When calculating clip rect/visible rect, since using target information from effect tree, we cannot cache values on clip nodes.
Possible results to cache:
- clip node's clip in certain target effect node space, for calculating clip in target space
- clip node's combined clip in screen space (?) for calculating visible rects
- combining clip in screen space will lead to inflation of clips when there is rotation. This is easier to implement the cache, though might not be the best way to do calculation.
Comment 1 by weiliangc@chromium.org
, Aug 31 2016