MaxGekk commented on PR #59018: URL: https://github.com/apache/spark/pull/59018#issuecomment-5825323603
Thank you for the approval and the nits, @dongjoon-hyun. All four are addressed in the follow-up commit: - The empty line before the closing brace is removed. - **The walk is shared now.** Rather than going back to a `HashSet` next to a second copy, I moved the walk into one helper, `debug.codegenSubtrees` (`private[sql]`), which returns the subtrees sorted by stage id. Both `codegenStringSeq` and `BenchmarkQueryTest.generatedCode` call it, so they can't drift apart. - `withRowBroadcasts` is now `private`, and the unused default is gone. - **It rewrites and checks only the stage.** The copy stops at `InputAdapter`: a broadcast read right there is replaced, and nothing below it is copied or checked. The assertion message also names the stage id. The six suites plus `DebuggingSuite` pass (642 tests). `WholeStageCodegenSizeBenchmark` gives the same output before and after the change, so the committed results files stand. -- 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]
