promote.yml carries the stable feed to main through a PR, and fails loudly when it cannot (nxf 6j6v.exga) - #3
Merged
Conversation
…oudly when it cannot (nxf 6j6v.exga) main's ruleset allows no direct push, not even for GitHub Actions (owner rule: every change through a pull request). promote.yml still pushed the regenerated release-notes.json straight to the line, and a rejected push only warned: the promotion stayed green while main's feed and the public changelog silently lacked the stable entry. The feed commit now goes to release-notes/promote-v<v>, and the job opens a PR against the line (skip-changelog label). A workflow-opened PR starts no workflows, so the owner closes and re-opens it to run the required checks, then merges. The owner chose this over a GitHub App token. The job summary prints the commands. The merge is a human push touching release-notes.json, which fires publish-content.yml by itself, so the explicit dispatch (and actions: write) is gone. Failing to open the PR fails the job. Spec §5.5 and the release runbook, the backport runbook step 6 and the release-cut-runbook memory describe the new step. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NhwrYjEtdnS3jCneb57X3k
cabcookie
added
the
skip-changelog
cabcookie
added a commit
that referenced
this pull request
Sep 27, 2026
The promotion itself ran (promote.yml run 36309436401: S3 copy, stable manifest, sidecars verified), but the release event runs the workflow as of the tag, and v0.200.0 predates PR #3: its best-effort step pushed the feed straight to main and the ruleset declined it. This is that same entry, regenerated locally with `cargo xtask changelog promote 0.200.0` (same 17+/1- diff as the rejected commit). Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Merged
cabcookie
added a commit
that referenced
this pull request
Oct 2, 2026
A re-ship with no user impact: since v0.200.0 only CI and the changelog feed changed (#2-#6). It is the first release through staged npm publishing (6j6v.c7dd) and the first tag whose promote.yml opens the stable feed PR (#3). Claude-Session: https://claude.ai/code/session_018Sipejn8YQhKgngxeHpMhT Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Why
Since 2026-09-24,
maincarries ruleset 23939298: every change through a pull request, no bypass, GitHub Actions included.promote.ymlstill pushed the regeneratedrelease-notes.jsonstraight to the line, and a rejected push only produced::warning::. The first stable promotion from the public repo would have stayed green whilemain's feed and the public changelog silently lacked the stable entry.What
release-notes/promote-v<version>, and the job opens a PR against the line (mainorrelease/x.Y), labelledskip-changelog. A re-run force-updates its own branch and reuses an open PR.GITHUB_TOKENstarts no workflows, so its required checks would never report. The owner closes and re-opens it, waits for the checks and merges. The job summary prints the threeghcommands. The alternative, a GitHub App token, was declined: a stable promotion is already a deliberate human gesture, and the app would add a secret to maintain.continue-on-error. If the branch push or the PR fails, the job goes red; the promotion itself has landed by then.publish-contentdispatch (and noactions: write). It would now run before the merge. The owner's merge is a human push touchingrelease-notes.json, which firespublish-content.ymlby itself.release-cut-runbookmemory (NEXUS_MEMORY.md).Evidence
actionlintis clean. The step itself only runs on a real stable promotion, the first being 0.200.0 → stable.promote-ci.ymlcoverspromote-release.sh, which this diff does not touch.skip-changelog:promote.ymlis a carved-out delivery file, but nothing a user installs changes.🤖 Generated with Claude Code
https://claude.ai/code/session_01NhwrYjEtdnS3jCneb57X3k