[
https://issues.apache.org/jira/browse/HADOOP-14686?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16101829#comment-16101829
]
Sean Busbey commented on HADOOP-14686:
--------------------------------------
the patch will fail until the updated .gitignore is in place because of how
Hadoop's use of precommit works.
Essentially, the {{patchprocess}} directory is where Yetus Test Patch keeps its
working files, like the output logs it writes to. On the Hadoop project, we
keep that directory inside of the working directory of the source checkout (due
to legacy reasons that [~aw] has mentioned a few times). As a start-up step,
Yetus cleans up the local git working directory to ensure that it is testing
the patch against the then-current version of whatever relevant branch the
patch is for.
On most hadoop branches, .gitignore ensure that git doesn't mess with the
{{patchprocess}} directory when this cleanup is happening. On branch-2.7, the
ignore directive is missing and so the "clean up git" step wipes out the
in-progress log files that Yetus expects to keep writing to. This leads to the
failures you've seen.
I'm happy to update my patch for whatever review feedback you have, but a blind
copy seems like the fastest way to builds working again. {{.gitignore}}ing
files that don't exist and won't be created by the build process on a
particular branch does no harm, after all.
> Branch-2.7 .gitignore is out of date
> ------------------------------------
>
> Key: HADOOP-14686
> URL: https://issues.apache.org/jira/browse/HADOOP-14686
> Project: Hadoop Common
> Issue Type: Bug
> Components: build, precommit
> Affects Versions: 2.7.4
> Reporter: Sean Busbey
> Assignee: Sean Busbey
> Priority: Blocker
> Attachments: HADOOP-14686-branch-2.7.v0.patch
>
>
> .gitignore is out of date on branch-2.7, which is causing issues in precommit
> checks for that branch.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org
