Record the red-baseline deadlock in the handoff

ci-baseline-green also runs inside ci-mirror's clean worktree, so a red main
blocks the very commit that fixes it. CI_GREEN_SKIP=1 has to be set on the
push, not just the commit. Cost a failed push to discover.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PJrqKERAo5QUeEHWS5Asgp
This commit is contained in:
Krzysztof Rudnicki 2026-08-23 20:11:09 +02:00
parent ac8e098c03
commit 042acc820b

View File

@ -102,6 +102,12 @@ by md5, DNS verified between them. Backups at
(it measures only the parent); read `memory.peak` from a cgroup.
- **`sudo -n` works here and `guardctl` is at `/usr/local/bin/guardctl`.** The
old "the classifier blocks me" note was stale.
- **A red `main` deadlocks its own fix.** `ci-baseline-green` refuses to build
on a red baseline, and ci-mirror runs `pre-commit` inside a clean worktree,
so the hook fires there too — the commit that FIXES the red baseline cannot
be pushed. `CI_GREEN_SKIP=1` must be set on the **push**, not just the
commit: `CI_GREEN_SKIP=1 git push`. Use it only when the push really is the
fix.
- **The pre-push gate is not the pre-commit gate.** It adds prettier and
ci-mirror. Two docs were non-conforming before this session and blocked
_every_ push; prettier normalises `*emphasis*` to `_emphasis_`.