New wash-mode painting implementation in 'krita-testing-kazakov'

[Dmitry Kazakov]

Hi!

Last two weeks I was working on fixing the artifacts that were seen in the
tail of the wash-mode strokes. The artifacts are explained in the first
part of the bug [0]. This happened because the our ALPHA_DARKEN composite
op implementation was not designed to handle the case when the opacity
drops gradually. Now the composition formula is reversed when the opacity
value drops.

This fix changes quite much in the heart of Kirta composition code,
including the SSE/AVX optimized code, so it would be really cool if someone
could test the branch before I merge it into master. You just need to check
that Wash Mode painting works as expected, and no new artifacts appeared.
Of course, you can also notice the absence of white gaps in the tail of the
stroke ;)


[0] - https://bugs.kde.org/show_bug.cgi?id=320651

-- 
Dmitry Kazakov
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kimageshop/attachments/20130629/e0670961/attachment.html>