[
https://issues.apache.org/jira/browse/MAPREDUCE-6225?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14322651#comment-14322651
]
Junping Du commented on MAPREDUCE-6225:
---------------------------------------
Thanks all for good discussion here!
I think mostly likely we don't need this in either of two places because
JobContext.WORKING_DIR being accessed (by internal of hadoop) all via
job.getWorkingDirectory(). So if we don't replace null with something default
here, it could be replaced later when we use it. However, [~ozawa] mentioned a
good reason (for behavior compatibility on JobConf which is public to other
projects, like: Oozie, etc.) to keep at least 1 of them in the whole code flow,
so I think the question now is which one should be kept here. Now, I agree that
copyAndConfigureFiles() is better because we don't care which one get called
first, but we need to make sure it get called - the other one
(addLog4jToDistributedCache) is under a condition so not proper.
[~ozawa] and [~ajisakaa], any other thoughts here?
> Fix new findbug warnings in hadoop-mapreduce-client-core
> --------------------------------------------------------
>
> Key: MAPREDUCE-6225
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6225
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Affects Versions: 2.7.0
> Reporter: Jason Lowe
> Assignee: Varun Saxena
> Attachments: MAPREDUCE-6225.001.patch, MAPREDUCE-6225.002.patch
>
>
> Recent precommit builds in hadoop-mapreduce-client-core are flagging findbug
> warnings that appear to be new with the recent findbugs upgrade. These need
> to be cleaned up.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
