ajantha-bhat opened a new pull request, #16946: URL: https://github.com/apache/iceberg/pull/16946
depends on https://github.com/apache/iceberg/pull/16945 ### Summary This updates Spark CI to build an incremental PR matrix based on the changed Spark version path. For ordinary PRs, Spark CI now uses the JVM plan from the earlier Java CI PR and then narrows the Spark matrix when changes are scoped to a known Spark version path: - `spark/v3.5/**` runs Spark 3.5 only - `spark/v4.0/**` runs Spark 4.0 only - `spark/v4.1/**` runs Spark 4.1 only - unknown Spark version paths, unknown Spark paths, missing changed-file detection, `full-ci`, and push/main/release/tag runs fall back to the full Spark matrix Known Spark versions are read from `knownSparkVersions` in `gradle.properties`, so adding/removing Spark versions does not require hard-coding them in the planner. This keeps the PR path conservative: stale or unknown planner inputs run more CI, not less. ### Testing - `actionlint .github/workflows/spark-ci.yml` - `bash -n .github/scripts/ci-plan-common.sh .github/scripts/plan-spark-ci.sh` - YAML parse for `jvm-ci-plan.yml` and `spark-ci.yml` - `git diff --check` - Local planner checks: - `spark/v4.1/**` -> Spark 4.1 only - `spark/v3.5/**` -> Spark 3.5 only - unknown Spark version/path -> all Spark versions - `full-ci` / push -> all Spark versions with the full JVM matrix -- 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]
