The algorithm I came up with is O(n^2) but given the small numbers of rects
we're typically working with, it doesn't really matter. May need to revisit
this in the future if we find ourselves with a huge number of rects.
Previously we'd blit every pixel in every window that intersected any dirty
rect to the back buffer. With this patch, we limit ourselves to blitting the
pixels inside the actual dirty rects.
There's still a lot of optimizations to make in this code.
