[jbosstools-dev] Test failures/timeout are very bad

Snjezana Peco snjezana.peco at redhat.com
Wed May 22 16:42:41 EDT 2013


We would need to apply 
https://github.com/jbosstools/jbosstools-build/pull/88. See 
https://bugs.eclipse.org/bugs/show_bug.cgi?id=405456
The test failure at 
https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/JBossTools/view/JBossTools_Trunk/job/jbosstools-base_master/260/testReport/org.jboss.tools.common.model.test/ClassPathTest/testSharingJarContent/ 
is caused by this bug.

Snjeza

On 5/21/2013 2:42 PM, Max Rydahl Andersen wrote:
> On Tue, May 21, 2013 at 10:46:24AM +0200, Mickael Istria wrote:
>
> Mickael - problem is that these tests passes fine on a local machine.
>
> Fails on jenkins which unfortunately keeps being extremly unreliable and slow.
>
> But Yes - we need to fix these and adjust the tests to not be reliant
> on multiple jobs/threads etc. so the tests can run to completeness more reliable.
>
> thats the worst thing I see in our testsuite, the tests are relying on concurrent jobs
> and aren't simply testing a unit of code. it's testing a combination (i.e. big integration test)
> and that's not something we should have in our core tests - thats too unreliable.
>
> /max
>
>> Hi all,
>>
>> On 05/18/2013 08:53 PM, Nick Boldt wrote:
>>> base: test fail
>> Base (which is used by almost all our component) has been red for more
>> than a week because of test failures or timeout. It has even been
>> branched for a revision that has been red for a week.
>> What does that mean? It means that we will create a Beta1 based on a
>> component something that either has a bug, or that we don't know
>> whether it works or not. If I were Len or anyone else from the QE
>> team, I would be really annoyed by the fact that QE is about to be
>> shipped a build where base component is not guaranteed to work...
>>
>> This is a general bad habit we have, as a team. regarding automated
>> tests. Automated tests, on whichever component, are here to ensure the
>> quality of what we're doing, and to avoid respins or worse, bugs on
>> end-usage.
>> These failed tests have to take priority #1 when they happen, and you
>> should monitor mails from CI or at the CI dashboard to check
>> everything you're working on seems to behave well. Leaving jobs into a
>> Red state is just taking a huge risk, which can have pretty bad
>> consequences (respins) and that will cost much more time and effort to
>> fix later.
>>
>> So please try to keep the CI jobs blue. It's our responsibility to all
>> of us to do good stuff, and to ensure we're doing good stuff.
>> Cheers,
>> -- 
>> Mickael Istria
>> Eclipse developer at JBoss, by Red Hat <http://www.jboss.org/tools>
>> My blog <http://mickaelistria.wordpress.com> - My Tweets
>> <http://twitter.com/mickaelistria>
>> _______________________________________________
>> jbosstools-dev mailing list
>> jbosstools-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jbosstools-dev
> _______________________________________________
> jbosstools-dev mailing list
> jbosstools-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jbosstools-dev



More information about the jbosstools-dev mailing list