Repository navigation
Conversation
The v* tag ruleset already limits stable releases to admins, so the release-environment approval was a second click by the same person. Removing it also unblocks tag-nightly, which a skipped approve job had skipped on every nightly. The schedule runs once a day and waits for main's CI instead of polling every three hours. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
There was a problem hiding this comment.
All reported issues were addressed across 5 files
Reply with feedback, questions, or to request a fix.
View guided diff | Turn on auto-fix | Re-trigger cubic
Size Report
Startup median (7 runs, lower is better):
|
Release and nightly publishing require green CI on the exact commit, but pushes to main cancelled each other's runs, leaving a tagged commit with no releasable result (v0.21.28's run was cancelled by the next merge). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
This PR is ready at 5c7bc7c. The change reads correctly: the publish chain now runs on a plain success, which is what the "Tag nightly never ran" fix needs. All 29 checks pass, and the PR's own Release dry run covers the Plan and Build routes. The publish, tag-nightly and stable paths cannot run in a PR, so CI does not exercise them, and the tag-nightly fix can only be confirmed by the first scheduled run after merge. I also could not read the live GitHub settings (release environment, tag ruleset, npm-publish policy), so the "ruleset is already admin-only" claim rests on Plan's runtime check. Not blocking, and you can take or leave these: the deleted approve-gate test has no replacement, so a new test in test/ci/release-workflow.test.ts could assert that On the open threads, the cubic-dev-ai P2 on a duplicate nightly after a failed tag-nightly does not apply, so you can resolve it: the run number is in the version string, so it is never republished, and the next scheduled run is the next UTC day. #3621 (comment) Before merge, nothing else is needed. Keep the release environment until v0.21.28 ships, then delete it. |
Summary
Every release run spent seconds deciding and minutes queued: Plan executes in ~12s, while recent schedule runs waited 30s–31min for an
ubuntu-latestrunner. Seven of the eight daily cron runs only reported "shipped today".v*tag ruleset (creation/update/deletion, org-admin bypass only) already gates stable releases, and Plan still refuses to publish if that ruleset or thenpm-publishbranch policy weakens. Thereleaseenvironment approval was a second click by the same admin. Drops theapprovejob and the reviewer check.Tag nightly, which never ran: a skippedapproveskipped every job downstream ofpublishon nightlies, so nonightly/*tag exists.23 1 * * *). Plan waits up to 20 minutes for main's HEAD CI, as tags already did; failed, cancelled, or still-running CI skips that day. Run Release onmainto publish one by hand. Removes the same-day npm lookup.ci.ymlkeyed its concurrency group by ref, so each main push cancelled the previous commit's run; v0.21.28's Plan failed when the next merge cancelled CI on its tag. Main runs are now grouped by SHA; pull requests still cancel superseded runs. Costs more runner time on busy days.Docs:
CONTRIBUTING.mdrelease channels and one-time setup. 6 files.Keep the
releaseenvironment until v0.21.28 ships: its tag predates this change, so any retry on it still runs the approval job. Delete the environment afterwards.Validation
At e60e96f and 8ba6f31:
pnpm check:affected --runpassed;scripts/__tests__/release-plan.test.tsandtest/ci/release-workflow.test.tspass (13 tests). A fresh-context review found the skip wording too narrow (fixed) and a dead test input (fixed). The PR's own Release dry run exercises the workflow.🤖 Generated with Claude Code