Repository navigation
π± Forward-merge v5 into v6 (2 commits) - #16688
Conversation
* π fix: label narrow dashboard navigation Make the drawer entry point recognizable when the sidebar is hidden, and keep its accessible name in sync with the drawer state. Hive-Run: #16654 Hive-Plan: dashboard-narrow-navigation Hive-Spec: issue-16654#navigation-access Co-authored-by: MikeSpreitzer <14296719+MikeSpreitzer@users.noreply.github.com> Signed-off-by: Karibhou <Karibhou@users.noreply.github.com> * fix: apply responsive navigation chip classes Signed-off-by: Karibhou <Karibhou@users.noreply.github.com> --------- Signed-off-by: Karibhou <Karibhou@users.noreply.github.com> Co-authored-by: Karibhou <Karibhou@users.noreply.github.com> Co-authored-by: MikeSpreitzer <14296719+MikeSpreitzer@users.noreply.github.com> Co-authored-by: hivecommons-hive[bot] <280983584+hivecommons-hive[bot]@users.noreply.github.com>
β¦ce row (#16684) running-at-level-6.md G8 called github.app_signed_commits YAML-only, but the dashboard has had the switch under Settings β Forge App since the toggle landed, and operator-reference/README.md had no row for the key at all. Add the row (default off, App-only, what it rewrites, fallbacks, Settings location, link to github-app-setup.md) and name the dashboard location in github-app-setup.md next to the YAML snippet. Signed-off-by: Andrew Anderson <andy@clubanderson.com> Co-authored-by: hivecommons-hive[bot] <280983584+hivecommons-hive[bot]@users.noreply.github.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 |
|
/ok-to-test |
|
@hivecommons-hive[bot]: Cannot trigger testing until a trusted user reviews the PR and leaves an DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
|
All three commits in this PR carry Fix needed from a maintainer (one of):
Deferring β no PR-local repair possible. cc @clubanderson π Hive Agent: β hive: agent=scanner backend=copilot model=claude-fable-5.1 copilot=1.0.88 |
|
Hive status β updated in place; edits do not notify. Full history is in the Activity Monitor. Hive filed review findings that do not block this PR as backlog items:
Cap: 10 backlog item(s) per PR per day. Review findings from
Fix: delete Other red checks, for the record: I have not merged or closed anything. Review findings from Forward-merge looks correct to me, including the one hand-made commit on top of it. What I verified on
Open question for whoever merges: Review findings from docs-currency (delta 3cb2e22 β c74cb26, read at head c74cb26): no in-scope blocker. The new One carried-wording gap to fix in v5 (does not block this forward-merge):
Prior [intent-alignment] note on Review findings from Reviewer adjudication: HOLD β no hive-side verdict is lossless and green; this needs a maintainer action, not another fix attempt. Evidence:
Still wanted / lossless: v6 is 7 commits behind v5 ( What unblocks it (maintainer-only, either one):
Changes: none pushed. Tests: none run (no code change); verified via GitHub API compares and check-run logs. Remaining risk: none beyond v6 ageing against v5 until a maintainer acts. Leaving Review findings from Reviewer adjudication: NO-VERDICT (stays needs-human; maintainer action required) β evidence: head c74cb26 is an exact merge of v5 tip c142df4 (7 commits, 0 behind v6, lossless, still wanted: v6 is 7 commits behind v5). 68/69 checks green. native-queue-policy currently fails as 'held' because of the needs-human label itself; with the label removed it fails on probot DCO = action_required for v5 commit 05ba6f5 (author 'Andy Anderson' vs Signed-off-by 'Andrew Anderson', same email). That is a merged maintainer commit on v5: rewriting its author in this PR breaks the v5βv6 invariant (daily v6-topup would re-carry 05ba6f5 and fail DCO forever) and no .github/dco.yml exists on v5/v6 so remediation commits are not accepted; adding one is a governance change the hive should not make. RECOMMEND-CLOSE is wrong (wanted + lossless; the workflow regenerates the same PR). changes: none. tests: gh api compare v6...sync/v5-to-v6 (ahead 8/behind 0), check-runs on c74cb26 (DCO action_required, all others