Repository navigation
Conversation
The engine only reads a contributor's repository permission once adoption is due (adoptContributorPR gates maintainerHoldsAdoption behind the due check), so a stage/072 row under maintainer_policy: never had no AdoptionBlocked record before that moment and nothing to tell it apart, in the dashboard's own data, from a row the hive will truly auto-adopt. The tracker read "hive adopts it in Nm" right up to the due moment for a maintainer's own PR, then flipped to "not adopted: maintainer_policy is never". stageTrackerAdoptionText now names the exception for the never policy's auto-adopt countdown instead of promising an outcome it cannot back for every author: "… unless the author maintains this repository". Hive-Run: hivecommons#15692 Hive-Plan: review-backlog-correctness Hive-Spec: issue-15692#stage_tracker_adoption_never_policy_text Signed-off-by: onkar717 <shelkeonka@gmail.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
|
There was a problem hiding this comment.
Read this against the tree at ea23fb2b, not just the diff — looks correct to me.
src/pkg/dashboard/stage_tracker_adoption.go:217-224: the two newnevercases sit afterAction/BlockedByand before the genericAutoDue/AutoAdoptAtarms, so a maintainer-approved head or an explicit blocker still wins, and the non-neverwording is unchanged.Policyis populated at all three constructor sites (:120,:134,:151) andconfig.AdoptionMaintainerPolicyNeveris the existing constant (config.go:3012).- The claim in the comment — that the engine only reads the author's permission once due — matches why the pill cannot promise adoption here; the row names the exception instead of asserting the outcome.
stage_tracker_adoption_test.gopins both captions (due, and 2m countdown) undernever, andsrc/docs/labels-and-control-signals.md:1934-1938carries the operator-facing wording.
One heads-up for whoever merges: #15704 also edits stage_tracker_adoption.go (3 lines, a different case) — not a duplicate, but whichever lands second will need a trivial rebase.
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Read this at head ea23fb2 against the engine and the tracker; it looks correct to me.
- Premise verified in the engine:
src/pkg/github/automerge/automerge_contributor_adopt.go:640-642returns before anything else when!due, andmaintainerHoldsAdoption/NoteAdoptionBlocked(…, AdoptBlockedAuthorMaintainer, …)only run at:661-664, after the due gate. So anever-policy row genuinely has noAdoptionBlockedrecord until due fires, and the pre-fix countdown text was indeed a promise the engine would then refuse for a maintainer's head. - Case ordering is right: the two new
case a.Policy == never && …arms instage_tracker_adoption.go:217-224come after everya.BlockedByarm, so the moment the engine records the block the row still readsnot adopted: maintainer_policy is never(the:211-212arm wins), and the hedge only shows in the window before that.adopt/askfall through to the unchanged:225-228arms. TestStageTrackerAdoptionNeverHedgesBeforeDuepins both the fired (AutoDue) and pending (AutoAdoptAt) texts with noAdoptionBlockedrecord present, which is exactly the window this fixes.- Docs and changelog fragment (
fixed-15692-…) are updated consistently with the new wording.
On the conflicts-with-open-pr label: the overlap is with #15704, which also edits stage_tracker_adoption.go, but at a different function (stageTrackerAdoption, hunk at :128) plus an import line; this PR touches only the switch in stageTrackerAdoptionText. Whichever lands second should rebase mechanically — no semantic conflict that I can see.
— hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Ready for a human merge: verdict approved @ea23fb2, 58/58 checks green, DCO yes, soak 30m0s elapsed. Blocked only by The hive never approves on your behalf: |
|
@onkar717 this PR is back in your court ( Review asks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
@clubanderson asked the hive to adopt this pull request now, without waiting for its author's hive. The hive opens its own PR carrying the author's commits and sign-off unchanged on the pass now running; this PR closes when that copy merges. |
|
Supersede failed: the hive could not open a fresh PR with these commits. Merging Files changed on both sides:
Rebase this branch onto |
Summary
Closes #15692.
adoptContributorPRonly reads a contributor's repository permission once adoption isdue(maintainerHoldsAdoptionis gated behind theduecheck inautomerge_contributor_adopt.go), so a stage/072 row undermaintainer_policy: neverhas noAdoptionBlockedrecord before that moment — nothing in the dashboard's own data tells it apart from a row the hive will truly auto-adopt.… — hive adopts it in Nmright up to the due moment for a maintainer's own PR, then flipped to… — not adopted: maintainer_policy is neverthe instant due fired — a confident claim that turned out false.stageTrackerAdoptionTextnow names the exception for theneverpolicy's auto-adopt countdown instead of promising an outcome it can't back for every author:… unless the author maintains this repository. Theadopt(default) andaskpolicies are unaffected — under those the hive genuinely does adopt on the next sweep, or at worst offers the person an explicitAdopt?decision instead of silently flipping.TestStageTrackerAdoptionNeverHedgesBeforeDuecovering both the fired (AutoDue) and pending (AutoAdoptAt) countdown text before anyAdoptionBlockedrecord exists, and updatedsrc/docs/labels-and-control-signals.mdto document thenever-policy carve-out.I considered instead moving the maintainer-permission check earlier in the sweep engine (so the record exists before
due), butrec.AdoptionBlocked() != ""also feedsTakeover.GraceElapsed,stageTrackerContributorGraceElapsed(WaitingOnYou) and the delegated-decision input'sAdoptionGraceElapsed— recording the hold ahead of the real grace window would make those report "grace elapsed" far too early for a maintainer-authored row. The text-only fix in the dashboard avoids that ripple entirely while still never asserting a false outcome.Filed from a low-severity (P3) correctness finding in the Hive review backlog (originally surfaced reviewing #15662, which has since merged).
Test plan
go test,gofmt, changelog-fragment guard, docs-citations guard)gofmt -lon all changed files (clean)adoptContributorPR/stageTrackerAdoption/stageTrackerAdoptionTextconfirming the new branches only fire whenPolicy == neverand no block is yet recorded, and don't touchGraceElapsed/WaitingOnYou/decision logic— hive: backend=claude model=claude-sonnet-5-5 effort=medium