face counts:
- 48: the pulley's annular end faces arrived as 240° and 120° sectors.
- 54: the side faces and holes of a body glued to its mirror image came
in halves.
Changes:
- Planes merge: PLANE faces and flat B-spline sheets, as well as revolved
planes. A plane keeps several loops (an annulus, a face with holes), no
seam. Flat B-spline sheets are written as PLANE.
- Pieces of an extruded cylinder (B-spline sheets) merge when their edges
are circles square to the axis. The merged face is written as a
CYLINDRICAL_SURFACE, with its parameter seam on the face's seam edge.
- A full turn's loop is built explicitly: ring, seam, other ring (or the
apex on the axis), seam back. Before, it depended on edge order, and
after 81c85b5 the pulley's last pieces no longer merged.
- The read-back check moved to the whole file. If the merged file reads
with warnings and the unmerged one does not, the unmerged one is
written.
Results: 48 has 12 faces (Fusion 12), 54 has 12 (Fusion 12). Oracle
results are unchanged (73 PASS) and the STEP corpus passes 76/76. New
test: a plate with a hole glued to its mirror exports 8 faces, all
planes and cylinders, and round-trips.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
