change srcover (and dstover) to use the faster thought slightly less accurate
computation of the scale factor (+1 instead of SkAlpha255To256). This has the
upside of guaranteeing that if the dst is opaque (alpha==255), it will always
stay opaque.



git-svn-id: http://skia.googlecode.com/svn/trunk@104 2bbb7eff-a529-9590-31e7-b0007b416f81
1 file changed