[hivecommons-hive]

Test Improvement

Adds __tests__/cronJobTest/sweepDelayFailures.test.ts (new file; no existing test is edited) for the catch of the delayed-reviewers loop in sweep (src/cronJobs/sweep.ts, line 85 at main@e3119ad): result.failures.push({ number, message: \blunderbuss: ${e instanceof Error ? e.message : String(e)}` })`.

Two due delayed-only pull requests (created_at/updated_at 70 min ago, blunderbuss.delay: 1h) on a base branch with an OWNERS file, where one of them fails:

case	fault	asserts
records a failed review read as that pull request's failure and still serves the other	GET /pulls/1/reviews → 500 { message: 'reviews unavailable' }	sweep rejects with sweep: 1 pull request(s) failed: #1 (blunderbuss: reviews unavailable); POST /pulls/2/requested_reviewers was still made and nothing was posted for #1
stringifies a non-Error rejection while requesting reviewers	requestOwnersReviewers rejects once with the string 'secondary rate limit' (vi.mock('../../src/plugins/blunderbuss', { spy: true }), the technique of __tests__/utils/configSweep.test.ts)	the failure reads blunderbuss: secondary rate limit; the other pull request was still requested; requestOwnersReviewers was called twice

The first case also pins that the loop continues past a failing pull request instead of aborting (forEachLimited + per-item try/catch).

Files/functions claimed: __tests__/cronJobTest/sweepDelayFailures.test.ts only; source under test is the delayed-only forEachLimited body of sweep. Disjoint from #553 (auth.ts), #555/#559 (__tests__/bundle/blunderbussDelay.test.ts), #556 (e2e floor), #557 (tide.test.ts), #558 (__tests__/cronJobTest/sweep.test.ts, which states line 85 is out of its scope), #561 (line 80, new sweepDelayWithoutOwners files), #562 (line 208, new sweepDelayBareListing files), #563 (config.ts), #564 (owners.ts).

Coverage evidence

• Unit, npm run test:coverage on main@e3119ad (2330 tests, 99.91 % lines / 99.35 % branches): coverage/lcov.info BRDA:85,5,0,0 and BRDA:85,5,1,0 — neither arm reached; sweep.ts 98.33 / 92.3, uncovered 85,244. With this PR (2332 tests): both BRDA:85,5,*,1; sweep.ts 99.16 / 94.87, uncovered 244 (that is test(sweep): pin the inclusive due-window bounds and the timeline-read failure of delayed reviewers #558's timeline-read case); all files 99.35 → 99.46 % branches.
• End-to-end, npm run test:coverage:e2e on main@e3119ad (528 tests, 100 % / 99.46 %): coverage-e2e/lcov.info BRDA:85,5,0,1 / BRDA:85,5,1,0 — the Error arm is hit once by the bundle suite, the String(e) arm never. This PR adds no bundle scenario: octokit only ever rejects with RequestError, so the String(e) arm is bundle-unreachable; whether to v8 ignore it for the e2e run as chore(coverage): v8-ignore lgtm.ts's unreachable String(e) arm after bindLgtm (bindLgtm only ever throws Error) #537/chore(coverage): v8-ignore the bundle-unreachable fallback arms in run.ts:40, command.ts:23 and config.ts:657 #544/chore(coverage): v8-ignore requireMatchingLabel's bundle-unreachable non-Error arm at the rule loop's catch #549 did is a production-file change outside this test PR.
• The two runs are read separately on purpose (different transforms, see vitest.e2e-coverage.config.mjs).

Related Issue

Refs #560 — lands the folded sweep.ts:85 variant the hive attached to that tracker (comment <!-- hive-finding-folded --> of 2026-10-22); the tracker's original sweep.ts:80 finding is delivered by #561 and the folded :208 variant by #562, so #560 stays open until those merge.

Tests

• npx vitest run __tests__/cronJobTest/sweepDelayFailures.test.ts — 2 passed, no ::warning:: output
• npx eslint __tests__/cronJobTest/sweepDelayFailures.test.ts — clean
• npm run test:coverage — 2332 passed, 75 skipped; thresholds met

---

Filed by quality agent (hold-gated mode). Human review required.

— hive: agent=quality backend=copilot model=claude-fable-5.1

[hivecommons-hive]

Important

Held for human review by the hive's ACMM level gate.

This PR was opened by the "quality" agent while Hive policy required a human checkpoint for that agent. Agent PRs are held at ACMM L3–L5; an outreach PR that targets a repository this hive does not manage is held at every level because it publishes project-facing communication into someone else's project.

Hive will keep the hold label until a human removes it. Operators can make a deliberate one-off release during an ACMM level change with release_level_holds=true, but level changes never release this hold automatically.

[hivecommons-hive]

/kind cleanup

---

🐝 Hive Agent: quality | Instance: hosted-available-lke648397-260827-5q9t | SHA: e3119ad

— hive: agent=quality backend=copilot model=claude-fable-5.1 copilot=1.0.88