Skip to content

[finding] Console Pin Gate first runs in the merge queue for a spec change that moves what the console bundles, and it is not required, so a red there merges (PR #20695: main red for about 2 h) #20765

Description

@objectstack-fleet

Filed by the domain:spec seat 2 PM (session_014EJ1ED8X4MMrT18BhVx4tx) at sign-off, as a mechanizable item from the shift (the handover-report class "可机械化项"). ⛔ Filed bare: routing and grading are triage's. ⛔ Not a claim.

What happened (measured)

The gap

A check that a PR head skips by path filter, and that is not required, is judged for the first time in the queue, and its verdict there cannot stop the merge. The seat's pre-enqueue rule (every check green or a roster skip) passes such a PR by construction.

Shapes a fix could take (measure first; none is a ruling)

  • Widen the console filter to the spec paths that decide what the console bundles, for example packages/spec/package.json (the exports map), packages/spec/src/index.ts and packages/spec/tsup.config.ts. Then a head that moves the entry layout runs the gate before it enqueues. This is an existing gate's parameter, not a new gate. The job's cost on such heads is one console build.
  • Or make Console Pin Gate a required context. That is a repository-settings change, the maintainer's.
  • Or have the seat protocol watch a queued PR's non-required queue checks and dequeue on red. The platform reading on the dequeue channel is filed separately (see the draft-conversion finding filed alongside this one).

Generated by Claude Code

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions