The general 3-D outcrop cap: a zone may meet any boundary

[lmoresi]

Stacked on #582 (base is its branch); only the last three commits are this PR.

A fault zone specified past the mesh boundary now embeds whatever that boundary is: a tilted or rotated wall, a box edge, a faceted sphere. The NotImplementedError at the old _trace_wall_code — "the zone outcrops on a non-axis-aligned boundary" — is gone, and the function with it.

What the cap is now

The cap over the outcrop bowl is re-triangulated per coplanar boundary region, each piece meshed flat in its region's own plane, so the cap sits exactly on the faceted surface and volume conservation survives to 1e-12. Where the band crosses a crease between regions, the two sides segment the crease line differently — the bowl's nodes are mesh vertices, the band's crossing nodes are assembly nodes — so a 1-D overlay merges both node sets by arc coordinate and keeps each elementary sub-segment for the side(s) whose bowl covers it and the band does not. A sub-segment covered from one side only must be a whole mesh edge, because a cavity-shell face matches it in the fill's closed surface; anything else is a refusal, not a crack.

The frame rule (_outcrop_frame_3d) separates two notions the third dimension forces apart. Where the cavity may open is by smooth wall: coplanar regions grouped across low-dihedral creases (under 45 degrees), so a faceted sphere is one wall and its bowl legitimately spills onto facets beside the band's own, while a box's other walls stay refused across their sharp edges. What the carve may delete is by shape: band-covered vertices, single-region interiors, and interior vertices of straight creases between touched regions. A vertex where three or more regions meet — a box corner, every vertex of a faceted sphere — is the domain's shape and is never deleted off the band. A protected vertex whose whole cell star drops is not stranded: it is a collar node, and the gap fill's tets reference it. That is how a curved boundary keeps its faceting through the surgery.

Wall labels are restored per removed face (the 2-D per-segment rule of the corner fix, one level up), so a bowl spanning Top and Front restores each wall's own labels. The Euler gate now demands conservation of the input's Euler number rather than the value 1: a spherical shell is S²×I, Euler 2, and was refused before for its topology rather than for any defect. The same assumption remains in place_sheet and remove_embedded, marked TODO(BUG).

Evidence

• The box oracle did not move: the full placement suite (0853–0856, 0859) passes unchanged, 62 tests.
• New test_0860_outcrop_general_boundary_3d: a rotated box (no wall axis-aligned), a band across the Top/Front box edge (trace on both walls, each wall's labels restored on its side), and a spherical shell outer surface. Each runs its negative control first — the identical census on an interior twin counts no trace — and each ends in the P2 Poisson oracle, exact through the zone.
• Parallel: the rotated-box outcrop added to ptest_0855 passes at np=2 and np=4; the trace count (28 facets) is identical at np=1, 2 and 4 with the info dict identical on every rank.

Known limits, stated as refusals

• A band running along a domain crease (rather than across it) refuses with the reason.
• A band outline node landing within rounding of a kept mesh vertex refuses: the 2-D imprint collapse (_collapse_boundary_imprints) has no 3-D counterpart yet.
• place_sheet's outcrop stays box-framed, per the plan's own advice — unify after, not during.

Underworld development team with AI support from Claude Code

[lmoresi]

Adversarial review (self)

We reviewed the three commits against the plan's gates and our own claims. Findings, most severe first.

1. A rank-local raise shipped in the first commit and would hang at np≥2. The collar-vs-bowl consistency check raised on the surgery rank inside the relabel block, which is outside any try — exactly the defect class the file's own comments warn about. Found by re-reading the diff, not by any test: the parallel tests pass because the check never fires on healthy geometry. Fixed in the third commit by routing through the block's collective failure. A test that forces this refusal at np≥2 does not exist and would need deliberately inconsistent geometry; we did not construct one.

2. The general outcrops are 2–4× more slivery than the box outcrop. Minimum dihedral angles, same mesh scale (cellSize 0.12, width 0.05): interior zone 9.5°, axis-aligned box outcrop 13.8°, rotated box 6.4°, box-edge crossing 3.6°, spherical shell 3.1°. The mechanism is the one the 2-D corner work already met: the band outline lands where it lands, and a collar polygon between an outline node and a nearby kept mesh vertex is thin. The 2-D cure (_collapse_boundary_imprints) has no 3-D counterpart; until it does, near-coincidence refuses at rounding distance but produces slivers just above it. All fills pass the conformity gates and the P2 oracle is exact — the slivers are a quality cost, not a correctness one.

