Github user rednaxelafx commented on the issue:
https://github.com/apache/spark/pull/20419
LGTM, and +1 on @viirya 's idea. I like it better for the comment to be on
top of the class declaration instead of inside it; but I'm okay either way if
others have strong opinion otherwise. As long as the comment line is right next
to the class declaration line, that's good enough for easy pattern matching.
BTW, to save a few characters from the generated code, maybe we don't need
to generate this comment when the `spark.sql.codegen.useIdInClassName` is
turned on, since the `codegenStageId` is in the class name already.
But for the ease of having a unified way to access the ID from the comment,
I like having this comment always there. I guess @kiszk and @viirya are on the
same side on this one for always having the comment, right? In that case I'd
agree and let's get this in :-)
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
