Repository navigation
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 |
|
@Karibhou this PR is back in your court ( Failing checks:
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 (2 waiting): fixes are handed out before new work as soon as you have capacity. |
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of No conflicting open PRs or same-file overlaps were found. |
|
CI triage (scanner, PR-local): Cause:
Fix: extend both lines to const STAGE_TRACKER_ISSUE_DECISION_COLUMN = 10, STAGE_TRACKER_TRIAGE_NEEDED = 15, STAGE_TRACKER_POSSIBLY_EXISTS = 17, STAGE_TRACKER_OPEN_FOR_DISCUSSION = 18;( π Hive Agent: β hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88 |
There was a problem hiding this comment.
docs-currency (read at head 91e8bfbad)
The doc surfaces touched here agree with the code: the caption quoted in src/docs/dashboard-sections/activity-monitor.md:71 matches stage_tracker_caption.go (Open for discussion (<label>); no agent work is queued β claim it to take it on), the "not a gate" / precedence wording in stage-labels.md:98 matches unclaimedIssueStage in derive_issue.go, and governor-thresholds.md:5 states the queue-depth decision #17150 asked for explicitly. The #17054 title-only limitation is stated honestly in stage-engine-plan.md ("reads labels only, never titles").
One advisory gap, low:
ADR-0025 has no 018 extension record. src/docs/adr/0025-central-stage-engine.md:420-426 ("Scope and extension points") records every post-epic stage addition β 017 has its own paragraph there β and the ADR's line 11 points readers at that exception list. 018 is the same kind of addition and #17150 asks to update "any stage-tracker / stage-label reference". Suggested paragraph after line 426:
Leaving 017 for 010 or closed is never a regression.
+
+**Open for discussion (`stage/018-open-for-discussion`, #17150).** Added
+after the epic: an issue-only stage between 017 and 020 with no dwell.
+Unlike 017 it is derived, not written by a gate: `DeriveIssue` returns it
+for an open issue labelled `proposal` or `kind/design` with no live claim
+(`IssueFacts.DiscussionLabel`, resolved in `pkg/github`), reason
+`open for discussion: <label>`. It is not a suppress: the claim gate is
+unchanged and a claim stages 020/030. `040 β 018` under
+`hive/partially-addressed` is progress; leaving 018 is never a regression.β hive: agent=ci-maintainer backend=copilot model=claude-fable-5.1 copilot=1.0.88
Review follow-ups (below
|
|
Hive status β updated in place; edits do not notify. Full history is in the Activity Monitor. Review findings from Intent alignment (#17150) β read at head 91e8bfb. The diff implements the issue's proposed rule faithfully: One advisory on acceptance criterion 5 ("both #17054 and #17020 land on the new stage"):
β hive review: perspective=intent-alignment backend=omp model=anthropic/claude-opus-5-5 Review findings from Correctness review (head The engine, store and transition-table changes check out: 1. Four Node harnesses fail:
|
|
Pushed
π Hive Agent: |
There was a problem hiding this comment.
Decision needed: @clubanderson should decide whether #17150 is fully resolved by a label-only rule (so this PR may Fixes #17150 and #17054 gets a proposal label by hand), or whether π-title matching must land before the issue closes (then Refs #17150).
Code looks correct to me; one intent question for the issue author.
I read the diff against the v5 tree (merge-base df1061a) plus the callers, not just the patch.
Verified correct
src/pkg/stagelabels/stagelabels.go:448βBefore()comparesNumber(), not enum order, so appendingOpenForDiscussionto the enum while giving it number 18 keepsSettlesTriage()(:460) returning false for 018, as the body claims.src/pkg/stagelabels/store.go:1213βIsRegression(OpenForDiscussion, *)is false (parking bay). Falls into 018 reuse the existingClaimed/PRInProgress/CoveredByPRcases, so 030β018 and 040β018 are regressions exactly as 030β010 / 040β010 are, andissueProgress(derive_issue.go:266) exempts thehive/partially-addressedcase. Consistent with the transition-table comments.src/pkg/stageengine/derive_issue.go:155-230β likely-done, hold, blocked, needs-human, 017 and 045 parks all return beforeunclaimedIssueStageis reached, so the stated precedence holds; umbrella + proposal with a stray claim now derives 018 rather than 010, which matches "mirror the epic behaviour".src/pkg/github/issue_stage_labels.go:306βissueDiscussionLabelis read-only on labels; the claim gate is untouched and the newissue_claim_gate_test.gorows pin that.- Dashboard:
stageByNumber(stage_tracker.go:1099) walksTable, so 18 resolves;stage_tracker_human_attention.go:47andstage_tracker_presence.go:121list 017 but not 018 β correct, since 018 always carries a reason and the latter is PR-only. The remaining 017-only references (config/issue_filter.go:66,overview_status.go:467, prior-art gate) are hard-suppress specifics that 018 deliberately does not share.
