Report aa bloat in triangulating path renderer
Ops should report HasAABloat::kYes if they use coverage AA approaches.
This was confusing the new GrClipStack into thinking a subpixel path
(that should have been coverage) was non-AA, and in certain transforms
this would then not touch pixel centers and the clip stack would reject
the draw early.
Bug: chromium:1159852
Change-Id: If8ad041af92b915e126ab50e4169c959201c1d95
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/345717
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
diff --git a/src/gpu/ops/GrTriangulatingPathRenderer.cpp b/src/gpu/ops/GrTriangulatingPathRenderer.cpp
index ffaf5d0..993e45e 100644
--- a/src/gpu/ops/GrTriangulatingPathRenderer.cpp
+++ b/src/gpu/ops/GrTriangulatingPathRenderer.cpp
@@ -290,7 +290,7 @@
// include the path bounds.
devBounds.join(SkRect::Make(fDevClipBounds));
}
- this->setBounds(devBounds, HasAABloat::kNo, IsHairline::kNo);
+ this->setBounds(devBounds, HasAABloat(fAntiAlias), IsHairline::kNo);
}
FixedFunctionFlags fixedFunctionFlags() const override { return fHelper.fixedFunctionFlags(); }