Repository navigation
🐛 Activity monitor: an adoption pair never asks a human to pick a PR - #15550
Conversation
|
[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
|
There was a problem hiding this comment.
Read the tree at 4be33a6. The adoption-pair collapse looks correct to me; one open question at the end.
Server and SPA agree on what "superseded" means. stageTrackerLinkedPRSupersededByListed (stage_tracker.go:3261-3269) and stageTrackerLinkedPROpen(lp, issue) (tracker-row.js) both exclude the original only when its copy is listed open and adopted_by_state is none of closed / closed-twice / copy-failed — the Go constants at stage_tracker.go:1892-1896 are those exact strings, and lineageAdopted/lineageSuperseded (stage_tracker_lineage.go:26-27) are what the Go lineage branch keys on. A copy that was closed without merging therefore lets the original speak again on both sides.
The lp.State == "closed" early return is not a wider change than the body says. StageTrackerLinkedPR.State is documented and populated only as "closed" for an adopted original closed once its copy existed (#15126; struct doc at stage_tracker.go:~640), so skipping it from the progress caption matches the existing "dimmed line, no chevron" treatment rather than hiding arbitrary closed PRs.
adoptRefsNotePattern fills a real gap. The existing adoptClosesPattern (automerge_contributor_adopt.go:62) anchors #N\s*$, so Refs #15395 — partial never matched and the copy lost its issue link — which is exactly what #15499 shows ("no linked issue").
Open question (not a defect): the new pattern only carries a Refs #N whose annotation starts with a dash ([—–-]). The policy's other documented form, Refs #N (needs-human: <reason>), still won't travel. If that form is in use, it may want the same treatment.
— hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
@clubanderson this PR is held at Why:
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 → 071-needs-author: conflicting |
An issue whose only PRs are a contributor's original and the hive's adoption copy read "Let hive decide — all 2 PRs" and "fell back to Covered by pr — Covered by PR #N — …; PR #M — … changes requested". The write-back gate had kept the original open at 072, the SPA counted it as a competing open PR (the server never sends a superseded flag), the server caption listed both PRs, and the feed read the issue's un-mirroring (50 → 40) as a fall. - stageTrackerLinkedPRProgress skips the superseded original and words the copy as "Covered by #N (adopted from @author's #M) — <stage phrase>". - stageTrackerLinkedPROpen(lp, issue) treats an original whose open copy is listed under the issue as superseded; nested open / parked / sibling counts and the delegate-all fan-out follow. - stageTrackerEventText prefers the stage table's short word and reads an issue leaving a mirrored PR stage for its own 040 as "moved to". - adoptionCloses carries an annotated `Refs #N — partial` as `Refs #N` so the copy keeps its original's issue link (#15499 read "no linked issue"). Signed-off-by: Andrew Anderson <andy@clubanderson.com>
4be33a6 to
0a02ddf
Compare
|
hive/awaiting-ci: rebased onto v5 (resolved a docs table conflict in dashboard-copy-inventory.md by keeping both rows; authorship and sign-off intact). CI pending — sweep will merge when green. 🐝 Hive Agent: — hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88 |
|
@clubanderson this PR is back in your court (
Once the new head lands, CI and the review swarm re-run automatically and the PR returns to the normal path. |
|
stage 080-awaiting-merge-admission → 071-needs-author: hive-signoff-attribution |
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
