Ma77Ball opened a new pull request, #5671:
URL: https://github.com/apache/texera/pull/5671

   ### What changes were proposed in this PR?
   - Gate the `github-action-benchmark` gh-pages publish (`auto-push` / 
`save-data-file`) to the weekly `schedule` run only; PR and push-to-main runs 
no longer write bot commits, they still emit the job summary and uploaded 
artifact. This cuts the gh-pages commit churn from per-PR/per-merge down to 
roughly two commits per week.
   - Add a "Benchmark main baseline in the same runner" step to 
`benchmarks.yml`: a PR run benches the PR head, then re-runs the identical 
trimmed grid against the base `main` commit in the same runner and adds 
`arrow-flight-e2e-main.csv` (plus a base-SHA sidecar) to the artifact. The step 
is fail-soft via a `trap` that restores the PR results and original checkout if 
the main re-run dies.
   - Update `benchmarks-pr-comment.yml` to prefer the same-runner main CSV as 
the comparison baseline (keyed 1:1 against the deterministic PR grid), with 
gh-pages still supplying the 7-day average column and a full fallback when the 
same-runner CSV is absent. The comment note flags when the comparison is 
same-runner.
   ### Any related issues, documentation, discussions?
   Closes: #5670
   ### How was this PR tested?
   - Locally validated both workflows parse: `python3 -c "import yaml; 
yaml.safe_load(open('.github/workflows/benchmarks.yml')); 
yaml.safe_load(open('.github/workflows/benchmarks-pr-comment.yml'))"`, expect 
no error.
   - Locally syntax-checked the embedded github-script in 
`benchmarks-pr-comment.yml` with `node --check`, expect no error.
   - Behavior is CI-only and was not exercised locally. Reviewer: open a test 
PR carrying an amber-integration trigger label (e.g. `ci`), confirm the 
Benchmarks run executes the "Benchmark main baseline in the same runner" step, 
the run artifact contains `arrow-flight-e2e-main.csv`, and the PR comment shows 
the "this same runner" note with a populated main-vs-branch table.
   - Reviewer: confirm a PR run and a merge to main produce no new `gh-pages` 
commit, and that only the weekly scheduled run pushes to `gh-pages`.
   ### Was this PR authored or co-authored using generative AI tooling?
   Co-authored with Claude Opus 4.8 in compliance with ASF


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to