Repository navigation
π tracker: merge queue ordinals, one column, readable train icon, no conflicting riders - #15449
Conversation
β¦ing riders - Train position reads 'Merge queued Β· 1st/2nd/β¦' everywhere; the slugs 'front' and 'queued-behind' never reach a chevron, hover, header chip or decision sentence (tests enumerate every train state). - Every merge-train state β the 1st rider included β anchors in the Merge-queue column, never CI/Review. - Column copy 'Merge gate' β 'Merge queue' (ids/classes/JSON unchanged). - Header chip and column header share a readable inline-SVG train icon (14px, currentColor, title 'Merge queue'). - Conflicting (DIRTY) and returned-to-contributor PRs are excluded from train membership and size, so they get no stall comments. Signed-off-by: Andrew Anderson <andy@clubanderson.com>
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of
|
|
@clubanderson is auto-promoted for contributor-lane merging on this PR: the last 5 merged PRs by this author in this hive's repositories have no revert or sentinel-alert evidence. CI, reviewer verdict, diff fences, soak, holds and branch protection still apply. |
|
@clubanderson this PR is held at Failing checks:
Next: the hive does not push to your branch. Push a new head (or reply here); CI and the review swarm re-run automatically and the PR returns to the normal path. |
|
stage 080-awaiting-merge-admission β 071-needs-author: optional-check-failing:relative links and anchors resolve |
Signed-off-by: Andrew Anderson <andy@clubanderson.com>
There was a problem hiding this comment.
Read the Go changes and the tracker JS against v5 at 2d97036; the behaviour change looks correct to me. One sequencing note for a maintainer at the end.
Verified in the tree:
planMergeTrains(src/pkg/github/automerge/automerge_merge_train.go:366-369, 431-437): returned-to-contributor PRs are dropped beforeperBaseis built, and conflict outsiders are zeroed (m.Number = 0) after the live-head refresh.groupMergeTrains(automerge_merge_train.go:196) already skipsm.Number <= 0, so an excluded member neither takes a slot nor counts towardsize.mergeTrainConflictOutsider(automerge_merge_train.go:477-487) keeps a dirty member only whenmergeTrainLastFront> 0, a resolver is wired, and0 < attempts < MaxRebaseAttemptsβ which is exactly the "the train is repairing it" exception the body describes.mergeTrainSkipLoggedgoes throughmarkTerminalHoldLogged(automerge_held_runs.go:672), keyed on head SHA + reason, so the INFO line is once per head as claimed.mergeTrainReturnedToContributorreads the label viastagelabels.ReturnedToContributor.Label()and falls back tostore.Get(stagelabels.Key(...))β both exist onv5(stagelabels.go:64, 273).stageTrackerAnchor(tracker-row.js:17-21) now anchors anymerge_train.statepast the PR column in the Merge-queue column; the glossary Go/JS copy swap is wording only, slugs and aliases untouched.go build ./pkg/github/automerge/ ./pkg/stageengine/prstage/ ./pkg/delegatedecision/ ./pkg/dashboard/passes on the head. Go tests are CI's (not runnable from this pod).
Overlap with #15400 (open, @Karibhou, Refs #15395). It inserts an early return in the merge-gate word function (tracker-words.js, after the laneReason line, in the queued-behind && src.merge_train case) that renders 'In merge queue Β· ' + position + ' of ' + size. This PR does not touch that function (its nearest hunk is at tracker-words.js:712), so the two merge cleanly β but if #15400 lands as well, that early return fires before stageTrackerTrainWord and the chevron would read "In merge queue Β· 2 of 6" instead of this PR's "Merge queued Β· 2nd", undoing the single-vocabulary goal stated here. Not a defect in this diff; flagging so whoever merges decides whether #15400 is superseded by this one or needs rewording on top of it. I have not touched #15400.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
v5 already ships the queue-position chip (hivecommons#15427, hivecommons#15449): 'Merge queued Β· 2nd', sweep-stable label, spinning cue and a tooltip naming the train and PRs ahead. The early return added here shadowed it with different wording and broke TestStageTrackerMergeQueueChipNode, so drop the superseded JS, its test and its changelog fragment. Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com>
Follow-up to #15427 (merged) with the owner's round of feedback on the live train (hive #15400 / #15425 rows). Same concern β the merge-train chevron β one PR.
Fixes #15395 (follow-up round).
What changed
Ordinal wording, no slugs. The train position now reads
Merge queued Β· 1st/2nd/3rdβ¦11th21ston the chevron, with hover "1st of N in the merge queue Β· merges next Β· ETA ~Xm" / "2nd of N in the merge queue Β· behind #15400 Β· ETA ~Xm". The header chip ismerge train Β· 2 PRs Β· 1st: #15400. The stringsfrontandqueued-behindno longer appear anywhere user-facing β chevron, hover, caption, header chip, popover, stage comment sentence, decision copy. A test enumerates everymerge_train.stateand fails on either slug (and on ordinals rendering wrong for 1/2/3/11/12/13/21/22/23). "merge sweep running" is folded into the hover only; the chevron stays a single label (#15416).1st-in-train row lit the CI/Review column.
stageTrackerAnchoronly pulledmerging/merge-failedinto the merge column; a train head whose stage cursor sat at 50/60/70 (CI still re-running after rebase) anchored in the review column. Now any item withmerge_train.statepast the PR column anchors in the Merge-queue column; a test asserts every train state at stages 50β95 resolves there."Merge gate" β "Merge queue" in all user-facing copy: column header, hover/caption, legend heading, tracker-words, deferred note, decision glossary (Go + JS mirror, byte-identical), docs. Internal ids/classes/JSON keys (
merge_gate,.ham-gate-*,merge-gate) unchanged.Readable train icon. Header chip and Merge-queue column header share an inline Octicon-style SVG (14px,
currentColor, side-view railcar with two wheels,<title>Merge queue</title>,aria-label).STAGE_TRACKER_ICONS[80]carries the same path; a test asserts the two literals are equal.Conflicting / returned PRs never ride. hive #15307 (mergeable=CONFLICTING, returned to contributor at 072) got a
stage 080 β¦ merge-train:queued-behind#15424:4/6comment. The planner now excludes DIRTY/CONFLICTING PRs and anyone carrying the returned-to-contributor label (or stage-store observation) from train membership and from the train size N; the only exception is a dirty member the train is itself repairing (resolver wired, attempt in progress). Excluded PRs get a once-per-headmerge train: PR left out of every train β¦ why=conflicting|returned-to-contributorlog line and no stall comments. Tests:TestPlanMergeTrainsExcludesReturnedAndConflicting, updated front/sibling/rebase-guard tables.Screenshot
Docs / changelog
src/docs/dashboard-sections.md,dashboard-copy-inventory.md(new "Merge queue" section),operator-reference.md,api-reference/dashboard-misc.md,labels-and-control-signals.md;changelog.d/added-merge-queue-position.mdupdated,changelog.d/fixed-merge-train-conflicting-excluded.mdadded.