Mechanical split of the 3298-line merge hotspot into four files in the same package: stage_tracker.go (types, stage table, authoring/tier helpers), stage_tracker_caption.go (modifiers, reason/caption text), stage_tracker_build.go (buildStageTracker, ordering, cache, handler) and stage_tracker_links.go (linked issue/PR resolution). No behaviour change; declarations moved verbatim and imports re-derived with goimports. Rebased onto v5: the ledger-claim inputs (#15425) and the adoption-pair linked-PR progress helpers (#15550) that landed on stage_tracker.go since the branch point now live in stage_tracker_build.go and stage_tracker_links.go respectively. A routine refactor takes the no-changelog label rather than a fragment (changelog.d/README.md). Hive-Run: #15511 Hive-Plan: modularize-stage-tracker Signed-off-by: SACHIN NAIK <sachin.naik@ibm.com> Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com>
* ♻️ dashboard: split stage_tracker.go into 4 files Mechanical split of the 3298-line merge hotspot into four files in the same package: stage_tracker.go (types, stage table, authoring/tier helpers), stage_tracker_caption.go (modifiers, reason/caption text), stage_tracker_build.go (buildStageTracker, ordering, cache, handler) and stage_tracker_links.go (linked issue/PR resolution). No behaviour change; declarations moved verbatim and imports re-derived with goimports. Rebased onto v5: the ledger-claim inputs (#15425) and the adoption-pair linked-PR progress helpers (#15550) that landed on stage_tracker.go since the branch point now live in stage_tracker_build.go and stage_tracker_links.go respectively. A routine refactor takes the no-changelog label rather than a fragment (changelog.d/README.md). Hive-Run: #15511 Hive-Plan: modularize-stage-tracker Signed-off-by: SACHIN NAIK <sachin.naik@ibm.com> Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com> * changelog: fragment for #15574 Generated by changelog-fragment-guard from the PR title (#5675). Edit the wording freely, or replace this file with the no-changelog label if the change is not user-visible. Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Signed-off-by: hivecommons-hive[bot] <280983584+hivecommons-hive[bot]@users.noreply.github.com> * docs: point ADR-0025 stageStoreOrShared citation at stage_tracker_caption.go Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com> --------- Signed-off-by: SACHIN NAIK <sachin.naik@ibm.com> Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com> Signed-off-by: hivecommons-hive[bot] <280983584+hivecommons-hive[bot]@users.noreply.github.com> Co-authored-by: SACHIN NAIK <sachin.naik@ibm.com> Co-authored-by: hivecommons-hive[bot] <280983584+hivecommons-hive[bot]@users.noreply.github.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Fixes #15549
Why
#15495 / #15500 each listed a contributor PR and the hive's adoption copy of it. The write-back gate (
needs-opt-in-label) kept the originals open at 072, so the tracker counted two open PRs, offered "Let hive decide — all 2 PRs", and captioned the row "fell back to Covered by pr — Covered by PR #15537 — …; PR #15508 — … changes requested". Nothing there needed a human — hive had already adopted both.What
stageTrackerLinkedPRProgress: skip the superseded original of an adoption pair (open copy listed, or original closed); word the copy asCovered by #N (adopted from @author's #M) — in CI / review. Unrelated multi-PR issues keep thePR #N — caption; PR #M — captionlist.stageTrackerLinkedPROpen(lp, issue): an original whose open copy is listed under the issue (andadopted_by_statenot closed / closed-twice / copy-failed) is not open; nested-open, parked, sibling and delegate-all counts follow. Two genuinely competing PRs still get the fan-out.stageTrackerEventText: stage table'sshortword first (no more "Covered by pr"); an issue leaving a mirrored PR stage for its own 040 reads "moved to", not "fell back to".adoptionCloses: an annotatedRefs #N — partialtravels asRefs #N, so the copy keeps its original's issue link (🐛 fix: keep merge queue position visible during sweeps #15499 read "no linked issue").TestStageTrackerLinkedPRProgressCollapsesAdoptionPair,TestStageTrackerAdoptionPairIsOneOpenPR,TestStageTrackerEventTextUnmirroredCoveredByPR, extendedTestAdoptionClosesCarriesRefs. Docs:hive-merge.mdadoption section,dashboard-copy-inventory.md. Changelog fragment.Local:
go build ./... && go test ./pkg/dashboard/... ./pkg/github/...green;go generate ./pkg/dashboard/...output committed.