success/skipped), read native-queue-policy.cjs assertEligible/BLOCKED_LABEL. Maintainer fix (one click): on the probot DCO check for head c74cb26 use 'Set DCO to pass' (email matches, only display name differs), then remove needs-human β native-queue-policy admits it on the next run. remaining risk: none beyond the human step; the next v5 push regenerates the head and the override must be repeated unless the author name is aligned in future commits. Review findings from Reviewer adjudication: DE-ESCALATE β evidence: the listed failure ("#16688: head/base moved, draft, closed, or held", native-queue-policy.cjs:91 assertEligible) is the check tripping on the needs-human label itself (BLOCKED_LABEL, :22), i.e. circular; the underlying cause that triggered the fix-loop escalation is probot DCO = action_required on 05ba6f5 (author "Andy Anderson", Signed-off-by "Andrew Anderson", same email) β a maintainer commit already merged on v5, not a defect in this branch. The branch is lossless and still wanted: compare v6...v5 = ahead 7, and the PR carries exactly those 7 commits (ff60153..c142df4) plus the merge commit; compare v6...sync/v5-to-v6 = ahead 8 / behind 0, so no rebase push is needed (and rewriting a forward-merge to re-sign a maintainer commit would change carried SHAs β not a hive move). Prior commits with the identical name mismatch (8565377, 8b824c4, 0a7212e) reached v6 only via maintainer-merged top-up #16683. changes: label-only β needs-human removed so native-queue-policy stops failing on the hold and will admit the PR the moment a maintainer applies the DCO app "Set DCO to pass" override on this PR (email matches; only display name differs) or merges it directly as with #16683; reviewer-passed added. No code pushed. tests: 68/69 checks green on c74cb26 incl. all test shards, coverage, patch-coverage, prow dco=success; only probot DCO action_required. remaining risk: none to the merge content; needs one maintainer action (DCO override) β until then the PR will re-escalate and, per the one-pass ladder, route to a human rather than another reviewer pass. Review findings from Reviewer adjudication: NEEDS-HUMAN STANDS (no REPAIR / DE-ESCALATE / RECOMMEND-CLOSE applies) β evidence: Why no verdict: REPAIR would require re-signing a v5 commit on this branch, rewriting its SHA and desynchronising the v5βv6 chain (lossy, and the next top-up would re-carry the original); DE-ESCALATE/rebase re-carries the same commit; RECOMMEND-CLOSE is wrong because the top-up is wanted and v6-topup.yml would reopen it. The previous top-up #16683 had the identical pattern (DCO action_required + native-queue-policy failure) and was admin-merged by a human at 01:18Z. Needs a human (or an ISSUES_PRS_MERGE-tier actor) to admin-merge as with #16683, or a policy decision on how Another agent tried to open a PR from |
There was a problem hiding this comment.
Decision needed: @clubanderson should decide whether to admin-merge #16688 as-is (probot DCO action_required on carried commits ff60153/05ba6f5 is what fails native-queue-policy) or exempt forward-merges from the DCO evidence gate in .github/scripts/native-queue-policy.cjs:109.
Clean forward-merge, blocked by a check the hive cannot repair. Needs a maintainer.
- Content: 8 files, +14/β7, two carried v5 commits (docs for App-signed commits + the #16654 navigation label) plus the merge commit;
mergeable = MERGEABLE, no conflicts. native-queue-policyfails withError: DCO: action_requiredthrown at.github/scripts/native-queue-policy.cjs:109(evidenceForHeadβassertEvidence, job 126984881523 log line 218). It is reading the probotDCOstatus, which isaction_requiredbecause the carried commits' author andSigned-off-bynames differ:ff60153authored ashivecommons-hive[bot]but signed off by Karibhou,05ba6f5authored "Andy Anderson" signed "Andrew Anderson". The prowdcocheck on the same head passes.- A forward-merge cannot change those commits without rewriting v5 history, so this will recur on every top-up carrying such commits β same pattern as #11001, except that
native-queue-policynow turns the cosmetic probot status into a hard failure.
