[
https://issues.apache.org/jira/browse/HBASE-7582?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13555273#comment-13555273
]
nkeywal commented on HBASE-7582:
--------------------------------
I don't think it's immediately available.
As far as I know, Surefire has a random order execution, but my understanding
is that it's on a per class basis, so it's not what we need.
Withing JUnit, it's likely possible to do what we want by writing a custom
runner, but it's not available in the core product (it may have been partly
implemented, but I'm not sure the seed would be there, even if it is a good
point).
The closest stuff I have, from what I understood but not tested, is to use
MethodSorters.JVM on a JDK 1.7: it will be random but without seed.
> Unit test execution mode that randomizes test order
> ---------------------------------------------------
>
> Key: HBASE-7582
> URL: https://issues.apache.org/jira/browse/HBASE-7582
> Project: HBase
> Issue Type: Improvement
> Components: test
> Reporter: Andrew Purtell
> Priority: Minor
>
> I've also come around to agreeing with Jon Hseih that we should have a test
> mode that somehow randomizes test order to catch issues where unit test cases
> are written with the assumption of a certain order. It should not be on by
> default but can be enabled when evaluating patches for commit that include
> new tests.
> Case in point is HBASE-7581, which seems to be due to a new test added in
> HBASE-5498. The patch on HBASE-5498 tested out all green in local tests prior
> to commit.
--
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
