[
https://issues.apache.org/jira/browse/JSPWIKI-821?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Juan Pablo Santos RodrÃguez resolved JSPWIKI-821.
-------------------------------------------------
Resolution: Fixed
fixed in 2.10.1-svn-8, was an easy one :-)
> TestEngine sometimes creates testrepositories with double timestamps after
> r1567444
> -----------------------------------------------------------------------------------
>
> Key: JSPWIKI-821
> URL: https://issues.apache.org/jira/browse/JSPWIKI-821
> Project: JSPWiki
> Issue Type: Bug
> Affects Versions: 2.10
> Environment: linux openjdk 1.7.0_51
> Reporter: Brian Burch
> Priority: Minor
> Labels: test
> Original Estimate: 24h
> Remaining Estimate: 24h
>
> r1567444 changed TestEngine.cleanTestProps so that it now concatenates
> System.currentTimeMillis to the property jspwiki.fileSystemProvider.pageDir.
> This is a good idea, because each test then has a unique name for its pageDir
> and so post-mortems of failing tests become easier.
> Unfortunately, this created an obscure problem because some of the methods
> and variables in this class are static, while others are instance, but they
> all play around with the caller's Properties, which might be static or
> instance, and also with its own static combinedProperties variable.
> This leads to several, but a minority, getting double timestamps, eg
> testrepository13923117464781392311746502.
> I discovered this behaviour when holding jspwiki-vers-custom.properties in a
> static variable and also creating a new TestEngine after changing one of the
> properties... my resulting testrepository had at least 12 concatenated
> timestamps - probably one for each test case, plus one more! Since changing
> my class to assign the properties to an instance variable, its testrepository
> only collects two timestamps.
> This weirdness does not prevent any of the tests running successfully because
> the directory name is used consistently throughout the suite. However, it is
> confusing when working on a problem that requires examination of the
> testrepository contents. This code should be fixed when someone finds both
> the time and enthusiasm!
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
