* Boolean nudge tried both ways: no sliver where a cut's end touches a round side (#897, #827)
Robustness seed 14 cuts a cylinder whose end lies exactly on the apex of the
plate's arc side. Since the display tessellation plans (#805) the generator sees
the arc's true extent, so the cut's base lands on -10 instead of -10.000000000000002
and the end is tangent to the arc. The direct boolean fails there and the
last-resort tool nudge pushed the end into the arc, leaving a planar sliver a
tenth of a millimetre wide; shell then picked that sliver and refused it. The
13 October merges were not the cause (the bisect lands on #805).
The nudge is now tried both ways and the simpler result (fewer faces) is kept,
so the opposite nudge parts the touching faces instead. The same change joins
the hinge barrel in #827, whose top lies tangent to the lid's underside where
the mounting leaf ends. Each result still passes the same checks as before.
Goldens: seeds 14, 26 and 85 now build to the end, 31 and 78 come out with
fewer faces. Seeds 6, 43 and 73 still stop at shell: their joins now succeed and
shell reaches corners it can't offset (a cusp where a box face touches an arc,
coplanar faces facing opposite ways), which is unchanged.
Fixes#897Fixes#827
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Nudge the other way only when the first nudge fails or leaves a sliver
Trying both nudges and keeping the one with fewer faces could prefer a result
that keeps two pieces joined by a hair over one that separates them (the
centred coplanar revolve cut). The opposite nudge now runs only when the first
one failed or left a face narrower than 1/200 of the part, and it is taken
when it works with fewer of those.
Goldens re-recorded on current main: seeds 14, 26 and 85 build to the end,
31 and 78 with fewer faces. The other changed seeds are main's own since its
last full record, identical with and without this change.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Nudge the other way only for slivers the nudge made
On the sensor enclosure's lid the first nudge left thin faces the operands
already had (thin walls, small rounds), so the opposite nudge ran a second
boolean for nothing and the press/pull took 79 s against main's 55 s. The
opposite nudge now runs only when the first one failed or left more slivers
than the operands brought; the press/pull test takes 36 s against 42 s on main
(same machine, one after the other).
Goldens re-recorded on top of #940.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Goldens on current main: seeds 31, 43 and 78 with the two-way nudge
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Keep the hinge and tangent-cut tests beside the #871 fillet test
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Goldens: seeds 33 and 91 as main's fillet changes build them
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Goldens: keep main's #950 record
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
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 <[email protected]>
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.
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.
A coincident face is pushed apart only when the space it moves through holds nothing that changes the result; otherwise the boolean falls through to the exact strategies.
Co-authored-by: NOWN-DEV <[email protected]>
A cut or an intersection between prisms (a plate and the tools through it) is now decided in 2D before either operand is meshed. A body's size and volume are measured once, a placement reads its orientation from its matrix, the evaluation cache is bounded to 64 MiB, and STEP import copies fewer shells. Hole plates evaluate 20-30% faster with less memory, and models are unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A cut built slab by slab (the exact prism boolean) sewed the result from horizontal
slabs and only joins merged the coplanar pieces back. So a floor vent cut through the
Raspberry Pi case split the long wall into strips at the ports' tops and bottoms
(87 faces instead of 79 at 32 mm), and at 24 mm the end opening split the end wall.
- Prism booleans merge coplanar neighbours for every operation (faces only: edges,
curves and surfaces stay exact), keeping each face's construction source where all
its pieces share it, Split Face lines and regularization protection. Painted or
protected operands keep their authored pieces as before.
- The prism history now tells apart a wall's several edges with its caps (a wall
with openings meets each opening's floor), so cuts beside existing openings keep
exact history and faces keep their names.
- Tests: the Pi case's faces pinned after every feature at 24, at 24 changed to 32,
and built at 32 (each outer side one face); random openings and bosses on a
shelled case never split another side; every corpus design rebuilds a parameter
edit like a fresh build and reverts exactly.
- Action matrix baseline: demo/pi-case re-blessed (fewer, merged faces; no
regressions), demo/pi-case-0.8 recorded, and 88 new passes on
regressions/lid_ribs_flush_with_window.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Two L's extruded up and down from one 2 x 2 grid now join into one solid (the planar rebuild gives each fan of faces around a pinched vertex its own vertex copy), and a pad whose end lies half past the edge of the face it stands on fuses exactly: contact faces are imprinted with each other's outline and the common faces glued away, still held to the retry guard. Face splitting handles a line crossing an edge twice; glue matches differently parametrised curves and keeps several pieces apart. Robustness seed 16 now fuses with the exact volume.
