Repository navigation
test(doc-count): enforce the count guard in the cold gate - #752
Conversation
…en-remembered) scripts/check-doc-counts.sh hardened to 9/9 form coverage (#738/#749/#750/#751) but only ran when someone remembered to. Wire it into test:offline via a thin test that runs the guard and asserts it passes — so a Move test added WITHOUT 'npm run sync:count' fails 'npm test' (with that one-command fix surfaced in the failure), instead of a judge running 'sui move test' and seeing a number that mismatches a doc. Mirrors test:doc-links (the doc-link consistency check already in the gate). The fleet's count flurry has settled at 695, so this won't cause transient mid-sync failures. Verified: test passes on the consistent tree; test:cold-gate-coverage accepts the new wired test; package.json valid. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b500e47839
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| let out = ""; | ||
| try { | ||
| // The guard cd's to the repo root itself; run from there for good measure. | ||
| out = execFileSync("bash", ["scripts/check-doc-counts.sh"], { |
There was a problem hiding this comment.
Accept four-digit Move test counts
When the Move suite reaches 1000 tests, this new cold-gate test will fail even after the documented npm run sync:count fix: sync-move-count.mjs already accepts and writes 3–4 digit counts, but scripts/check-doc-counts.sh only extracts [0-9]{3} groups, so synced text like 1000/1000 is parsed as 000 and 100 and exits with a disagreement. Since this invocation is now part of npm test, please make the checker accept four-digit counts before gating on it.
Useful? React with 👍 / 👎.
…) + add SAFETY.md ref guard (#756) Two cold-gate guards, both wiring into test:offline: - test:doc-counts — re-instates #752, which was silently reverted when a later PR with a pre-#752 package.json was squash-merged (--admin bypasses the up-to-date check), dropping the script AND the wiring. The durable guard (check-doc-counts.sh) survived; only the cold-gate enforcement was lost. Restored the test file + wiring. - test:safety-refs — NEW: asserts every test named in move/SAFETY.md exists as a Move fun, so a renamed/removed test doesn't leave a judge's 'sui move test <name>' finding 0 tests. 155 refs verified present. Both mirror test:doc-links. Verified: both pass; cold-gate-coverage accepts both; JSON valid. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…gged) (#761) The doc-count reality guard (#738) caught the drift: a5's #758 (post-rug-open drain fix) added a Move test (count → 696) but the docs stayed at 695, so the cold-gate-enforced guard (#752) was failing. Ran 'npm run sync:count' to rewrite all four docs (README badge+body, move/SAFETY.md, deployments/ADDRESSES.md, the runbook) to 696. Verified: guard green — consistent across all docs AND matches the 696 #[test] functions in move/. This is the machinery working as designed (catch + one- command fix), unblocking the cold gate. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…gged) (#769) The doc-count reality guard (#738) caught the drift: a5's #767 (gate enable_rug on the vault admin cap — closes the rug bootstrap front-run) added a Move test (count → 697), but the docs stayed at 696, so the cold-gate-enforced guard (#752) was failing — and CI goes red on it (as it did for the last count drift, fixed by my #761). Ran 'npm run sync:count' to rewrite all four docs to 697. Verified: guard green — consistent across all docs AND matches the 697 #[test] functions. Unblocks the cold gate and the CI badge. Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…ffline (#773) verify:offline is the judge-facing zero-network proof command (DEMO.md). My #765 made it delegate to test:offline so it can't drift stale (it previously carried a hand-maintained list that silently dropped the /pro fairness proofs). But that fix is a single package.json line — exactly the kind a stale-base --admin merge silently reverts (cf. the #752 revert, where my doc-count wiring vanished the same way). The existing leaf-coverage assertion can't catch a verify:offline revert (it's not a test:* leaf), so this adds a direct guard: verify:offline must include 'npm run test:offline'. A revert to a parallel hand-list now fails the cold gate. Verified: both cold-gate-coverage tests pass (2/2). Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
scripts/check-doc-counts.sh was hardened to 9/9 form coverage (#738/#749/#750/#751) but only ran when someone remembered to. This wires it into test:offline via a thin test that runs the guard and asserts it passes — so a Move test added without
npm run sync:countfailsnpm test(with that one-command fix surfaced in the failure message), instead of a judge runningsui move testand seeing a doc mismatch.Mirrors test:doc-links (the doc-link consistency check already in the gate). The fleet's count flurry has settled at 695, so no transient mid-sync failures.
Verified: test passes on the consistent tree; test:cold-gate-coverage accepts the new wired test; package.json valid.
🤖 Generated with Claude Code