Github user JoshRosen commented on the pull request:
https://github.com/apache/spark/pull/3561#issuecomment-65728956
This sounds good to me, although there's one edge case that I'm curious
about: what if my main JAR is hosted in a shared cluster filesystem and has a
URI like `file://path/to/my/library.jar`? In that case, I think we'll run into
problems when trying to read the main class. That's a problem in the current
code, though, and this patch seems like a strict improvement over the current
behavior.
I wonder how hard it would be to support this for JARs hosted in HDFS,
since that might be useful when submitting jobs under cluster deploy mode. I
suppose we could try downloading the JAR to read its manifest. @andrewor14, any
thoughts 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]
