Github user aarondav commented on the pull request:
https://github.com/apache/spark/pull/7028#issuecomment-115526307
I've updated the code to move the injection to runJob. This means that
exceptions thrown from any path inside the DAGScheduler that fail the job will
now have their stack trace supplemented, if the job was submitted via runJob
(i.e., not an approximate job or submitted via SparkContext.submitJob).
After looking over the usages of special call site, I found that streaming
is a significant user and thus opted to keep the callsite parsing code. I could
definitely see the benefit of removing the 100 lines of code or so that's
required to do this for the sake of simplicity at the cost of streaming stack
traces not being quite as helpful as they could be, so feel free to push back
and I can change it.
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
