2010/11/16 Josh Andler <scislac@...400...>:
The Cairo branch is wonderful on many levels. It fixes our seam & pattern issues where gaps and other precision problems have plagued us forever. It's a bit faster and less resource hungry in my limited testing. There are also small other benefits here and there that are like "Yeah, that feels more right" too (feel in a visual sense). However, there are upstream fixes that need to take place for it to even be considered (gradient problems and raster image scaling problems are the big ones as far as I know). To me it's a shame that we can't get more extensive testing in trunk to potentially find other issues, but as things stand, the rule of "don't commit obviously broken stuff" is what prevents that. So yes, lets encourage testing of the branch but know that it just may tease people for now. :)
Cheers, Josh
The gradient precision problems are fixed upstream now. The only remaining big task is converting Pixman to optionally use doubles or floats instead of fixed point to store transformation matrices and gradient coordinates. It's quite a big task unfortunately.
Regards, Krzysztof