sim c444051bef test(#3339): fix orthogonal-review findings — Wave 7 fold ...

Standards/Spec-axis review + Memtrace graph pass found real issues in
the just-folded suites, all fixed here:

- tests/phase.test.cjs: the issue explicitly asked to dedupe overlapping
  fixtures between issue-2945/issue-2949 — the fold preserved both test
  sets correctly (genuinely distinct code paths, 0 tests dropped, already
  verified correct) but left each fold block with its own near-identical
  copy of a runVerifiedPhaseComplete(args, tmpDir) helper, the fixture-
  level dedup the issue actually asked for. Consolidated into one shared
  definition without disturbing the file's other, unrelated same-named
  helper at module scope (would have collided if hoisted directly).
  Removed two now-unused local runGsdTools destructurings left behind by
  the consolidation (both blocks already close over the module-scope
  import at line 26).
- tests/review-lane-descriptor.test.cjs: two .find() results dereferenced
  without a presence guard (same defect class Wave 3/#3335 already found
  and fixed once in this epic) — added assert.ok() guards matching the
  repo's established style.
- tests/model-resolver.test.cjs: documented the 1-of-80 dropped duplicate
  test with an inline comment, matching this same wave's host-integration
  fold's convention of citing drops by exact reference instead of leaving
  a reviewer to reconstruct the justification via git archaeology.

No test() count changed in any file. No production code touched.

2026-08-12 08:25:14 -04:00

247 KiB

Raw Blame History

View Raw