The question (intent-alignment, not a defect)
#17150's acceptance criteria say "Both #17054 and #17020, replayed through the engine, land on the new stage." The PR body candidly says #17054, as labelled today (Epic, approved-direction, no proposal/kind/design), still derives 010, because titles are not plumbed into IssueStageInput. The code does exactly what the body says; the gap is between the body and the issue. With Fixes #17150 the issue closes on merge with that criterion unmet unless someone adds proposal to #17054. Either outcome is fine β it just needs the issue author to pick one.
Also noted: governor-thresholds.md chooses "018 issues still count toward queue depth"; the issue allowed either answer as long as it is stated, and it is.
Nothing here blocks landing from my side. I have not approved, closed or labelled anything.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
CI triage (scanner): the only red check, π Hive Agent: β hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88 |
|
Failure:
Baseline helper verdict: Test review (quality): the added π Hive Agent: β hive: agent=quality backend=copilot model=claude-fable-5.1 copilot=1.0.88 |
|
@Karibhou the hive has adopted this PR ( |
β¦for-discussion An open issue labelled `proposal` or `kind/design` with no live claim now derives the new issue-only stage/018-open-for-discussion instead of stage/010-actionable, so a proposal whose Refs RFC PR merged no longer reads as idle agent work (hivecommons#17020 fell 040 -> 010). - stagelabels: OpenForDiscussion (018, open-ended); leaving it is never a regression. - stageengine: IssueFacts.DiscussionLabel; unclaimedIssueStage applies it in the actionable/reporter-triage buckets and to an umbrella whose claim is not staged; reason "open for discussion: <label>"; issueProgress books 040 -> 018 under hive/partially-addressed as progress; transition table + diagram gain the 018 edges. - pkg/github resolves the label; claim gate unchanged (proposal, kind/design and the 018 label are not refused); goldens replay hivecommons#17020/hivecommons#17054. - dashboard: Actionable column word/caption "Open for discussion", shared row asset + hub globals carry STAGE_TRACKER_OPEN_FOR_DISCUSSION. - docs: stage labels, plan amendment, governor thresholds (018 issues still count toward queue depth), activity monitor, CONTRIBUTING. Fixes hivecommons#17150 Co-authored-by: Andy Anderson <407614+clubanderson@users.noreply.github.com> Hive-Run: hivecommons#17150 Hive-Plan: stage-engine-plan.md#amendments-after-pr-e Hive-Spec: hivecommons#17150#proposed-rule Signed-off-by: Karibhou <Karibhou@users.noreply.github.com>
β¦ ADR-0025 records 018 Signed-off-by: Karibhou <Karibhou@users.noreply.github.com>
be84e9c to
1143dcf
Compare
There was a problem hiding this comment.
Read the engine, label and dashboard code at 1143dcf before the body. The derivation change looks correct to me; one acceptance-criterion gap the body already declares is worth a maintainer's eyes.
What I verified
stageengine/derive_issue.go:276-282βunclaimedIssueStageonly short-circuits to 018 whenDiscussionLabelis set; every other path falls through to the unchangedactionableIssueStage, so 010/015 derivation for non-proposal issues is untouched. It is wired into exactly the three unclaimed sites (actionable/reporter-triage bucket at:227, unstaged umbrella claim at:220);BucketClaimed/BucketInProgressare not routed through it, so a live claim still stages 020/030.stagelabels/stagelabels.go:448-450βBeforecomparesNumber(), not the enum value, so appendingOpenForDiscussionto the persisted enum while giving it number 18 keepsSettlesTriage(018 < 020β false) and the dashboard< trackerNumberClaimedguard (stage_tracker_caption.go:138) consistent.stagelabels/store.go:1216β leaving 018 is never a regression;040 β 018is a regression unlessissueProgress(derive_issue.go:266-271) seeshive/partially-addressed, matching the issue's AC.github/issue_stage_labels.go:306-317β label match is case-insensitive and returns the issue's spelling, whichReasonDiscussionLabelthen surfaces in the caption.issue_claim_gate.gois not in the diff β confirmed.- Not a defect, noting for completeness:
stage_tracker_human_attention.go:47doesn't list 018 in the actor-reason switch, but 018 is also not instageTrackerNeedsHumanStages(stage_tracker_build.go:620), so an 018 item never reaches that code.
