Repository navigation
🐛 Review and CI chevrons state what the head waits on, never "activity unknown" - #16662
Conversation
…, never "activity unknown" A Review track with no verdict and no reviewer in flight now derives a concrete statement from the PR's lifecycle and its CI track (stage_tracker_review_fallback.go): closed/merged, draft, waiting for CI, not started (+ who is next), or — only when nothing was fetched for the head — "no data for head <sha7> yet · refreshing". The UI drops the "activity unknown · last observed …" prefix: settled verdicts stay known as observations age (dated in the tooltip), and a settled review explains itself — who decided, how long ago, each perspective's verdict with its first finding (perspectives field, dissent preserved). An aggregate approval the verified bundle does not back reads "evidence not verified" instead of unknown. Fixture invariant: no CI/Review tooltip contains "activity unknown" / "CI unknown" / "Review unknown". Supersedes #16634. Fixes #16617 Fixes #16661 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
|
There was a problem hiding this comment.
Intent alignment vs #16617 / #16661 (read at head 2a89b93)
The fallback ladder, perspectives[], the settled-stays-known UI change and the fixture invariant all land as described. One in-scope gap:
1. · needs reviewer (#16661 rule 4) is unreachable — ReviewerLane is always true.
stage_tracker_build.go:478 and :555 pass len(expectedPerspectives) > 0 as the "repo has a hive reviewer lane" signal, but stageTrackerExpectedPerspectives (stage_tracker_activity.go:356-366) falls back to review.DefaultPerspectives (pkg/review/review.go:38) whenever config names none, so it is never empty. The needs reviewer branch in stage_tracker_review_fallback.go:82-88 is covered only by the unit table that sets the flag directly; in production every CI-done head reads hive reviewer next, including hives with no reviewer configured — the PR body's "· needs reviewer when no hive reviewer lane is configured" is not implemented. The configured lane is cfg.Review.ReviewerAgents (pkg/config/config.go:7857), which is in scope at the same call site (cfg is used at stage_tracker_build.go:58):
--- a/src/pkg/dashboard/stage_tracker_build.go
+++ b/src/pkg/dashboard/stage_tracker_build.go
@@ -58 +58,2 @@
expectedPerspectives := stageTrackerExpectedPerspectives(cfg)
+ reviewerLane := cfg != nil && len(cfg.Review.ReviewerAgents) > 0
@@ -478 +479 @@
- stageTrackerReviewFallbackState(it.ReviewActivity, stageTrackerReviewFallbackInputsFor(&it, pr, len(expectedPerspectives) > 0))
+ stageTrackerReviewFallbackState(it.ReviewActivity, stageTrackerReviewFallbackInputsFor(&it, pr, reviewerLane))
@@ -555 +556 @@
- stageTrackerReviewFallbackState(it.ReviewActivity, stageTrackerReviewFallbackInputsFor(it, currentPR, len(expectedPerspectives) > 0))
+ stageTrackerReviewFallbackState(it.ReviewActivity, stageTrackerReviewFallbackInputsFor(it, currentPR, reviewerLane))(Still hive-wide rather than per-repo as #16661 words it; if a per-repo signal exists, prefer it. Either way a build-level test asserting needs reviewer with empty ReviewerAgents would keep the branch live.)
2. (advisory, low) #16617 fix item 2 asks that the caption name the first failing perspective (correctness ✗: <summary>). Here the caption stays Review findings · awaiting author (stage_tracker_activity.go:613) and the perspective lines ride only in the tooltip (tracker-row-strip.js:330). The tooltip does answer "which perspective and why", so this is acceptable, but the PR body/changelog should not imply the chevron words do.
— hive: agent=ci-maintainer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Hive status — updated in place; edits do not notify. Full history is in the Activity Monitor. Review findings from docs-currency — read at head
Review findings from Correctness review (head 2a89b93) — no blocker; the fallback ordering (verdict → dispatch → review-pending → fallback,
@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. Evidence bundle `#166625790865` sha256:144d2c8748dd8d73391121e3fa89207175308978c5c2be2604dbb712a933aacd — download from the dashboard or `hivectl review evidence #16662`. |
|
merged by Hive trusted-author auto-merge: author @clubanderson holds role owner and GitHub permission admin on repo; CI green at 5790865. |
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
There was a problem hiding this comment.
Intent-alignment re-review at head 5790865 (delta vs 2a89b93 is only the v5 merge — the PR's own diff is byte-identical apart from hunk offsets, so the two prior intent findings stand unchanged).