Fixes#679, fixes#674.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Joining ribs flush with a lid's window left a zero-thickness slit far from the rib (the planar fallback glued back only one of the faces its pieces shared) and displaced the tool on the other joins. Gluing now joins every shared face pair, coincident-face separation checks there is material behind a pushed face for every join, and any join or cut that needed a retry must be closed, keep volume monotonic and leave the part unchanged outside the tool's box, or it is refused with an error. Feature command replies carry the feature's warning.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Fixes#396. A union that would come back as shells touching over an area is rejected and the face-in-face case is fused exactly; line and point contact stay as before.
When sample points say two bodies are apart or one is inside the other, the facetings are now checked for surfaces passing through each other; if they do, the boolean is computed instead of shortcut. If it still can't be intersected, the sampled answer is kept as a flagged, unverified result, and Interference lists such pairs as "could not be checked" with the reason. Fixes#442 and #443.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A Union dropped every construction certificate, so after any join no face
could take a durable sketch: sketch.create failed with "this face has no
exact planar source" even on a face the join never touched (#420, and the
"CREATE SKETCH: operation failed" reports #344, #195).
- kernel: map planar sources through a Union from both operands, with the
same exact-support rule Cut uses. Coplanar supports of the two operands
carry distinct identities, so a face merged from both stays without one.
- doc: Join extrude and Join box tools are certified like new bodies; a
cap moved off its designed height (into the target, or past a curved To
face) is not certified. Join producers are accepted when certifying.
Tests: both operands' faces attach and follow an upstream edit and a
reopen; a merged coplanar face is refused and fixed:true works; a face a
join covers whole is lost without moving to the new roof. Tests that used
a joined face as their "uncertified" example now use a merged face; the
trimmed-face case in a_sketch_on_a_face_stays_on_that_face now stays
attached to the surviving piece.
Claude-Session: https://claude.ai/code/session_01YDDPmNdRoBgkHcLeUDKVfb
Co-authored-by: Claude Opus 5.5 <[email protected]>
Pattern copies of a post joined to a body already holding the original post had
their tops in its plane, so every join first pushed faces apart (about 190 ms,
then discarded). Faces now coincide only where their boxes meet, the push-apart
retry runs once per boolean instead of twice, and splitting prism outlines skips
the sampled arc projection for points off the arc's circle.
Sensor enclosure evaluation: 3.07 s -> 1.63 s, same bodies.
Co-authored-by: Claude Opus 5.5 <[email protected]>
* Tangent hole contacts: every symmetric contact gives the same faces (#307)
Three things made the four symmetric contacts of reference case 42 differ:
- The 2D prism boolean's ray-parity test miscounted a ray grazing the bottom
of a patterned hole, whose two halves are read back with centres a few ulps
apart: one half counted the touch, the other missed it. The loop chaining
then failed and those cuts fell back to the general boolean, which treats
circles tangent within its tolerance as overlapping and splits the hole.
The ray is now moved off vertices and arc extremes by far less than any
modelling tolerance.
- The prism result's volume check used a coarser mesh than the operands, so
a plate with many holes read too large and correct results were rejected.
- The merged topology did not join a hole's two pieces across the line where
two tangent holes touch (four faces meet there).
Case 42 now gives 39 faces (33 holes, 6 planes); its expected.json pins it.
Fixes#307
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01RjGw4gA5ZvKrU8MQP6sBgT
* Kernel test: tangent hole pinch no longer splits the wall (#307)
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01RjGw4gA5ZvKrU8MQP6sBgT
---------
Co-authored-by: Claude Opus 5.5 <[email protected]>
union_apart joins bodies whose boxes are clearly apart (many pattern tools,
#261) without any intersection work: their shells side by side. Each face
and edge of the result is its piece's own, unchanged, and the result now
says so (a history over the pieces), so a pattern's cut keeps the exact
names of every copy's faces.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01J53sR6gEaVt56soFDpTck8
Fillets (local, loop, rolling ball, corner, variable, chord), chamfers
(cutters, local corners, loops, two sides), shell, draft, offset faces,
split face/body, delete face, patch/copy/stitch/trim/extend and
transforms record their history over [body] while they build: faces
rebuilt from the body's, blends in place of edges (corner patches in
place of the edges at the corner), shell walls as offsets, edges as
pieces of the body's or where two faces meet. Multi-step operations
compose each step. The push-face offset fallback now has history, and
the bore union composes its narrowed tool.
A reflection reorders a body's edges, so transforms now match edges by
identity (mirrored copies reported the wrong edges).