For the maintainer (@clubanderson, as the issue author)
#17150's acceptance criteria say both #17054 and #17020 should land on the new stage. The PR body is explicit that it keys on labels only, so #17054 (Epic, approved-direction, no proposal/kind/design) still derives 010 until someone adds proposal β pinned by golden regression/17054-rfc-epic-without-proposal-stays-010. That is consistent with the issue's own "Proposed rule" (which is label-based), but not with the AC line. If labelling #17054 is acceptable, this is complete; if π-title matching is required, it needs the title plumbed through IssueStageInput, which the author deliberately left out. Flagging so the choice is deliberate, not accidental. I have not changed anything.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
Fixes #17150
What changed
An open issue labelled
proposalorkind/designwith no live claim now derives a new issue-only stage,stage/018-open-for-discussion, instead ofstage/010-actionable. The Activity monitor shows it as Open for discussion in the Actionable column, and the caption names the label. It stays claimable: a claim stages it 020/030 as on any other issue.stagelabelsOpenForDiscussion(018, no dwell limit, label description is 100 chars or fewer).IsRegressiontreats leaving 018 like leaving any other parking stage: never a regression.stageengineIssueFacts.DiscussionLabel. The newunclaimedIssueStageapplies it in the actionable and reporter-triage buckets, and to an umbrella issue whose claim isn't staged. The reason isopen for discussion: <label>.issueProgressnow books040 β 018underhive/partially-addressedas progress. Adds the 018 edges to the transition table and the mermaid diagram (they're kept in sync byTestTransitionTableMatchesDiagram). AddsReasonDiscussionLabelfor presentation code.pkg/githubresolveIssueLabelFactsresolves the label. The claim gate is unchanged. New test rows show thatproposal,kind/designand the 018 label are not refused.< 020for "before a claim". JS words/anchor/detour are updated,index.htmlis regenerated, and the sharedtrackerrowasset,RequiredGlobalsand hubactivity.jsall carrySTAGE_TRACKER_OPEN_FOR_DISCUSSION = 18. Node harnesses that already liftedSTAGE_TRACKER_POSSIBLY_EXISTSnow lift the new constant too.Precedence: likely-done, hold,
blocked, parks (045/017) and a covering PR still outrank 018. 018 outranks the untriaged 015, because a person owns the next step either way. 018 doesn't settle triage, soneeds-triagestays on.Queue depth: the governor counts the enumerated issue queue (
actionable.Issues.Count), not stage labels. So 018 issues still count toward actionable queue depth.governor-thresholds.mdnow says so.Acceptance criteria
proposal/kind/designderives 018; with a live claim it derives 020/030 (TestDeriveIssueTable, goldensregression/17150-*).TestStageTracker_OpenForDiscussionStage, vocabulary/pin tests).040 β 018after a mergedRefsPR is booked as progress, not regression (TestDeriveIssueProgress,TestReconcileProgressFallBooksNoRegression).issue_claim_gate.gois unchanged (TestIssueClaimBlockReasonrows).Epic,approved-direction, noproposal/kind/design) still derives 010. The proposed rule keys on labels, and the stage engine never sees titles:IssueStageInputcarries no title, and the out-of-cycle FactSink path doesn't either. The goldens pin both sides:regression/17054-rfc-epic-labelled-proposal-is-018andregression/17054-rfc-epic-without-proposal-stays-010. Addingproposalto π RFC: LLM-ranked knowledge advisor (OMP-style)Β #17054 moves it. Matching π RFC titles would need a title plumbed through enumeration and FactSink, so I left it out of this PR.Testing
Per
AGENTS.md, I didn't run the Go test suite locally; CI is the verdict.gofmt -lis clean on the changed files.index.htmlwithgo run ./pkg/dashboard/webstatic/gen.go run, which compiledpkg/github,pkg/stageengine,pkg/stagelabelsandpkg/dashboard. Output:stage/018-open-for-discussion, reasonopen for discussion: proposalstage/010-actionable; withproposalβ 018kind/designβstage/020-claimedIsRegression(040β018)=true(unmerged close),IsRegression(018β010)=false,SettlesTriage(018)=falseindex.html:stageTrackerAnchor(18)β 10 (Actionable column)STAGE_TRACKER_ACTIVE_WORD[18]= "Open for discussion"stageTrackerHeadline= "Open for discussion"humancovered.jsonrows for the new Table stage, identical to the 017 rows because both are below 040, plus 7regressions.jsonrows. CI'sTestIssueStageGoldenschecks them byte for byte.β hive: backend=omp model=anthropic/claude-opus-5-5
π Hive Agent:
contributor| SHA:91e8bfbad