1. Still present — ReviewerLane is always true, so · needs reviewer is dead (src/pkg/dashboard/stage_tracker_build.go:481, stage_tracker_review_fallback.go:78-81)
The flag is passed as len(expectedPerspectives) > 0, but stageTrackerExpectedPerspectives (stage_tracker_activity.go:368-377) always falls back to review.DefaultPerspectives, so the length is never 0 and the PR body's "· needs reviewer when no hive reviewer lane is configured" cannot occur. Wire it to the actual reviewer configuration:
stageTrackerReviewFallbackState(it.ReviewActivity, stageTrackerReviewFallbackInputsFor(&it, pr, cfg != nil && len(cfg.Review.ReviewerAgents) > 0))
(or, if the hive can dispatch review without reviewer_agents, drop the branch and the · needs reviewer wording from the body/docs so the text matches what can ship).
2. Still present — the caption does not name the failing perspective (src/pkg/dashboard/stage_tracker_activity.go:613,624)
#16617 fix item 2: "the caption names the first failing perspective (correctness ✗: <summary>)". Both changes_requested branches still emit the fixed "Review findings · awaiting author"; the perspective only appears in perspectives[]/detail, so the chevron's own caption still can't say which perspective failed. out.Perspectives is already populated at :587 before this point, so:
case "changes_requested", "reject":
out.State, out.Caption = "changes_requested", "Review findings · awaiting author"
+ for _, p := range out.Perspectives {
+ if p.Verdict == "changes_requested" || p.Verdict == "reject" {
+ out.Caption = "Review findings · " + p.Name + " ✗ · awaiting author"
+ break
+ }
+ }
out.Blocker = stageTrackerFixerExhaustedSummary(agg)and the same loop in the bundle-fallback branch at :624. Update the Review findings · awaiting author expectations in stage_tracker_review_fallback_test.go / the #16661 node fixture accordingly.
— hive: agent=adjudicator-3 backend=copilot model=claude-fable-5.1 copilot=1.0.88
v5 (#16662) landed its own StageTrackerPerspective, trackerReviewPerspectives and the settled-state/ago-word tooltip for #16617. Keep the v5 shapes and wording (single-line ' · ' perspectives, 'stale · <iso> (Nm ago)', never 'activity unknown', waiting_ci is settled) and carry over what this branch still adds on top: the dissent-first tie-break in trackerLatestBundleVerdicts, the first finding appended to the changes_requested/human_decision caption on both sides, the blocker dedupe, and gating tooltip perspectives on a known display state. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: quality <280983584+hivecommons-hive[bot]@users.noreply.github.com>
Supersedes #16634, Fixes #16617 and Fixes #16661
Operator: an icon-only Review chevron whose tooltip said "Review activity unknown" — "not helpful to anyone."
What changes
stage_tracker_review_fallback.go(new):stageTrackerReviewFallbackStateturns a Review track that is stillunknownafter the verdict / dispatch / review-pending derivations into a statement, in order: merged/closed →Review · n/a (closed); draft →Review · draft, not requested; CI on the head running/queued/not all reported (or a partial ring) →Review · waiting for CI; CI done →Review · not started · hive reviewer next(· needs reviewerwhen no hive reviewer lane is configured); nothing fetched for the head →Review · no data for head <sha7> yet · refreshing(the only no-data wording, and it names the next step). CI's own no-data caption becomesCI · no checks observed for head <sha7> yet · refreshing.review_activity.perspectives[](name, verdict, first finding; a blocking verdict outranks another backend's approval on the same perspective) and adetailsentence ("Changes requested by the hive reviewer: 1 perspective posted findings …"). An aggregate approval the verified bundle does not back is stateunverified("Review · aggregate approves · evidence not verified"), notunknown.tracker-row-strip.js: the"<Track> activity unknown · last observed <state>"prefix is gone; settled states stay known past the freshness window (staleness is dated:stale · <ts> (12m ago)), a settled verdict's age and perspective lines ride in the tooltip, derived states carry their caption as words when the column is the stage's own, legacyReview unknown · reasoncaptions from an older spoke are rewritten, and the code-lane detour's "activity unknown" no longer leaks into the Review tooltip.TestStageTrackerReviewFallbackStateDerivesAConcreteStatement), inputs reader, perspectives dissent/detail; node fixture#16661(waiting_ci / not_started / draft / closed / unverified / none / legacy / settled-with-perspectives); prelude invariant over all row fixtures: no CI/Review tooltip containsactivity unknown,CI unknownorReview unknown.#16634
Its two intents — settled review/CI verdicts persist as observations age, and the tooltip lists each perspective's verdict + finding — are included here (same
perspectivesshape). It was red on an unrelatedapi_repo_item_decision_test.goexpectation; this PR supersedes it.