andygrove opened a new pull request, #4408: URL: https://github.com/apache/datafusion-comet/pull/4408
## Summary - Splits the monolithic `spark_sql_test.yml` (a 7-module x 4-version matrix) into one workflow per Spark version. - Shared job logic lives in a new `spark_sql_test_reusable.yml` invoked via `workflow_call`; each per-version caller (`spark_sql_test_3_4.yml`, `_3_5.yml`, `_4_0.yml`, `_4_1.yml`) is ~70 lines and just supplies `spark-short`, `spark-full`, and `java`. - Spark 3.5 and 4.0 run on PRs and main; Spark 3.4 and 4.1 run on main only. Each version is now its own check in PRs, which makes it easier to spot/re-run a single version. The build-native job runs once per workflow invocation (so up to 4x on a main push) but is fully cached. The commit uses `[skip ci]` so this PR will not consume CI cycles before review. ## Test plan - [ ] actionlint passes on all 5 new workflow files (verified locally with `actionlint -color --shellcheck=off`). - [ ] After merge, observe that the Spark 3.5 and Spark 4.0 workflows appear as checks on subsequent PRs. - [ ] After merge, observe that all 4 workflows run on the next eligible push to `main`. - [ ] Manually dispatch each per-version workflow with `collect-fallback-logs=true` and confirm the merged artifact uploads with the version-namespaced name. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
