Inline fetch functions in pixman-transformed.c
authorSøren Sandmann Pedersen <sandmann@daimi.au.dk>
Thu, 18 Sep 2008 23:28:31 +0000 (19:28 -0400)
committerSøren Sandmann Pedersen <sandmann@daimi.au.dk>
Thu, 18 Sep 2008 23:28:31 +0000 (19:28 -0400)
commitb5a9002d2e3f0d961bcfe2acee6e9a0f1c583d5d
treee12acc92f616544afe30ca725ab31742ad647aa2
parenta57e7bf34a2f312285c8065b8b1328bd5650788a
Inline fetch functions in pixman-transformed.c

There was significant performance overhead associated with indirect
call to the fetch functions, so inline them.

Also, the old code assumed that if the number of rectangles in the
source clip was 1, then the clip was identical to the image
boundaries. Fix that by running a full region_contains_point()
whenever the source clip is different from pict->common.full_region.

Based on a patch from Antoine Azar.
pixman/pixman-region.c
pixman/pixman-transformed.c