Decision: admin-merge this one, and/or have native-queue-policy.cjs treat probot DCO: action_required on v6-topup PRs as non-blocking (the prow dco check already covers sign-off). I have not merged or changed anything.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
c8a7d6e to
cc8d968
Compare
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. |
There was a problem hiding this comment.
Decision needed: @clubanderson should decide whether to merge this top-up as-is (it carries 8277c2c4c, which v5 no longer has after its tip was rewritten to 05ba6f563) or close it and let the v6-topup workflow regenerate from the current v5.
Content is correct; ancestry is not. The merge commit cc8d96803 is a pure merge: git merge-tree v6 8277c2c4c produces tree 5faccb209β¦, identical to this PR's tree, so nothing was hand-edited. The two carried changes (docs for github.app_signed_commits, the narrow-dashboard navigation label in topbar.html / chat.js / regenerated static/index.html) are the same bytes v5 has.
But the second parent, 8277c2c4c ("docs: app_signed_commits is a Settings toggle⦠(#16684)", author "Andrew Anderson", 2026-11-05 01:40Z), is no longer on v5. origin/v5 now tips at 05ba6f563, same title, same tree, same parent ff60153d8, re-committed at 01:20Z UTC as "Andy Anderson". So v5's tip was rewritten after this PR was opened. Merging this PR would put 8277c2c4c into v6's history while v5 carries 05ba6f563 for the same change. The next top-up would then merge 05ba6f563 as a content-empty 1-commit forward-merge, and the ancestry that forward-merges exist to preserve (running-at-level-6.md, G6) would diverge for that commit.
Two clean options, both a maintainer's call: (a) close this PR and let .github/workflows/v6-topup.yml regenerate sync/v5-to-v6 from the current v5 tip, or (b) merge as-is and accept the duplicate commit. I have not closed anything.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Needs a decision, not a merge: the automerge lane will not merge this unattended. Head: verdict approved @6421cb6, checks unknown, soak 0s of 10m0s. How to release it:
Clear the cause and the hive re-evaluates on its next pass; the stage moves to Updated 2026-11-05 16:12 UTC. |
β¦epair (#16698) #16648 let a hive-authored PR on a pushable base-repo head be amended in place whatever its unsigned commits were linked to, because commits the hive's agents push as project.ai_author counted as "other" (only App-bot linkage made a commit the hive's). As a result a collaborator with push access who added an unsigned commit to a hive branch got a Signed-off-by asserted for them by the App. The hive/other split now also counts commits GitHub links (by author login) to one of the hive's identities (StageLabelPolicy.HiveAuthor), so #16648's case still amends in place, and in-place repair again requires every unsigned commit to be the hive's. A hive-authored PR carrying anyone else's unsigned commit waits for a person; it is still never superseded. Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com> Co-authored-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
cc8d968 to
f72bdc4
Compare
There was a problem hiding this comment.
Looks correct to me as a forward-merge β verified the merge carries v5 faithfully, not inferred from the body.
What I checked (head f72bdc4):
compare/v6...f72bdc4andcompare/v6...v5touch the same 13 files with identical per-file additions/deletions; for the four files where v6 has its own surrounding changes (src/docs/operator-reference/README.md,src/docs/running-at-level-6.md,src/pkg/dashboard/static-src/html/topbar.html,src/pkg/dashboard/static/index.html) the+/-lines are byte-identical to v5's β only context differs. No hunk dropped or conflict mis-resolved.- The carried #16698 code resolves on the v6 tree:
automerge_stage_dco.go:150callsdcoCommitStatsForHive(..., policy.HiveAuthor)andStageLabelPolicy.HiveAuthorexists atautomerge_stage_labels.go:150;automerge_contributor_adopt.go:342gates the in-place amend onDCOUnsignedOther == 0, matching the v5 fix. Alltest (β¦)shards, golangci-lint and gosec are green.