3. A dead end we built and removed. Our first treatment of stranded protected wall vertices kept one cell of the vertex's star ("rescue"). It produced floating kept islands the fill wrapped as bubbles (global Euler defect, measured on the sphere) and was replaced wholesale: a protected vertex needs no kept cell, because the collar re-triangulation references it and the gap fill's tets give it cells. The final code is simpler than what it replaced; the connectivity gate we wrote for the islands went with the rescue that caused them.

4. The Euler gate was wrong before this PR, for domains nobody could reach. != 1 encodes a ball; a spherical shell is Euler 2. The general clip made shells reachable, so the thin-volume gate now demands conservation of the input's number. The identical assumption remains live in place_sheet and remove_embedded — a sheet placed inside a spherical shell refuses today for its topology. Marked TODO(BUG) at both sites rather than fixed, per scope discipline.

5. Untested paths we know about. The polygon-nesting logic in _outcrop_collar_3d is exercised only at depth ≤ 1 (outer ring plus hole); depth-2 nesting (an island inside a hole) is written but unreached. The three refusal guards — band along a crease, outline node within rounding of a kept vertex, a band node splitting a one-side crease edge — fire on geometry none of our cases construct; they are read-verified only. The 45° smooth-wall threshold means a curved boundary faceted coarser than ~45° facet-to-facet splits into separate walls and refuses; documented at the constant, not tested.

6. Partition independence is asserted, not proven, for the sphere. The np=1/2/4 identity (trace 28) is measured on the rotated box only. The sphere runs serial-only; its parallel behaviour rests on the same gather-first mechanism the box exercises, which is an argument, not a measurement.

Underworld development team with AI support from Claude Code

[lmoresi]

Review finding 4 is resolved in 41eb60a rather than left as a TODO: place_sheet and remove_embedded now gate on conservation of the input's Euler number, the same fix place_thin_volume carries. A sheet embeds mid-gap in a spherical shell and removes again, pinned by a regression test in test_0854. Every placement primitive now works in non-ball domains.

Underworld development team with AI support from Claude Code

[lmoresi]

Review addendum: what a 1000 km outcrop band taught us

We drove the general cap at Earth scale — a 1000 km, 4 km-wide megathrust band against a spherical shell, background mesh ~350 km — and it found four defects the 300 km tests could not see. Three are fixed on this branch, one is characterised and open.

Fixed. (1) OCC's boolean mass on a many-faceted tool is only accurate to ~5e-7 relative, so the 1e-9 meshed-vs-CAD gate refused correct assemblies; the gate now runs before the snap and allows 1e-6 (807a912, restructured in db68864). (2) OCC leaves clipped nodes up to ~4e-7 off the tool's own planes; the snap now covers that noise, masked to the boundary facets so infinite-plane projection cannot reach interior nodes (34ea192). (3) The band outline grazes the wall's own vertices — measured, a 2.4 m gap the collar meshed as a sliver — and _collapse_boundary_imprints_3d now reroutes the outline through such vertices, with a volume guard so a move or split can never flatten a cell; on-crease outline edges are reassigned to the far collar rather than refused (db68864). All 31 placement oracles pass throughout.

Open, characterised. Refining a spherical boundary projects new facet vertices onto the true sphere; the creases between the resulting nearly-coplanar sub-facets are laterally fuzzy at (placement noise)/sin(theta) ~ 1e-4 — wider than a 4 km band's own feature spacing — and the crease overlay cannot classify crossings against them. A fuzz-aware tolerance was tried and withdrawn: it traded the sliver for misclassified chains on the honest-crease tests. Until the crease representation is rethought, outcrops on adapt-refined spherical boundaries refuse at the fill's own gates; the plain-shell 1000 km case also still trips a skin/cap facet overlap that is not yet diagnosed. Negative results reported as found; the collar refusal now names its pinch (gap and node kinds), which is the diagnostic that separated all four causes.

