[
https://issues.apache.org/jira/browse/HADOOP-9583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13736912#comment-13736912
]
Kihwal Lee commented on HADOOP-9583:
------------------------------------
I misunderstood what HADOOP-9853 does. After experimenting with it, I realized
it would be helpful if test-patch runs tests in multiple modules in a single
build. If a patch touches two modules and there was a known test timeout in the
first module, the build won't continue to the second module. The author may be
misled to believe that the known failure is the only failure. With HADOOP-9853,
surefire timeout cases can be dealt with, but 1) the script cannot rely on the
exit code for failures and 2) any other types of test process failures are not
covered (e.g. something causing System.exit()).
Regardless of HADOOP-9853, it might be better to have the script also post
something indicating that not all tests were run if something like this happens.
> test-patch gives +1 despite build failure when running tests
> ------------------------------------------------------------
>
> Key: HADOOP-9583
> URL: https://issues.apache.org/jira/browse/HADOOP-9583
> Project: Hadoop Common
> Issue Type: Bug
> Affects Versions: 3.0.0
> Reporter: Jason Lowe
> Assignee: Jason Lowe
> Priority: Critical
> Attachments: HADOOP-9583.another_dummy.patch,
> HADOOP-9583-dummy.patch, HADOOP-9583-dummy.patch,
> HADOOP-9583-dummy-without-changes.patch, HADOOP-9583.patch
>
>
> I've seen a couple of checkins recently where tests have timed out resulting
> in a Maven build failure yet test-patch reports an overall +1 on the patch.
> This is encouraging commits of patches that subsequently break builds.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
