You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
scripts/root-index.test.ts has ejected 2 pull requests from the merge
queue within a rolling 24 hours — 1 independent hit 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.
⚠️Queue depth is not evidence. The 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 root 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.
Closed completed: a semantic conflict between two queued PRs, not a flake; both have landed, and main agrees with itself · triage seat · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-16T04:54Z
Read in this act: the two victims' triage comments, PRs #23783 and #23758, and content/docs/references/index.mdx on origin/main.
The REASON line. It was an assertion, not a timeout: scripts/root-index.test.ts > … states no count that its own rows contradict, with ['ui', 16, 168] against ['ui', 16, 169].
On main the UI row reads 16 | 169 and the header agrees.
Queue CI since 03:10Z reads 8 success and 1 running, with no red.
The structural note for whoever owns the generator (domain:spec, packages/spec/scripts): an aggregate count in a generated header conflicts whenever two PRs add schemas to the same module in parallel. One independent hit in 24 h is no measured cost. If the anchor refreshes again, triage files the hardening card with that count.
Activity
objectstack-fleet commentedon Oct 16, 2026
Closed
completed: a semantic conflict between two queued PRs, not a flake; both have landed, andmainagrees with itself · triage seat ·session_01AavokzJ5DndAwitDXvKy4U· 2026-10-16T04:54ZRead in this act: the two victims' triage comments, PRs #23783 and #23758, and
content/docs/references/index.mdxonorigin/main.scripts/root-index.test.ts > … states no count that its own rows contradict, with['ui', 16, 168]against['ui', 16, 169].record:json, a new UI schema) regeneratedindex.mdxon a base from before another schema landed. The queue stacked it, and the tree's ui header said 168 while its rows held 169.inheritedbystander: its diff touches onlyskills/objectstack-ui/rules/pages.md.mainthe UI row reads16 | 169and the header agrees.domain:spec,packages/spec/scripts): an aggregate count in a generated header conflicts whenever two PRs add schemas to the same module in parallel. One independent hit in 24 h is no measured cost. If the anchor refreshes again, triage files the hardening card with that count.⛔ Not a claim.