Underworld development team with AI support from Claude Code

[lmoresi]

Adversarial review — fault-placement stack (#582, #589)

Reviewed together. Each already carries its own review; this adds the
relationship between them, which neither says.

S1. #589 is stacked on #582. pr582 is an ancestor of pr589. Both edit
src/underworld3/utilities/place_surface.py and
tests/parallel/ptest_0855_place_thin_volume_parallel.py; #589 adds
test_0854_place_sheet.py and test_0860_outcrop_general_boundary_3d.py on top
of #582's test_0855_place_thin_volume.py and test_0859_domain_boundary_tool.py.

Consequences for the backlog:

• Merging The general 3-D outcrop cap: a zone may meet any boundary #589 merges Clip fault assemblies against the mesh's own boundary; label the outcrop trace #582. Merging Clip fault assemblies against the mesh's own boundary; label the outcrop trace #582 first leaves The general 3-D outcrop cap: a zone may meet any boundary #589 as the general 3-D
  cap alone, which is how the work was scoped (Outcrop on curved boundaries: the wall descriptor is (axis, value) everywhere #555's remaining piece).
• The reviews are not independent evidence. Anything approved on Clip fault assemblies against the mesh's own boundary; label the outcrop trace #582's
  place_surface.py is being approved again as part of The general 3-D outcrop cap: a zone may meet any boundary #589, and a finding
  raised on Clip fault assemblies against the mesh's own boundary; label the outcrop trace #582 and fixed on The general 3-D outcrop cap: a zone may meet any boundary #589 will look unaddressed on Clip fault assemblies against the mesh's own boundary; label the outcrop trace #582.
• Whichever merges second needs no rebase, but Clip fault assemblies against the mesh's own boundary; label the outcrop trace #582 will show as fully merged
  rather than closing on its own commits.

Recommended order: #582, then #589.

S2. Both touch ptest_0855, which is a ptest_ file. Those are not
collected by a default pytest tests/ run — they are driven under mpirun
explicitly. Two PRs editing the same manually-driven parallel test is the case
where a regression is least likely to be noticed by CI. Worth confirming the
combined file runs at np=2 and np=4 after both land, rather than after each.

Underworld development team with AI support from Claude Code

[lmoresi]

The sheet path joins the general machinery (111ae09..24cbd56)

place_sheet now clips, carves and caps against the mesh's own boundary — the Sutherland–Hodgman box clip, the wall-code frame and the single-loop rim cap are deleted. The discrete clip cuts the authored triangulation directly (no OCC): per-component outward orientation, signed-distance classification, sequential cuts by the coplanar regions' planes, cut nodes interned by (original-edge, region) identity. The carve takes the frame's two masks like the volume's carve; the collar embeds the trace chain through its pieces instead of cutting a hole, with crossings on the creases via the existing 1-D overlay. The trace chain's edges carry <label>_trace through to the wall — the splittable trace the split machinery needs for a daylighting fault (test_0861 holds the kinematic acceptance criterion, gated on the split side's refusal until feature/fault-split-node lands).

Evidence: box oracle unchanged (test_0854 8/8; clip area matches Sutherland–Hodgman bit-for-bit); the spherical sheet outcrop that previously could not run at all embeds with the trace labelled, volume conserved to 1e-12 and the P2 oracle exact (test_0860, 4/4 with interior negative control); trace counts identical at np=1/2/4; full suite 1505 passed.

Adversarial review (our own)

Defects found and fixed before the push:

• Coplanar facets tore the sheet. Cutting per facet plane keyed the two triangles sharing a cut edge by different (coplanar) facets of one wall — a duplicated node on the box top wall. Cuts now key and compute by the coplanar REGION's plane.
• Sub-segment interpolation tore the sphere. A side truncated by an earlier plane interpolated its next cut from the sub-segment, its neighbour from the full edge: ~1e-17-apart duplicate pairs. Side cuts now re-derive from the original sheet edge's endpoints.
• The collar's overlay shadowed the new chain parameter (for chain in chains.values()), so the VOLUME path read a stale crease chain as a sheet trace and indexed with tuples — caught by test_0860's box-edge case, renamed crease_chain.
• petsc4py segfault in our own census: getStratumIS on an empty stratum hands back an IS whose getIndices SEGVs (PETSc traps it, so it presents as a silent pytest death, exit 59). Guarded by getStratumSize first — the same trap class as the Null_Boundary lesson.
• Pre-existing red test: 34ea192 widened the snap tolerance to 1e-6 and left test_0859's untouched-control at 2e-9, inside it. A branch push does not trigger CI, which is how it shipped red. Fixed (24cbd56).