The one red check is not this diff's fault: native-queue-policy fails with Error: DCO: action_required (.github/scripts/native-queue-policy.cjs:109) because probot DCO reports action_required on the upstream-authored commits in the range β the prow dco context is SUCCESS and dco-signoff: yes is set. Same pattern as prior v5βv6 top-ups; it needs a maintainer /approve + /lgtm for tide rather than a code change.
Note: the title says "2 commits" but the merge commit message and the commit list carry 3 (ff60153d, 05ba6f56, 4e7e4048) β the title appears to predate the last v5 push; harmless.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
β¦uash-unsafe checks (#16706) Count project.ai_author-linked commits as the hive's in the squash-unsafe check too (#16703), re-read DCO splits cached before hive identities counted (#16701), and document the email-derived login-linkage trust assumption for in-place re-signing (#16702). Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com> Co-authored-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com>
f72bdc4 to
141803c
Compare
There was a problem hiding this comment.
Verified this is a pure automatic forward-merge and the red check is a DCO artifact, not a content problem.
- Merge commit
141803cchas parentsb43b6b9(currentv6tip) ande3f54c2d(currentv5tip).git merge-tree --write-tree b43b6b9 e3f54c2dproduces tree8200f445β¦, which is byte-identical to the PR's tree β no hand-resolved hunks. - The 4 carried commits (
ff60153d,05ba6f56,4e7e4048,e3f54c2d) are all already-mergedv5PRs (#16678, #16684, #16698, #16706); nothing new is introduced here. native-queue-policyfails withError: DCO: action_required(.github/scripts/native-queue-policy.cjs:109) because the probotDCOcheck isaction_requiredon a merge carrying other authors' commits. The prowdcocheck passes. This is the same non-blocking pattern seen on previous forward-merges (e.g. #8793).
Looks correct to me. A maintainer needs to merge this past the probot DCO status; nothing in the diff warrants changes.
β hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
The --provider-file consumer side of cline.run only had a positive test: nothing exercised the guard that refuses a mounted provider file that resolves inside the Git workspace (directly or through a symlink), the unreadable-mount branch, or a narrowed file that fails credential narrowing. Each case now asserts the turn exits 1 with the specific denial, leaks no credential value, and never launches the Cline CLI. Signed-off-by: quality <280983584+hivecommons-hive[bot]@users.noreply.github.com> Co-authored-by: quality <280983584+hivecommons-hive[bot]@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
141803c to
fc9aad5
Compare
There was a problem hiding this comment.
HUMAN DECISION NEEDED β admit this forward-merge despite the probot DCO check staying action_required (native-queue-policy admission blocked), or wait for the DCO provider.
Security delta re-review at fc9aad5 (vs 141803c):
- The only content change between the two heads is v5 commit 629ca32 (#16705):
src/crux/cline/provider_file_test.go:199-264addsTestMountedProviderWorkerDenials. Test-only; the stub CLI writes a marker insidet.TempDir()and the assertions check that denied provider files neither launch the CLI nor leaktest-key. No production code, permissions, or trust boundary changes. The other 17 files listed by the compare endpoint are byte-identical at both heads (artifact of the re-merged base). - Prior finding still present:
native-queue-policyfails at this head ("Admission blocked: DCO provider unavailable",.github/scripts/native-queue-policy.cjs:109) because probotDCOisaction_requiredon the merge commit. This is a merge-admission/governance state, not a defect introduced by the diff; the carried commit already passed DCO when it landed on v5 via #16705.
β hive: agent=adjudicator-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
π Fix loop escalated β human attention neededThis PR has failed CI on 3 distinct fix attempts (new commits, still red). The hive has stopped dispatching further automated fixes for it. Failing checks: native-queue-policy Raw failure evidence (from check-run annotations): Remove the |
β¦ng (#16730) refetchHeadCI declines to list check runs when the process-wide REST budget is critical or the shared pacing window is engaged. The test read both globals without isolating them, so state leaked from an earlier test in the package could make the flush a no-op (listings = 0). Isolate the REST accounting store and reset pacing for the test. Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com> Co-authored-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com>
Signed-off-by: hive-release[bot] <hive-release@hive.kubestellar.io>
90f0d19 to
0d76754
Compare
There was a problem hiding this comment.
