test(bundle): drive tide's merge-queue skip and enqueue-refusal arms through dist/index.js

[hivecommons-hive]

Test Improvement

Files: __tests__/bundle/mergeQueueArms.test.ts (new, bundle e2e; imports configReads, helpersFor, ownersProbe, queueRead, repo, token from ./helpers). Cluster: tide's merge-queue path — src/plugins/tide.ts evaluateInQueue skip and enqueue-refusal arms, src/utils/mergeQueue.ts classify/errorMessage and the pullRequest: null state-query arm — driven through dist/index.js.

Ten pull_request labeled lgtm runs on a queue-enabled base branch against fakeGithub:

1. pr already in the queue with an entry → skipping pr #1: in the merge queue (position 2, AWAITING_CHECKS), no enqueue mutation, no PUT
2. pr in the queue without an entry → skipping pr #1: in the merge queue
3. mergeable_state: dirty → skipping pr #1: not mergeable (dirty), no enqueue
4. first read unknown, re-read after the 1 s backoff at a different head → skipping pr #1: head moved during evaluation, no enqueue
   5–7. enqueueError messages classified head_moved / already_queued / not_ready → each skipped, exit 0
5. an unclassified GraphQL refusal → ::error::could not enqueue pr #1: …, exit 1, no PUT
6. HTTP 500 on the mutation (no GraphQL errors array) → the transport's Server Error is reported, exit 1
7. state query answering repository.pullRequest: null → falls back to the direct PUT …/merge without the "falling back" warning

Each case asserts the exact request sequence (bind, pullRead, ownersProbe, bindingRead, graphql[, graphql | pullRead | merge]).

Evidence (main @ 3fc21f2, Node v26.10.0, vitest 5.0.3, npm run test:coverage:e2e): src/plugins/tide.ts e2e lines 81.02 → 86.15 % (303–304, 308, 313, 316, 335, 337, 339, 346–347 now hit); src/utils/mergeQueue.ts 84.09 → 97.72 % (only line 67, the test-only resetMergeQueueWarnings, remains). Bundle suite 6 files / 152 tests green; eslint and tsc --noEmit clean. Unit coverage unchanged (no src/ change).

Disjoint from the open hold-gated PRs: #289 (mergeGateLabels.test.ts, the missing_labels gate and labelMatch.ts), #293 (holdSweepSettings.test.ts), #310 (pullReqJobsArms.test.ts), #318 (lgtmBindingArms.test.ts) and the rest — none drives evaluateInQueue's skip arms or an enqueue refusal other than forbidden, which bundle.test.ts already covers.

Related Issue

Closes #321

---

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

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

[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. Non-outreach agents are held at ACMM L3–L5; the outreach agent is always held because it publishes project-facing communication.

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: 443e72f

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

[hivecommons-hive]

build-test is red here for a cause this PR's diff (__tests__/bundle/mergeQueueArms.test.ts only) does not touch: all 81 test files pass, and the job exits 1 on the single vitest Unhandled Error assert(err.code !== "ERR_TLS_CERT_ALTNAME_INVALID") originating in __tests__/issueCommentTest/meow.test.ts (runs 37278444109, 37278399048). This is the shared incident tracked in #294; the fix is #287 (held, awaiting human review). Deferring this PR until that lands — no retries or pushes here for this failure.

---

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

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