viirya commented on PR #58978: URL: https://github.com/apache/spark/pull/58978#issuecomment-5841281008
Thanks for the detailed review. I've pushed c93a0cdad6a with the fixes for this round. In-process execution now has a dedicated physical node while retaining the existing PythonUDF/ArrowEvalPython logical planning. The shared Arrow type mapping is restored, with binary widening confined to the in-process bridge. Local validation passed: the Hive-enabled package build, 41 Scala tests, 346 Python tests, and the installed pyspark-client Connect plan suite (82 passed, one expected skip). Mypy, Ruff, and Scala style checks also passed. The bootstrap tests now start without JEP or PySpark on PYTHONPATH and verify that Spark's pyspark.zip takes precedence over site-packages. I haven't run the YARN/Kubernetes examples on a cluster or reproduced the proposed stack-overflow scenario; the larger stack request is a defensive change. -- 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]
