Repository navigation
Conversation
…hanges The changelog/check job failed on PR #6118 with "This pull request modifies CHANGELOG.md directly" although the branch never touches CHANGELOG.md. Root cause: on pull_request events actions/checkout checks out GitHub's test-merge commit, whose tree carries the base branch's current CHANGELOG.md, and github.event.pull_request.base.sha can lag the base branch tip by days (observed: the event carried a 2026-10-02 base SHA while the run merged into a 2026-10-07 master tip). The script diffed from the recorded base SHA to the merge checkout, so the 2.44.0 release regenerating CHANGELOG.md on master was reported as a PR change. Any PR open across a release is falsely rejected, and the branch has no remedy: nothing on the branch touches CHANGELOG.md. Fix: pass the base ref and the PR head SHA from the workflow and diff origin/<base_ref>...<PR head SHA>, which counts only the PR's own changes (the merge base of base tip and PR head is the PR's fork point) and stays correct when the branch merges master after a release, when the base SHA is stale, and when CI checks out the PR head because the test merge conflicts. Task: extract the changelog fix from PR #6118 into its own draft pull request, per user request, so the CI fix lands independently of the coordinated-sampling feature. The fix itself was authored on PR #6118 to fix its failing changelog/check job. Passing the base ref and PR head SHA from the workflow instead of relying on the test-merge commit's parent order is my design decision. No changelog fragment: internal CI tooling. Verified when authored on PR #6118 by reproducing the CI state locally (test merge of the PR head into the current master tip, with the run's recorded base SHA and branch name): the old script exits 1 (reproduces the failure), the fixed script exits 0. A simulated branch that modifies CHANGELOG.md is still rejected; a bump_to_version_* branch is rejected when it does not modify CHANGELOG.md and accepted when it does; a changelog_fix_* branch is accepted; a branch that merged post-release master is accepted; missing BASE_REF or PR_HEAD_SHA errors. Re-verified on the current master tip: the fix applies cleanly (cherry-picked from PR #6118's casual-correlation branch), shellcheck and yamllint clean. actionlint is not installed and is absent from the Debian package lists, so it was not run.
p-datadog
added
the
AI Generated
|
✅ All CI checks and tests passed. 🎉 All green!🧪 All tests passed 🎯 Code Coverage (details) 🔗 Commit SHA: 4365104 | Docs | View more details | Give us feedback! |
|
@codex review |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
🟢 Approval recommended
The ref selection and full-history checkout correctly support the intended three-dot diff.
0 open findings
What changed in this PR
Fixes false-positive changelog checks caused by stale pull-request base SHAs.
Changes:
- Passes the base ref and actual PR head SHA into the check.
- Diffs the PR head against the current base branch merge-base.
| File | Description |
|---|---|
.github/workflows/check.yml |
Supplies accurate PR refs to the script. |
.github/scripts/check_changelog_release_only.sh |
Restricts detection to PR-authored changes. |
🧠 Review effort: Balanced
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
|
Codex Review: Didn't find any major issues. 🚀 Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
🛡️ Codex Security Review · Automatically triggeredSecurity review completed. No security issues were found in this pull request. Reviewed commit: Only the user who started this review can view the report in Codex. ℹ️ About Codex security reviews in GitHubThis is an experimental Codex feature. Security reviews are triggered when:
Once complete, Codex will leave suggestions, or a comment if no findings are found. |
|
@codex review |
|
Codex Review: Didn't find any major issues. What shall we delve into next? Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
🛡️ Codex Security Review · Automatically triggeredSecurity review completed. No security issues were found in this pull request. Reviewed commit: Only the user who started this review can view the report in Codex. ℹ️ About Codex security reviews in GitHubThis is an experimental Codex feature. Security reviews are triggered when:
Once complete, Codex will leave suggestions, or a comment if no findings are found. |
User request in session: delete the comment block added by this branch in .github/scripts/check_changelog_release_only.sh explaining why the changed-files diff is computed against the current base tip instead of the recorded base SHA. No functional change. All changes in this commit are user requirements from session input.
What does this PR do?
Fixes the changelog/check CI job. The check previously diffed the pull request event's recorded base SHA against the test-merge checkout, which attributes base-branch
CHANGELOG.mdedits to the pull request whenever the recorded base SHA lags the base tip or a release regeneratedCHANGELOG.mdon master. The script now receives the base ref and the pull request head SHA from the workflow and diffsorigin/<base_ref>...<PR head SHA>, so only changes made by the pull request itself are counted.Motivation:
The changelog/check job on PR #6118 failed with "This pull request modifies CHANGELOG.md directly" although that branch never touches
CHANGELOG.md:github.event.pull_request.base.shacarried a 2026-10-02 base SHA while the run merged into the 2026-10-07 master tip, so the 2.44.0 release regeneratingCHANGELOG.mdon master was reported as a pull request change. Any pull request open across a release is falsely rejected, and the branch has no way to remedy it.Change log entry
No. Internal CI fix.
Additional Notes:
How to test the change?