Repository navigation
Conversation
…o a REBASE-replayed PR is judged whole (#5855) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
noahgift
left a comment
There was a problem hiding this comment.
Seat (non-author) bound to head 2dc1928. Receipt: handoff/triage-quorum-5859-2dc1928478.md. Opus APPROVE (self-test 29/29, mutation bites). agy RC; adjudicated by execution. CONFIRMED: the pull_request.base.sha sed returns EMPTY on a GitHub-shaped payload (base.repo{owner{}} precedes sha, [^}]* stops early); only affects the line-112 shallow fallback, which refuses exactly as before (no regression), and no self-test row covers it. REFUTED: '|| true' + catch-all (the *) arm prints FAIL). The merge_group.base_sha fix for #5855 is correct and tested. Follow-up wanted: reuse #5858's parse for pull_request + add a row. Not unanimous; arming is a3/cop's call. Not armed by me.
…a real payload (#5859) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…a real payload (#5859) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
|
§13.11 rung 1 — quorum shadow verdict Shadow mode: this records a verdict and merges nothing. A refusal |
Part of #5855, separate from #5858. resolve_base.sh now names merge_group.base_sha (falling back to the PR base sha) so a REBASE-replayed PR is judged whole. Branch work/5855b-resolve-base @2dc1928478, authored by flow; self-test 29/29.
Gate-content change: needs a non-Sonnet quorum + seat and the cop's ANDON before arming. Not armed.
🤖 Generated with Claude Code