Semantic merge conflict: src/crux/serve/log_tail.go is left dead on the merged head, and golangci-lint (required in Go Security Analysis) fails on it.
What I verified at head 0d76754:
src/crux/serve/log_tail.goexists onv6only (404 onv5). It was introduced by v6 commit6323f19c5("reconcile forward-merge build and source guards") and definesnewTail(line 7) andequalLines(line 20).- The v5 commit carried here (
changelog.d/changed-crux-streamtail.md) moves that logic tosrc/crux/internal/streamtail/streamtail.goand rewires the only callers:src/crux/serve/serve_follow.go:78now callsstreamtail.Fresh(prev, cur), andsrc/crux/serve/serve_test.godrops thenewTailcases (TestIsTrueAndNewTailβTestIsTrue). - Git merged cleanly because v5 never knew about
log_tail.go, so the file survives with zero references. Lint output on this head:src/crux/serve/log_tail.go:7:6: func newTail is unusedand:20:6: func equalLines is unused.
Fix: delete src/crux/serve/log_tail.go on sync/v5-to-v6 (the streamtail package now carries that logic and its tests). I cannot push to the sync lane's branch, so flagging for whoever tops up this merge.
Separately, native-queue-policy fails with #16688: head/base moved, draft, closed, or held (.github/scripts/native-queue-policy.cjs:92) β that is the hold/needs-human state, not a defect in the diff; it should clear once the lint fix lands and the hold is released.
The remaining 37 files are straight v5 carries (automerge stage labels, agent_runtime_supervise, dashboard topbar, docs/changelog fragments); build-contributor, hub/agent/scheduler test shards and contract-tests are green on this head, so I found no other merge-induced breakage.
β hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
v6 moved newTail/equalLines into serve/log_tail.go while v5 switched the only caller (serve_follow.go) and its test to streamtail.Fresh. On the merged head both functions are unreferenced and golangci-lint reports them as unused. Delete the file. Hive-Fix-Attempt: 3/6 Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: scanner <280983584+hivecommons-hive[bot]@users.noreply.github.com>
There was a problem hiding this comment.
Decision needed: @clubanderson should decide whether to set the probot DCO check to pass for the name-mismatched v5 commit 05ba6f5 and remove needs-human so this v5βv6 top-up can merge.
Current head 6421cb6 looks correct to me; the remaining blockers are both label/check decisions, not code.
What I verified at 6421cb6:
- The one commit that is not a v5 cherry (
6421cb6, "Remove orphaned serve/log_tail.go") is right: on the merged treesrc/crux/serve/serve_follow.go:78already callsstreamtail.Fresh, and the only remainingequalLinesisstreamtail's own copy atsrc/crux/internal/streamtail/streamtail.go:20. Nothing insrc/crux/servereferencesnewTail/equalLines, so deleting the file is the correct resolution of the v5/v6 overlap rather than a lost change.go build ./crux/serve/ ./crux/internal/streamtail/ ./cmd/hive/ ./pkg/github/automerge/ ./pkg/stagelabels/passes on this head. native-queue-policyis red at this head for a different reason than the earlier DCO diagnosis: it now throws fromassertEligible(.github/scripts/native-queue-policy.cjs:92, "head/base moved, draft, closed, or held") becauseneeds-humanmatchesBLOCKED_LABELatnative-queue-policy.cjs:22. The hold label itself fails the check, so it will stay red until the label is removed; probotDCOis stillaction_requiredbehind it.
Nothing further for the hive to push here. This is the two-click decision already requested above, restated against the current head.
β hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
Automated v5 β v6 top-up (#7563, Track 7).
Carries 2 commit(s) from v5 onto v6.
This runs on every push to v5 (v5 only moves when a v4βv5 top-up or a v5 PR merges), plus a daily safety net, so v6 feature work never ages against a stale base.
Opened by .github/workflows/v6-topup.yml. If this merge conflicts the workflow fails instead of opening it β a conflicting forward-merge is a human decision (see #7199).