[prev in list] [next in list] [prev in thread] [next in thread] 

List:       koffice-devel
Subject:    Re: unit tests.
From:       Cyrille Berger <cberger () cberger ! net>
Date:       2008-07-12 10:01:11
Message-ID: 200807121201.11891.cberger () cberger ! net
[Download RAW message or body]

On Saturday 12 July 2008, Thomas Zander wrote:
> On Monday 30. June 2008 10:04:10 Boudewijn Rempt wrote:
> > > my idea is to have
> > > * javascript hide/show code in there so you initially see nothing until
> > > you click to see more.
>
> Done (still on http://www.koffice.org/testresults/status.html)
>
> I added regression detection (mark a red star if its a test thats been
> passing before but now fails), the compile run.
> And the tests actually run every night.
>
> Relevant to the results, my script removes the build dir and install dir
> before updating/recompiling.  So its always a clean build.
> Also, I use kdesupport, kdelibs and kdepimlibs from svn trunk.
>
> > > * have an rss feed with a path, so you can subscribe to only krita
> > > failures.
> >
> > Sounds wonderful! Especially an rss feed would be great.
>
> Done add this to your akregator;
> http://www.koffice.org/testresults/testresults.php?filter=krita
>
> the filter takes as an argument the path.  If the filter is found in the
> path to the test, its a hit. So use "krita" or "koffice/libs/flake" etc. 
> No filter means you see everything.

How the filter things work ? I managed to have a feed for "krita" and one 
for "lib". But I didn't manage to have one for krita's filters (they don't 
appear in "krita" and don't appear with "filters")

> Now, please don't write "thank you", I'd be much happer if the amount of
> tests that fail goes down ;)
About this. What would be helpfull as well is information about the system:  
CPU architecture, version of installed packages, etc... (and also if other 
could submit tests). I say this, because of the kis_tiff_test fail and 
TestConvolutionOpImpl tests, it works fine for me, this makes it impossible 
to fix it for you, but at least with system information I could collect 
information on what differences we have).

Same goes for krita-ui-kis_action_recorder_test (it goes without saying since 
I have relaxed the test so that it doesn't fail... but it should be made 
stricter later when the precision issue has been found and solved).

In fact the only tests that I feel remotely concerned about and that fails for 
me is TestColorConversionSystem.
-- 
Cyrille Berger
_______________________________________________
koffice-devel mailing list
koffice-devel@kde.org
https://mail.kde.org/mailman/listinfo/koffice-devel
[prev in list] [next in list] [prev in thread] [next in thread]