Refs #136
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01J53sR6gEaVt56soFDpTck8
ops::boolean's result carries its history over [a, b] whichever path made
it, recorded by the code that builds the faces:
- truck path (incl. jitter retries): shapeops history mapped through the
order-preserving copies and the cut's turned tool, then through healing
- heal records which input faces each healed face merges and which input
edges each joined or exact edge replaces (HealMap); opt-in, so heal_keep
callers pay nothing
- trivial shortcuts, the void, glue, faces pushed apart (push_faces records
its own history; the slab cut back on a union maps to the operands'
faces in that plane), the nudged retry, by_lumps and join composition
- prism booleans: section segments carry tags (operand cap edge, the wall
beside it, its vertices) through the 2D boolean; walls, caps and edges of
the extruded or sewn result follow from them
- planar and faceted polygon booleans: polygons carry their operand face;
faces join the pieces across edges, edges come from the pieces either side
History gains renumber/over (and consuming forms) for composition;
record gains regions/traced/build helpers. truck-shapeops
*_with_history now return the solid with an Option history. Paths that
cannot report (bore narrowing via offset, inexpressible lineage) leave the
result without history. Boolean results, tolerances and fallback order are
unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01J53sR6gEaVt56soFDpTck8
Bodies can now carry the exact history of the operation that made them:
for each result face its origins among the operation's operands (kept,
modified, split, merged, or generated from a profile segment, cap, edge or
offset) and for each edge the faces it separates and the operand edge or
face pair it came from. History is recorded by the operation, never
sampled, and checked against the body when attached.
- history.rs: the vocabulary (moved from provenance.rs), History::check,
identity, substitute (composes an operation with the one that made an
operand, including bodies made from a sketch) and face-source propagation.
- One edge order for a body (body::edge_order), shared by tessellation and
history, so history edge indices are the mesh's edge indices.
- Transforms report an identity history; placed_topology documents the
face/edge order guarantee it relies on.
- provenance.rs keeps the geometric trace for operations that do not yet
report history.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01J53sR6gEaVt56soFDpTck8
* test: reproduce #261 (Rectangular pattern of a hole cut scales super-linearly)
* Patterns of many hole cuts scale near-linearly (#261)
Two quadratic costs:
- Meshing a face tested every point against every segment of every boundary
loop, so a plate face with hundreds of holes meshed in quadratic time. The
vendored truck-meshalgo inside test now skips loops whose parameter box is
clear of the point (they wind zero times around it).
- A pattern cut each copy in turn, rebuilding and re-checking an ever more
complex target per copy. Copies whose boxes are clear of each other are now
one tool of many pieces (kernel::union_apart, no intersection work), cut once.
512 holes now take ~7 s instead of not finishing. The test's volume tolerance is
widened from 1e-6 relative to 0.1 % of each hole: the measure meshes the bores
and reads each hole ~0.06 % small, on main too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Hole-pattern test: a budget CI can meet; coarser apart-box meshes (#261)
CI runs every test binary at once and is about ten times slower than a
machine running this test alone (15 tests passed 60 s in the last run), so the
512-hole pattern (about 6 s alone) hit the 60 s timeout there. The budget now
only has to tell linear from the quadratic hang: 300 s timeout, 120 s budget.
union_apart meshes each copy coarsely (2 % of its size, box grown by twice that).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Hole-pattern test checks the scaling, not a wall-clock time (#261)
CI runs every test binary at once: the 512-hole pattern (about 6 s alone, 10 s
CPU single-threaded) took 177 s there, so no fixed time is reliable. The test
now patterns 128 and then 512 holes and requires the second to take under 8x
the first (linear is 4x, measured 3.5-4.2x here; the old quadratic cost was 16x
or more), each with the volume check and a 900 s hang guard.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
---------
Co-authored-by: bherbruck <[email protected]>
Co-authored-by: Claude Opus 5.5 <[email protected]>
Record ownership transfers in the timeline so earlier Join/Cut membership,
world placement, machining data and styles survive component moves, edits,
independent copying, undo and reopening. Preserve pending positions in history
snapshots and keep constructor transfers ordered during feature redefinition.
Retain shallow sheet pockets and exact planar intersection islands through
unfold/refold, with depth-aware flat previews and DXF export. Store explicit
3D wire coordinate spaces and migrate known legacy provenance without moving
saved points or analytic fits.
Infer unique eligible cylindrical Emboss targets, align Revolve manipulators
with actual extent geometry, and discard obsolete preview errors after input
correction. Certify constructed and immutable imported face styles using exact
source metadata while preserving legacy face names. Upgrade designs to format10
with bounded validation and geometry-preserving legacy migration.
Persist typed sketch dimensions and bounded ordered configuration rules in format 8. Use exact occurrence paths and source identity for editing, movement, appearance, and mass. Preserve sheet material regions through cutting, folding, and exact DXF export; retain nested cavities and material islands.
Validation: cargo xtask ci passed, including 368 engine tests, 178 UI tests, 68 UI scenarios, 136 kernel tests, and the 10-crate wasm check. Actual preview workflows and legacy format-7 origin-corner reopening were inspected.
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 <[email protected]>
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 <[email protected]>
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 <[email protected]>
- 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 <[email protected]>
Claude-Session: https://claude.ai/code/session_01AS5Tn3xieewUSg3uxrMVdL
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 <[email protected]>
Claude-Session: https://claude.ai/code/session_01RjGw4gA5ZvKrU8MQP6sBgT
phanoob