Broken unittests

[Boudewijn Rempt]

On Monday 30 August 2010, Cyrille Berger wrote:
> On Monday 30 August 2010, Boudewijn Rempt wrote:
> > Not sure what's going on here
> > 
> >         124 - krita-plugin-format-xcf_test (Failed)
> > 
> > Looks like our blending modes don't produce exactly the same results as
> > Gimp's...
> > 
> > 
> > The ctlcs tests are "fixed" by marking them as EXPECT_FAIL: apparently 50%
> > opacity is scaled to 128 now instead o 127. I don't know whether that's
> > bad or not...
> Well this is the same problem with the xcf actually. The error is of 1. I am 
> not sure whether it is acceptable or not. I am also unsure where it comes 
> from. Both tests pass on branch/2.2.

Hrm. This sounds pretty serious to me. It would mean that at some point we have started displaying our own files differently as well.

> (on a side note, I would prefer if the 
> test are not marked as EXPECT_FAIL, it is the best way to forget about them 
> alltogether...)

I sort of agree -- but on the other hand, many tests have ten or so subtests, and if one of them fails, people don't notice when more start failing.

-- 
Boudewijn Rempt | http://www.valdyas.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.kde.org/pipermail/kimageshop/attachments/20100830/ce199c3f/attachment-0001.htm