[
https://issues.apache.org/jira/browse/HADOOP-9330?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13585415#comment-13585415
]
Steve Loughran commented on HADOOP-9330:
----------------------------------------
@Luke: if that's all we need to do, that's all we need.
# We should be able to put this base test case into hadoop-common test & have
it imported by the other projects -which all import the -test JAR.
# What name & package?
Konstantin: I thought the switch from JUnit3 to JUnit4 was driven by the goal
of adding beforeclass/afterclass methods and so setup/teardown miniclusters
only once per test class, not per test.
The added benefit we get is being able to skip, not just in attributes, but by
raising {{AssumptionViolatedException}}, either from the {{assume()}}, method,
or in your own code. I'm using that in some of my tests already.
> Add custom JUnit4 test runner with configurable timeout
> -------------------------------------------------------
>
> Key: HADOOP-9330
> URL: https://issues.apache.org/jira/browse/HADOOP-9330
> Project: Hadoop Common
> Issue Type: Test
> Components: test
> Affects Versions: 3.0.0
> Reporter: Steve Loughran
>
> HADOOP-9112 has added a requirement for all new test methods to declare a
> timeout, so that jenkins/maven builds will have better information on a
> timeout.
> Hard coding timeouts into tests is dangerous as it will generate spurious
> failures on slower machines/networks and when debugging a test.
> I propose providing a custom JUnit4 test runner that test cases can declare
> as their test runner; this can provide timeouts specified at run-time, rather
> than in-source.
--
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
