test(sweep): call the exported report builder, not a name that never landed

[kkroo]

Thinking Path

I was investigating why my own PR #1201 had fallen from merge-queue position 1 to position 10, expecting a flake. It wasn't one.

The failing shard reported ReferenceError: buildHumanGatedAgeingReport is not defined. That is a missing symbol, not a flaky assertion, so the first question was whose it is. #1201 is CI-only (workflows plus two test files) and touches nothing near this code, which ruled it out. A repo-wide code search for buildHumanGatedAgeingReport returned zero matches — the function exists nowhere.

Checking the file on master: human-gated-ageing.test.ts calls it at lines 611 and 622, the import block at lines 2–14 does not import it, and it is not defined locally. So both tests throw on every run.

Then the important question — is this only my PR, or everything? Every merge-queue branch built on d63bbc66 fails the same shard: pr-1201, pr-1189 and pr-1089, three unrelated PRs, all ejected. I pulled the log for pr-1089's queue run and found the identical ReferenceError four times. The queue has grown to 14 entries. Nothing can merge.

For the intended API, selectAgedHumanGatedIssues is exported at human-gated-ageing.ts:360 returning HumanGatedAgeingReport, is already imported by this test at line 12, and is already called with the identical shape at lines 65 and 71. HumanGatedAgeingReport.scanned is the field both broken tests read. So the two call sites simply carry a stale name from an earlier draft of the module.

I checked whether the missing half was in flight before writing anything. My own #1241 (feat(sweep): add human-gated ageing escalation) is a DIRTY draft that adds the same file set and exports the same symbols — it does not define buildHumanGatedAgeingReport either. Nothing open supplies it.

What Changed

Two call sites in server/src/__tests__/human-gated-ageing.test.ts renamed from buildHumanGatedAgeingReport to selectAgedHumanGatedIssues. No import change — it was already imported. No production code touched.

Risks

Low. Test-only, two identifiers, and the replacement is the function the rest of the file already uses for exactly this purpose with exactly these arguments.

The one thing worth a reviewer's eye: I assumed the two tests intended to build a report the same way lines 65 and 71 do. If instead a genuinely distinct buildHumanGatedAgeingReport was meant to land — a different builder with different semantics — then this makes the tests pass while quietly weakening what they assert. I looked for that function in the open PR set and in the module and found nothing, but I cannot prove intent from a name that never existed. Both tests are about humanGatedAgeHistogram rejecting bad bucket bounds, and they only use the report to obtain .scanned, so the builder choice is incidental to what they verify.

Verification

Ran locally in a clean worktree off origin/master (d63bbc669):

• With the fix: Test Files 1 passed (1), Tests 35 passed (35).
• Negative control — stashed the fix, same command at base: Tests 2 failed | 33 passed (35), both failures ReferenceError: buildHumanGatedAgeingReport is not defined. Restored after.

So the two failures are precisely the ones this fixes, and nothing else in the file changes behaviour.

CI evidence that this is repo-wide rather than local: queue-branch runs for pr-1201, pr-1189 and pr-1089 all on base d63bbc66, all failing General tests (server 1/4), with the identical error confirmed in pr-1089's log.

• I searched open and merged PRs for similar or duplicate work before filing this — nothing supplies the missing symbol; the nearest match, my own draft feat(sweep): add human-gated ageing escalation (BLO-19130) #1241, defines the same exports without it.

Model Used

Claude Opus 5 (1M context)

[cursor]

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.

[allyblockcast]

Ally — Consolidated PR Review

Lenses: pr-review-toolkit (code, tests, comments, errors, types) + gstack/review + native-codex.
Reviewed head: 9908ccf

Critical Issues (0)

Important Issues (0)

Suggestions (0)

Strengths

• selectAgedHumanGatedIssues is already imported and is used with the same arguments elsewhere in this test file.
• The replacement returns HumanGatedAgeingReport, so .scanned continues to supply the exact population asserted by both histogram-validation tests.
• The patch is intentionally scoped to the two nonexistent references; it restores execution of the existing assertions without changing production behavior or weakening the bucket-bound checks.

Recommended Action

1. Merge when the remaining repository checks complete.

[allyblockcast]

Ally — Consolidated PR Review

Lenses: pr-review-toolkit (code, tests, comments, errors, types) + gstack/review + native-codex.
Reviewed head: 9908ccf

Critical Issues (0)

Important Issues (0)

Suggestions (0)

Strengths

• selectAgedHumanGatedIssues is already imported and is used with the same arguments elsewhere in this test file.
• The replacement returns HumanGatedAgeingReport, so .scanned continues to supply the exact population asserted by both histogram-validation tests.
• The patch is intentionally scoped to the two nonexistent references; it restores execution of the existing assertions without changing production behavior or weakening the bucket-bound checks.

Recommended Action

1. Merge when the remaining repository checks complete.

[allyblockcast]

Hey @kkroo! Before this PR can be reviewed, a few things need attention:

Missing or incomplete:

• No test files detected in this PR — please include a test that verifies the bug fix or new behavior. If this PR genuinely doesn't need a test (e.g. a refactor), please retitle with refactor: prefix.

Once updated, push a new commit and these checks will re-run automatically.

— commitperclip