Skip to content

Next Gen STEP AP242 Native-ish BREP Geometric Modeling Kernel

docs: record V2-V5 triangle migration blocker and revert decision

Yuechen Li93d4e3b10ec2
diff --git a/docs/frictionlab/v2-x8-triangle-hex-prism-profile-parity-lab.md b/docs/frictionlab/v2-x8-triangle-hex-prism-profile-parity-lab.md
index 814a7733..a3e1a844 100644
--- a/docs/frictionlab/v2-x8-triangle-hex-prism-profile-parity-lab.md
+++ b/docs/frictionlab/v2-x8-triangle-hex-prism-profile-parity-lab.md
@@ -56,3 +56,8 @@ Current lab rows are designed to provide bounded parity evidence only. Productio
- If parity rows remain fully green: proceed to V2-V5 triangle production migration candidate.
- Then V2-V6 hex production migration candidate.
- If mismatch appears: harden emitter/convention mapping before production migration.
+
+
+## Update note (consumed by V2-V5)
+
+Triangle parity evidence from this lab was evaluated in V2-V5; production triangle migration was reverted after downstream Firmament chamfer parity regression, so lab evidence remains preparatory rather than consumed for final routing change. Hex parity rows remain continuity evidence for a future bounded migration milestone.
diff --git a/docs/v2-a2-prismatic-operation-migration-audit.md b/docs/v2-a2-prismatic-operation-migration-audit.md
index 35489a85..1518fd35 100644
--- a/docs/v2-a2-prismatic-operation-migration-audit.md
+++ b/docs/v2-a2-prismatic-operation-migration-audit.md
@@ -210,3 +210,8 @@ This milestone explicitly does **not** do:
## Update note (V2-X8)
V2-X8 lab-only parity probe added for triangle/hex prism baseline-vs-emitter comparison (line-only loops), with deterministic topology + STEP smoke diagnostics and invalid input rejection rows. No production routing changed.
+
+
+## Update note (V2-V5)
+
+V2-V5 attempted to consume V2-X8 triangle evidence for production migration, but reverted after Firmament bounded-chamfer regression exposed a seam-level topology/adjacency parity blocker. Triangle production routing remains unchanged pending blocker resolution.
diff --git a/docs/v2-v5-triangle-prism-profile-extrude-production.md b/docs/v2-v5-triangle-prism-profile-extrude-production.md
new file mode 100644
index 00000000..7d337022
--- /dev/null
+++ b/docs/v2-v5-triangle-prism-profile-extrude-production.md
@@ -0,0 +1,53 @@
+# V2-V5 — Triangle prism production migration via line/arc profile extrusion
+
+## Purpose and scope
+V2-V5 evaluated production migration of the triangle prism route to `LineArcProfileExtrudeEmitter` under strict parity constraints.
+
+## References
+- V2 doctrine: `docs/aetheris-v2-sweep-first-architecture.md`
+- V2-A1 resolved profile contract: `docs/aetheris-v2-a1-resolved-profile2d-contract.md`
+- V2-A2 migration audit: `docs/v2-a2-prismatic-operation-migration-audit.md`
+- V2-X8 parity lab: `docs/frictionlab/v2-x8-triangle-hex-prism-profile-parity-lab.md`
+
+## Outcome
+**Acceptable Outcome B (honest stop): migration attempt was reverted.**
+
+## What was attempted
+- A Firmament execution seam migration was prototyped for `triangular_prism` to emit a one-loop, line-only profile through `LineArcProfileExtrudeEmitter` using the required coordinates:
+ - `(-w/2, -d/2)`
+ - `( +w/2, -d/2)`
+ - `(0, +d/2)`
+- Centered Z semantics were preserved in the attempted route.
+- No 3D boolean operation was introduced in the attempted triangle route.
+
+## Blocker classification
+1. **Topology parity regression** and
+6. **LineArcProfileExtrudeEmitter integration bug (for this production seam)**.
+
+Observed effect:
+- Non-orthogonal triangular prism chamfer scenarios regressed (`No bounded corner-resolution candidate was admissible`) in existing Firmament gates, indicating corner/adjacency semantics used by downstream bounded chamfer recognition changed relative to the legacy triangle primitive path.
+
+## Decision
+- To preserve production behavior and all hard constraints, the triangle production route was reverted to the existing `BrepPrimitives.CreateTriangularPrism(...)` path.
+- No weakening of topology/STEP assertions was applied.
+- No STEP exporter or boolean core changes were made.
+
+## Validation/admissibility boundary
+- Existing deterministic invalid-dimension behavior remains unchanged in production.
+
+## Tests run
+- Focused Core/Firmament/FrictionLab filtered gates were run.
+- Core and FrictionLab focused suites passed.
+- Firmament focused suite exposed the blocker and remains the evidence source for this stop.
+
+## Non-goals (unchanged)
+- No hex migration in V2-V5.
+- No slot/capsule migration.
+- No generic polygon prism migration.
+- No STEP exporter/importer changes.
+- No Boolean core changes.
+- No sketch solver, clipping engine, or NURBS/freeform additions.
+
+## Next recommended migration
+- Resolve triangle chamfer adjacency/parity at the emitter-integration seam first.
+- Keep V2-X8 hex parity continuity; then proceed to likely V2-V6 hex migration only after triangle seam parity strategy is proven safe.