Skip to content

πŸ› Close partially-addressed issues the scanner verified complete - #19301

Open
clubanderson wants to merge 2 commits into
v5from
fix-scanner-verified-close
Open

clubanderson wants to merge 2 commits into
v5from
fix-scanner-verified-close

Conversation

@clubanderson

Copy link
Copy Markdown
Member

Summary

Finished Refs-only issues now get closed. The scanner's verify-once note said "this can be closed" in free text (#17157, #19177) and nothing acted on it.

  • Marker: the verify-once kick asks the scanner to end its note with <!-- hive:verification verdict=complete prs=N[,M] --> or verdict=partial (github.VerificationMarker).
  • pkg/github/verified_close.go (closeVerifiedComplete, run from the issue stage sweep next to the tracker auto-close): for open hive/partially-addressed issues (or ones already in the lane), reads the marker from hive-authored timeline comments only, then on two reads one settle interval apart (reuses trackerReadDue / trackerSettled / trackerEvidenceHash from ✨ Close tracker and umbrella issues whose referenced work is done (opt-in)Β #19151):
    • close as completed, with an evidence comment citing the merged PRs, when every cited PR (marker plus merged claim) is merged and no timeline-linked PR is open;
    • verify-close lane (hive/verify-close, a hard-suppress park at stage/045-needs-decision, one status note naming the unmet guard) when a cited PR is not merged, a linked PR is open, notes disagree, no PR is cited, more than 20 PRs, or the reporter-confirmation gate applies;
    • verdict=partial adds hive/verified-open.
    • Never touched: hold, hive/tracking, umbrella or tracker, open sub-issue, reopened by a person, lane label removed by a person, no marker.
  • Dashboard: a lane issue's decision reads "The scanner verified this issue complete" and leads with one-click Close as completed (decide:close-issue).
  • Relaunch claims: kept as is, explained in the reference doc. The relaunch notice fires only for a fresh CLI, and crux never resumes the old session there, so "same agent resumes" does not reach that path. The marker now carries the verdict across a relaunch, and the verify-once gate still reads the ledger's last attempt.
  • Lane naming: the brief asked for stage/050-verify-close, but 050 is the PR stage stage/050-review-pending and the stage table keeps issue (010 to 045) and PR (050 to 100) stages disjoint. So the lane is a 045 park label instead of a new numbered stage. It can become its own stage later if wanted.
  • Docs (labels reference, stage labels, notifications inventory) and changelog fragment.

Tests

  • pkg/github/verified_close_test.go: marker parsing; every guard in decideVerifyClose; the candidate pre-filter; the sweep end to end (close only on the second read with evidence; lane for an unmerged cited PR, an open linked PR and disagreeing notes; partial adds verified-open; hold, tracking, umbrella, person reopen, person lane removal, foreign-author marker and free text never touched; a plain issue costs no timeline reads); the lane label parks at 045.
  • pkg/dashboard/stage_tracker_verify_close_test.go: lane popover leads with an enabled one-click close, plus a control.

Fixes #19299

The scanner's verify-once note said "this can be closed" in free text
and nothing acted on it, so finished Refs-only issues sat at 010 until a
person closed them (#17157, #19177). The verify-once kick now asks for a
machine-readable <!-- hive:verification verdict=complete|partial prs=N -->
marker; the issue stage sweep reads it from hive-authored comments and,
on two reads one settle interval apart, closes the issue with an evidence
comment when every cited PR merged and no linked PR is open, or parks it
in a hive/verify-close 045 lane the dashboard shows with a one-click
Close. verdict=partial adds hive/verified-open. Held, tracking, umbrella,
open-children, person-reopened and person-unlaned issues are skipped.

Fixes #19299

Signed-off-by: Andy Anderson <andy@clubanderson.com>
@kubestellar-prow

Copy link
Copy Markdown
Contributor

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign nicholasjackson for approval. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Details Needs approval from an approver in each of these files:

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

@kubestellar-prow kubestellar-prow Bot added the size/XL Denotes a PR that changes 500-999 lines, ignoring generated files. label Nov 19, 2026
@github-actions

github-actions Bot commented Nov 19, 2026 •

Copy link
Copy Markdown
Contributor

Open PR overlap check

Base branch: v5

Conflicts are checked with every PR merged onto the current tip of v5, so a PR that only needs a rebase is not reported as conflicting with its siblings.

PR Title Relation Files
#19290 πŸ› Keep human-only follow-ups out of the owner-delegate's default approve same files src/pkg/dashboard/stage_tracker_decision.go
#19266 πŸ› Sentinel judges the PR's current head and clears its own stale alert same files src/docs/notifications.md

@hivecommons-hive

hivecommons-hive Bot commented Nov 19, 2026 •

Copy link
Copy Markdown
Contributor

@clubanderson this PR is back in your court (stage/072-returned-to-contributor). The hive will not merge it until the head changes.

Failing checks:

  • test (rest 2/4)
  • test (rest 3/4)

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.

The lane note carries a hive marker so it is idempotent, the label is
added directly (its colour comes from the escalation label definitions),
and verified_close.go joins the direct-write baseline like the tracker
and decomposed-parent closes.

Signed-off-by: Andy Anderson <andy@clubanderson.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

dco-signoff: yes Indicates the PR's author has signed the DCO. size/XL Denotes a PR that changes 500-999 lines, ignoring generated files.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

πŸ› Close partially-addressed issues the scanner verified complete

1 participant