Known limitations, deliberately loud rather than fixed here:

• A crossing that can touch a locally concave crease (an inner boundary) refuses — the polyline cut is not built. On adapted spheres with fuzzy creases the refusal may now fire at the clip rather than the fill's gates; same posture as the crease-classification redesign note, earlier trigger.
• Tolerances are absolute (clip 1e-12, membership 1e-9), consistent with the subsystem; at 1000 km-in-metres scale the rounding floor approaches them. Campaign-wide, not new here.
• The bbox near-set is a superset heuristic: an over-keep is caught by the every-node-inside exit gate (refusal); an over-clip cannot happen where the boundary is locally convex; a torn configuration surfaces at the fill's PLC as a loud error. No silent path found.
• The shell sheet outcrop's min cell volume is ~1e-10 — the sliver follow-on already recorded for general outcrops. Sheets have no 3-D graze collapse; the collar's near-coincidence refusal is the guard.

🤖 Generated with Claude Code

https://claude.ai/code/session_01WDKP73LD3dCFQSNgEXbbw9

[lmoresi]

Merging as the second half of the #582 stack (#582 landed as 87232f0). Its CI was red when we reviewed; it re-ran against a development that now contains #582 and is green — the failure was the stack, not the change.

Per S2 in the cluster review, both PRs edit ptest_0855_place_thin_volume_parallel.py, which is a ptest_ file and so is not collected by a default pytest tests/ run. The combined file is worth a run at np=2 and np=4 now that both halves are in, rather than after each.

[lmoresi]

The daylighting ruling lands (aa79ec5, ae7f9f6)

Maintainer ruling (2026-08-18): split-node faults do not split through the surface. A fault stops an element or two below, blind, with a damaged region above carrying the deformation to the surface. This supersedes the split-through-the-wall acceptance test_0861 first carried; the body gated on a daylighting split is removed.

The kinematic acceptance is now validated by running, not gated. The listric handover rule (a contact tip inside weak material is free; a tip at the weak zone's edge is pinned) turned vertical against a FREE surface: a blind frictionless split fault two cells under the top, driven in antisymmetric dip-slip, solved bare / damage-abutting / damage-enclosing. Measured (probe in ~/+Simulations/blind_fault_surface_expression/): near-tip slip 0.45 / 0.65 / 0.80 of peak, surface localization 37% / 44% / 50% — strict ordering on both, enclosing the tip worth 1.77x over bare pinning, and the abutting seam (the hybrid-seam defect) measurably worse than the enclosed tip. The overlap margin is the deliberate design parameter.

place_sheet(..., setback=d) does the deliberate part: the job is still to figure out the intersection so the fault can stop before it hits it — it just never meshes or splits all the way there. The sheet clips against the boundary offset inward by d, arrives blind with a strictly interior rim, and split_along_label_3d takes it as it stands; the would-be intersection with the true boundary comes back as info['surface_trace'] — the damage region's locator. Verified end-to-end in test_0861: place with setback, trace located on the true wall, shallowest sheet vertex exactly a setback below it, split succeeds with the Plus side carrying every placed face.

Adversarial notes on our own additions: an all-rim corner face (two side-rim originals + cut nodes) admits no triangulation the split accepts, at any sheet density — fixed by longest-interior-edge midpoint splits. A centroid split was tried first and rejected on measurement: inside a sliver corner face it drove the gap fill to 1e-23 cell volumes and a P2 error of 0.64 on the shell outcrop, caught by the P2 oracle after every topological gate passed — the #520 lesson holding again. The blind-fault probe's margins are modest at unit scale (the far field dominates a 1x1 box); the committed thresholds hold generous slack under the measured values.

🤖 Generated with Claude Code

https://claude.ai/code/session_01WDKP73LD3dCFQSNgEXbbw9