GPU "fix" for thin stroked rects

This CL dims the interior "solid" portion of a stroked rect when the stroke width is < 1.0.

[bsalomon, 2013-06-26 15:47:49]

https://codereview.chromium.org/17845003/diff/1/src/gpu/GrAARectRenderer.cpp
File src/gpu/GrAARectRenderer.cpp (right):

https://codereview.chromium.org/17845003/diff/1/src/gpu/GrAARectRenderer.cpp#...
src/gpu/GrAARectRenderer.cpp:733: SkScalar inset = SkMinScalar(SK_Scalar1,
devOutside.fRight - devInside.fRight);
As in square scale? Have we checked that anywhere?

https://codereview.chromium.org/17845003/diff/1/src/gpu/GrAARectRenderer.cpp#...
src/gpu/GrAARectRenderer.cpp:764: innerColor = scale | (scale << 8) | (scale <<
16) | (scale << 24);
Can we use GrColorParckRGBA()?

https://codereview.chromium.org/17845003/diff/1/src/gpu/GrAARectRenderer.cpp#...
src/gpu/GrAARectRenderer.cpp:769: innerColor =
SkAlphaMulQ(target->getDrawState().getColor(), scale);
maybe just call this unconditionally?

[robertphillips, 2013-06-26 17:33:05]

https://codereview.chromium.org/17845003/diff/1/src/gpu/GrAARectRenderer.cpp
File src/gpu/GrAARectRenderer.cpp (right):

https://codereview.chromium.org/17845003/diff/1/src/gpu/GrAARectRenderer.cpp#...
src/gpu/GrAARectRenderer.cpp:733: SkScalar inset = SkMinScalar(SK_Scalar1,
devOutside.fRight - devInside.fRight);
This code path explicitly expects non-uniform stroking (see handling in
GrAARectRenderer::strokeAARect).

https://codereview.chromium.org/17845003/diff/1/src/gpu/GrAARectRenderer.cpp#...
src/gpu/GrAARectRenderer.cpp:764: innerColor = scale | (scale << 8) | (scale <<
16) | (scale << 24);
On 2013/06/26 15:47:49, bsalomon wrote:
> Can we use GrColorParckRGBA()?

Done.

https://codereview.chromium.org/17845003/diff/1/src/gpu/GrAARectRenderer.cpp#...
src/gpu/GrAARectRenderer.cpp:769: innerColor =
SkAlphaMulQ(target->getDrawState().getColor(), scale);
On 2013/06/26 15:47:49, bsalomon wrote:
> maybe just call this unconditionally?

Done.

[robertphillips, 2013-06-26 19:06:40]

What was the final decision on this?

[bsalomon, 2013-06-26 19:17:45]

On 2013/06/26 19:06:40, robertphillips wrote:
> What was the final decision on this?

lgtm

[robertphillips, 2013-06-28 17:57:18]

committed as r9812