9e2ef2c94dea3f1267deba570c9a423fb38c2f85
2 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Tom Boucher
|
a28dcec981 |
chore(#597): replace count-based ratchet guards with AST lint + named-set allowlists (#603)
The windows-test-parity ratchet greps test source for fs.rmSync-without-
maxRetries (and six other Windows-portability anti-patterns), failing when an
integer offender COUNT exceeds a frozen baseline (rmSync: 95). A count ratchet
is a Goodhart metric: fixing one offender and adding another keeps the count
constant, so a new defect slips through green. Replace it — and every other
count ratchet in the repo — with a layered, masking-proof design.
Behavioral seam test
- tests/helpers-cleanup.test.cjs proves helpers.cleanup() carries the Windows
EBUSY retry budget. cleanup() delegates retries to Node's fs.rmSync via
maxRetries (it owns no loop), so the test asserts the option contract
(recursive/force/maxRetries>0/retryDelay>0) + real-FS removal + the cwd-guard,
rather than a loop that does not exist. The EBUSY risk is now tested ONCE at
the helper, not approximated textually at every call site.
Write-time ESLint rule (AST-accurate, replaces the grep)
- eslint-rules/no-raw-rmsync-in-tests.cjs (error in tests/**/*.test.cjs) bans
raw fs.rmSync, steering to cleanup(). Catches member, computed (fs['rmSync']),
destructured and aliased forms; escape hatch is inline
`// eslint-disable-next-line local/no-raw-rmsync-in-tests -- <reason>` only.
- Migrated 336 raw fs.rmSync teardown calls across ~116 test files to cleanup().
~18 genuinely load-bearing sites (mid-test SUT/fault-injection removals,
error-swallowing or name-colliding local teardown helpers) keep the raw call
with an inline eslint-disable + reason.
Shared anti-ratchet primitive
- scripts/lib/allowlist-ratchet.cjs:
- assertWithinAllowlist: fails on NOVEL ids (new offender introduced) AND on
STALE ids (a known offender was fixed but not pruned) — identity, not count,
and a ratchet DOWN toward zero.
- assertTightCeiling: a size/length budget whose ceiling must stay within a
grace band of the high-water mark, so budgets may only tighten, never creep.
Ratchets converted onto the primitive
- windows-test-parity-guard.test.cjs: rmSync rule deleted (now ESLint-enforced);
the remaining six patterns moved from integer baselines to named-set
allowlists with ratchet-down.
- scripts/lint-test-file-count.{cjs,allowlist.json}: per-module integer counts →
named filename sets (closes the swap-a-file-keep-the-count blind spot); a
module dropping under cap now FAILS to force pruning its allowlist entry.
- enh-2790 skill-count `<= 63` → named skill allowlist (ratchets toward ~58).
Size budgets hardened (tighten-only)
- agent-size / workflow-size / feat-3039 help-tiered: ceilings lowered to the
current high-water mark and an assertTightCeiling anti-creep check added per
tier. Fixed external-contract limits (description ≤100 chars, agent ≤100 KB)
are intentionally left as-is — they are not grandfathered creeping budgets.
No user-facing behavior change (tests + tooling only); no USER_FACING_PREFIXES
touched, so no changeset fragment is required.
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
|
||
|
Tom Boucher
|
619b5c42e3 |
fix(3407): snapshot old release into gsd-pristine, not new (#87)
* fix(3407): snapshot old release into gsd-pristine, not new saveLocalPatches() was wiping gsd-pristine/ then re-populating it from pristineCtx.packageSrc — the NEW release source tree. For files that changed between old and new releases, this wrote NEW-release bytes as the pristine baseline while backup-meta.json recorded OLD-release hashes. The resulting hash mismatch triggered the #3657 verifier guard on every such file, causing it to skip the three-way diff baseline and fall back to the over-broad heuristic — effectively nullifying the #2998 feature for any file that changed upstream. Fix: preserve existing gsd-pristine/ entries that are already correct (sha256 on disk matches originalHash from manifest). These were written by the previous install with old-release bytes and remain valid. For files where no correct entry exists, leave gsd-pristine/ absent so the verifier falls back cleanly to over-broad mode — safe, never false-fails. OK_PRISTINE_DRIFT_DETECTED (added by #3657) is intentionally kept: it guards installations that already have drifted pristine from pre-fix runs and protects against other future stale-pristine scenarios. It is not removed because its guard is correct; only its trigger frequency drops. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * chore: add changeset for #3801 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(3407): hash-validated regeneration for missing gsd-pristine entries Codex adversarial review found that the #3407 fix left absent gsd-pristine/ entries permanently absent, causing persistent over-broad verification even when the file was unchanged between old and new releases. Add selective regeneration: for entries absent from gsd-pristine/, generate a candidate via populatePristineDir into a temp dir using new-release source, then only promote if sha256(candidate) === originalHash. When hashes match, the file was identical across releases so new-release bytes ARE the correct old-release pristine. Discard mismatches — over-broad fallback applies. Add regression test asserting regeneration occurs for unchanged-between- releases files whose pristine entry was absent. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix(3407): address review — restore mkdtempSync resilience, tighten tests, fix counter accounting Addresses sonnet adversarial MAJOR (mkdtempSync outside try/finally caused uncaught exception on broken /tmp), MIN-01–MIN-05 (misleading prose, counter double-count, missing stale-pristine test, permissive assertion, vacuous antipattern-hunt), sonnet MINOR (rmSync EISDIR), NIT (unused import, test count). F1: move mkdtempSync inside try block with catch/warn for graceful degradation F2: fix misleading prose — gsd-pristine/ is populated lazily by saveLocalPatches, not a separate install-time step F3: fix counter double-counting — track stalePaths/regeneratedPaths as Sets; removed = stale NOT regenerated (non-overlapping counts); update log message accordingly F4: add stale-pristine recovery test — pre-populates gsd-pristine/ with new-release bytes (exact pre-fix bug artifact), asserts absent after fix run F5: tighten Test 3 assertion from permissive if(exists)/notEqual to strict assert.strictEqual(exists, false) F6: remove vacuous antipattern-hunt describe block (typeof checks only, no behavioral coverage) — rationale noted in comment F7: use fs.rmSync with force:true/recursive:true for EISDIR resilience; only count removed after confirming file is actually gone via existsSync F8: remove unused afterEach from destructured import F9: test count updated to reflect 4 tests in describe block Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> |