sunchao commented on PR #5557: URL: https://github.com/apache/datafusion-comet/pull/5557#issuecomment-5470480068
Thanks for the careful review. I completed the requested split: - #5560 contains the existing-main dictionary shuffle crash fix, its ownership and real-worker regressions, and the low-cardinality shuffle benchmark. - #5561 contains only the Spark 4.2 worker framing fix and the 4.2 real-worker CI row. - This PR is now an 11-file large-offset-only diff. It also includes the small offset-list cleanup, explicit serializer mode, mixed nested-buffer regression, and shim whitespace cleanup from this review. I converted this PR to draft and marked both fixes as prerequisites. Once #5560 and #5561 merge, I will rebase this branch onto `main`, rerun the integrated validation, and mark it ready again. The current narrow head passes 21/21 focused JVM tests and all 128 Spark 4.0 real-worker cases. -- 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]
