Skip to content

Activity Monitor: merge-train header says 3 PRs over one row; untrusted-author PR fronts the train #15552

Closed

Description

@clubanderson

Owner screenshot (22:34Z, hive's hive on v5 3ef38e6): the tracker group header reads 🚆 merge train · 3 PRs · 1st: #15425 — overlapping: tracker-row.js, index.html, but only one row sits under it — #15425 (Ready for human merge · untrusted author). The next row (#15499, Needs human decision) is a separate group. Owner: "this says there are 3 PRs in the merge train but only 1 is shown".

From /api/stage-tracker, train v5#15425 has three members: #15425 (position 1, front, stage 090, human_merge_reason: untrusted-author), #15529 (position 2, linked to #15493) and #15530 (position 3, linked to #15523). The train's head_state is needs-human ("waiting for a maintainer (untrusted author)").

Two bugs:

  1. Header count != rows. 🐛 fix(dashboard): omit empty lineage separator #15529 and 🐛 fix(dashboard): issue-row linked PR passes null when the lineage chip is empty #15530 are drawn as sub-rows of their issues Review backlog (correctness): Issue-row caller still passes '' (not null) when the lineage chip renders empty, leaving a pre-existing trailing space #15493 / Review backlog (security): Issue-row caller still passes '' instead of null when lineage chip is empty (carried over) #15523 (nested_under). The SPA's train grouping (stageTrackerTrainKey / stageTrackerTrainContiguous / stageTrackerTrainRows) only sees items with their own merge_train, so the issue rows are not part of the run, while the header text uses the server's size (3).
  2. A PR that cannot be auto-merged fronts the train. 🐛 fix: identify tracker claims from the local ledger #15425 was rejected by the contributor lane (contributor-lane:no-linked-issue, an untrusted-author verdict) but mergeTrainBlockingReason only covers not-mergeable / fork / red CI / terminal holds, so the planner kept it as an unblocked member and, being the oldest clean PR, it fronted: 🐛 fix(dashboard): omit empty lineage separator #15529 and 🐛 fix(dashboard): issue-row linked PR passes null when the lineage chip is empty #15530 queued behind a merge that was never going to happen. 🐛 tracker: merge queue ordinals, one column, readable train icon, no conflicting riders #15449 ("head that can't merge never fronts") and ✨ auto-merge: breaking-titled or needs-owner-merge PRs wait for a human #15485 (human-merge PRs excluded) did not cover lane admission/trust verdicts.

Expected: the header counts the rows actually drawn under it (issue rows whose nested PR is a member join the run; hidden members are named), and a PR that requires a person's merge is never a train member or head.

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

    help wantedDenotes an issue that needs help from a contributor. Must meet "help wanted" guidelines.

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions