Thorsten Zachmann's report of two current problems
michael.thaler at physik.tu-muenchen.de
Thu Mar 2 18:59:11 CET 2006
On Thursday 02 March 2006 08:24, Boudewijn Rempt wrote:
> This is known; shear uses the old rotate code, which isn't cs independent.
> Casper is working on getting the new generic transform code ready. I'm not
> sure about his progress, though.
I just had a look at the shear code. It uses mixColors() everywhere, so it
should be colorspace independent.
For me, rotating doesn't work even for RGB8 images. The rotated image just
looks terrible (it has lots of stripes).
Maybe something is wrong with the mixColor() function?
More information about the kimageshop