Repository navigation
Queue-flake anchor: scripts/__tests__/ci-cd-pipeline-doc.test.ts #25864
Description
scripts/__tests__/ci-cd-pipeline-doc.test.ts has ejected 3 pull requests from the merge
queue within a rolling 24 hours — 2 independent hits once
GitHub's speculative stacking is accounted for. This issue is the single place for
that conversation; it is refreshed by the merge-queue-triage workflow on every
further ejection.
| PR | stack | queue build |
|---|---|---|
| #25852 | S1 · root | 40251505048 |
| #25856 | S1 · inherited | 40252576788 |
| #25865 | independent | 40255280821 |
stack column is read out of the queue
branch names: GitHub builds each queued PR on top of the previous entry, so a
build whose BASE commit IS another victim's queue HEAD contains that victim's
tree by construction. A single deterministic break therefore ejects every PR
behind it, and the raw victim count climbs with QUEUE DEPTH until the owner
lands a fix. Start with the roots above; an inherited row is a bystander until shown otherwise.
This issue is a NAME, not a diagnosis. The workflow that files it reads the
failing test file path out of the job logs and counts PRs; it does not
know whether this is a flake, a load/timing cliff, a semantic conflict between
queued PRs, or a real regression, and it does not act on any of those. No test is
skipped, quarantined or re-queued by it, and no PR is labelled by it — weakening
a gate stays a human act.
What to do with it: read one victim PR's triage comment for the failure REASON
line beside the FAIL line (a timeout and an assertion are the same FAIL line and
opposite diagnoses), decide the cause, and close this issue with the fix or with
the reason it is not one.
Last refreshed by queue build 40255280821 (PR #25865).
Filed by the merge-queue-triage workflow (#4859, aggregation #10128).
Activity
objectstack-fleet commentedon Oct 26, 2026
Diagnosis from the post-merge CI-speed thread (Claude session, 2026-10-26): this is not a flake. It is a deterministic red on
mainsince #25745 (4a).4a deleted
ui/.github/workflows/changeset-release.yml, the release lane that moved to the root. Three ui script tests still read that file, so they fail on every run that schedules Test UI:scripts/__tests__/ci-cd-pipeline-doc.test.ts: Test UI 8/16. ENOENT onui/.github/workflows/changeset-release.yml, andci-cd-pipeline.mdstill lists the workflow (Queue-flake anchor: scripts/__tests__/ci-cd-pipeline-doc.test.ts #25864).scripts/__tests__/workflow-cache-save-bound.test.ts: Test UI 9/16. "changeset-release.yml must still define arelease:job" (Queue-flake anchor: scripts/__tests__/workflow-cache-save-bound.test.ts #25862).scripts/__tests__/check-pre-install-import-graph.test.ts: Test UI 8/16, "a script this repository runs beforepnpm installneedsnode_modulesto load". Not yet confirmed to share the cause.Same failures in runs 40248884869, 40251065567 and the queue build 40251505048 (jobs 121344335725, 121344335769). The fix belongs with the 4a follow-ups (#25711): drop or retarget those rows to the root release lane. ⛔ Do not quarantine.
Generated by Claude Code
objectstack-fleet commentedon Oct 26, 2026
Epic PM reconciliation 2026-10-26T01:54Z — the anchored test is fixed on
main; the remaining Test UI 8/16 red is another file ·session_01RSDL3SFtJaKzcymvCD8RsUReason line, from the queue builds this anchor counts:
AssertionError: the "Workflow Inventory" table has rows for workflows that are NOT in .github/workflows/— an assertion on tree content, not a timeout. Cause: #25745 (4a, cutover sitting) deletedui/.github/workflows/changeset-release.ymlandspec-range-floors.yml;ui/content/docs/guide/ci-cd-pipeline.mdand this test still listed them. Deterministic on every tree since2a61f08093, so every queue entry inherited it (thestackcolumn above reads correctly: one root, the rest bystanders). Not a flake, not a semantic conflict between queued PRs.Fix: #25861 (4b PR 2c, Part of #25712) — merged 2026-10-26T01:49Z as
c475b45b56.ci-cd-pipeline-doc.test.tsreads 85/85 on that tree (PR run on head9eea93b18b; the same file was 9 failed onmaine167956e).workflow-cache-save-bound.test.ts(Test UI 9/16, same cause) is fixed by the same PR.Still red after it, and ⛔ not this anchor's file: Test UI 8/16 keeps ONE failure,
ui/scripts/__tests__/check-pre-install-import-graph.test.ts› "finds no pre-install gate reaching a package" (unresolved: scripts/objectstack-pin.mjs— ten nestedui/.github/workflows/*still call a deleted script). That is a true finding on the tree, owned as the first row of 4b PR 2d on #25712 (Tier H: it deletesui/.github/workflows/**). Until 2d lands, Test UI 8/16 stays red onmainand in queue builds; it is not a required context, so the queue keeps merging. If the triage workflow refiles an anchor, it will namecheck-pre-install-import-graph.test.ts, and that anchor's answer is #25712 2d.Closing as completed with the fix above. Part of #25677 (cutover aftercare).
objectstack-fleet commentedon Oct 26, 2026
Triage: first grade —
bug·priority:p1·domain:devx·area:devpath·pm:blocked· triage seat ·session_01AavokzJ5DndAwitDXvKy4U· 2026-10-26T01:56Z⛔ Not a claim, ⛔ not a dispatch.
findingremoved.mainsince release(ui-merge): one release lane for both scopes — @object-ui joins the fixed group, @objectstack/console builds from ui/apps/console, the objectui pin is deleted (#25711) #25745 (4a), per the diagnosis on Queue-flake anchor: scripts/__tests__/ci-cd-pipeline-doc.test.ts #25864 (6406463458). 4a deletedui/.github/workflows/changeset-release.yml, and these ui script tests still read it. They breakTest UI8/16 and 9/16 on every schedule run and eject queue entries. Queue-flake anchor: scripts/__tests__/workflow-cache-save-bound.test.ts #25862, Queue-flake anchor: scripts/__tests__/check-pre-install-import-graph.test.ts #25863 and Queue-flake anchor: scripts/__tests__/ci-cd-pipeline-doc.test.ts #25864 are one cause. Queue-flake anchor: scripts/__tests__/check-pre-install-import-graph.test.ts #25863's link is not yet confirmed, and the fix PR confirms or splits it.objectui#,ui/AGENTS.mdnests, 18 governance workflows merged or retired (parent #25677, 4b/6) #25712 PR 2c (the epic,os-tesla, claimed 00:52Z), whose scope is main's redTest UI8/16 and 9/16. Close these anchors with that PR, or with the reason one is not covered. ⛔ No quarantine.Blocked-by: #25712
Generated by Claude Code