If the next edge has 0 fDX, we should add a SLACK = 1 so two edges
with fX less than 1 pixel apart should be considered close, and
noRealBlitter should be true to force the use of AdditiveBlitter and
the cumulation of alpha. The changed GM will show bleed through if
SLACK is 0.
The artifact without the fix can be seen at:
https://fiddle.skia.org/c/f6912f1af6c14e054f5b5935a93380ea
Bug: skia:
Change-Id: I15f9c3aef25a0357cd11d447e7bf0b4fbac0ce67
Reviewed-on: https://skia-review.googlesource.com/67804
Commit-Queue: Yuqian Li <liyuqian@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>
If not, we sometimes would end up with only one edge for a convex path. That
either triggers SkASSERT(count >= 2) failure in debug build or SIGSEGV in
release build. After the change, we should return 0 edges for such a path
because everything is totally combined.
Note that this change also makes the SkAnalyticEdge's CombineVertical function
behave more similarly to SkEdge's CombineVertical function: SkEdge only
compares fFirstY and fLastY which are integer values, which is equivalent to
setting our tolerance to SK_Fixed1 (our current tolerance is 0x100, 1/256 of
SK_Fixed1). And this is intentional.
BUG=chrome:662914
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2477373002
Review-Url: https://codereview.chromium.org/2477373002
