Github user srowen commented on the pull request:
https://github.com/apache/spark/pull/5550#issuecomment-95133295
I like that you've introduced a set for skipped stages, since there is also
a problem with the same stage being skipped many times! But can
`numSkippedStages` now be removed?
It still feels funny to count a stage as completed and skipped at the same
time. How about just removing a stage ID from the set of completed stages in
`onStageSubmitted`? if a stage is being run (again), it can no longer be called
"completed", I think?
Then you don't need a new `diff` method or to subtract out the skipped
stages. I think the set is still probably correct for skipped stages since a
stage could conceivably completed, restart, and be skipped multiple times (?)
CC @JoshRosen to make sure we're not going off track here, as I think you
added the skipped-stage logic here.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
