unittests

[Boudewijn Rempt]

Hi,

I've tried to clean up the unittests. This has resulted in two new bugs that replace failing tests:

https://bugs.kde.org/show_bug.cgi?id=248680 (KisTransformWorker -- Marc is looking into it)
https://bugs.kde.org/show_bug.cgi?id=248684 (convolving on a subset of channels -- we completely lost this feature, it seems)

Dmitry is working on the KisTiledDataManagerTest, which fails randomly.

Then there are only some tests in the plugins that still fail:


4 - krita-libbrush-KisAutobrushTest (Failed)

apparently something changed in the way we generate a mask, the original result was soft, it's hard now. I wonder whether this is related to the report that our softness is weird. (I also think that we should make that a bug)

9 - krita-plugin-format-jpeg_test (Failed)
14 - krita-plugin-format-xcf_test (Failed)

These look like genuine regressions. The test first crashed because of changes to KoProgressUpdater and that masked the regression, perhaps.

16 - krita-cs-ko_ctl_channel_test (Failed)
17 - krita-cs-ko_rgb_32f_test (Failed)

Either real regressions, or some weirdness in my local installation? Cyrille, can you check these four tests?

-- 
Boudewijn Rempt | http://www.valdyas.org