Repository navigation
π± ci(dco): treat upstream v6 history as landed on fork sync PRs - #196
Conversation
A merge-method/merge PR into a tuna/* fork line carries upstream hivecommons/hive v6 commits unchanged, signed off by their own authors. Fetch upstream v6 and add it to LANDED_REFS for those PRs only, so the squash-attribution gate skips commits upstream already gated. Commits authored in the fork are not reachable from upstream v6 and are still checked. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: James Reilly <jreilly1821@gmail.com>
5b78b9e
βΉοΈ Sentinel notice β informational; author is trusted (owner), merge is not blockedHive flagged this PR (author @hanthor, head
Hive did not add the |
Open PR overlap checkBase branch: Conflicts are checked with every PR merged onto the current tip of
|
There was a problem hiding this comment.
Decision needed: @hanthor should confirm that commits reachable from upstream hivecommons/hive v6 may skip the sign-off gate on labelled merge-method/merge PRs into tuna/* branches.
Routing, not blocking: this edits a CI workflow (.github/workflows/dco-squash-attribution.yml) and loosens a sign-off gate, so a person should confirm it.
What I read. The new step (:63-80) only runs when the repo is not hivecommons/hive, the base starts with tuna/, and the PR carries merge-method/merge. It fetches upstream v6 and passes it through LANDED_REFS. The checker splits that variable by word (src/scripts/check-squash-signoff-attribution.sh:149) and skips commits reachable from each ref. The change works as described. When the step is skipped its output is empty, so the existing value is unchanged.
The decision. Anything reachable from upstream v6 is now exempt, and the label that enables this can be applied by anyone with triage rights. The body cites an owner decision on #195, so this looks intended. I still can't treat a gate relaxation in a workflow as approved. Commits written in the fork are still checked, as the body says.
Suggestion, not a defect. Neither the step nor the two-ref LANDED_REFS case has a test. test-check-squash-signoff-attribution.sh could cover the second ref.
I have not approved or merged anything.
β hive: agent=reviewer backend=claude model=claude-sonnet-5-5 effort=high claude=2.1.322
|
Needs a decision, not a merge:
Head: verdict requires_human @5b78b9e, 17/19 checks green, 2 pending, soak 0s of 30m0s. What to do: a trusted merger approves this head ( Clear the cause and the hive re-evaluates on its next pass; the stage moves to |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: James Reilly <jreilly1821@gmail.com>
|
Thank you for your contribution! Your PR has been merged. We'd love to hear how your experience was: share feedback |
Owner decision for #195: the "Sign-off survives the squash" gate exempts commits that are already in upstream
hivecommons/hivev6 history.merge-method/mergewhose base is atuna/*fork line, and never inhivecommons/hiveitself.v6intorefs/remotes/upstream/v6and adds that ref toLANDED_REFS. The checker already skips commits reachable fromLANDED_REFS, so the script itself is unchanged.π€ Generated with Claude Code