Fix for fat stroked roundrects.

Fix for fat stroked roundrects.

The original method did not correctly compute the distance value along
the inner edge, so the antialiasing was not quite correct. This fixes
that by adding additional vertices and setting a different outerRadius.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2275423002

Committed: https://skia.googlesource.com/skia/+/6a3976114d8a6aaa564c3c7571cfe7d1727a3a6e

[jvanverth1, 2016-08-25 18:33:55]

Description was changed from

==========
Fix for fat stroked roundrects.

The original method did not correctly compute the distance value along
the inner edge, so the antialiasing was not quite correct. This fixes
that by adding additional vertices and setting a different outerRadius.
==========

to

==========
Fix for fat stroked roundrects.

The original method did not correctly compute the distance value along
the inner edge, so the antialiasing was not quite correct. This fixes
that by adding additional vertices and setting a different outerRadius.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2275423002
==========

[jvanverth1, 2016-08-25 18:34:35]

jvanverth@google.com changed reviewers:
+ egdaniel@google.com

[egdaniel, 2016-08-25 20:24:27]

lgtm with the one question

https://codereview.chromium.org/2275423002/diff/1/src/gpu/GrOvalRenderer.cpp
File src/gpu/GrOvalRenderer.cpp (right):

https://codereview.chromium.org/2275423002/diff/1/src/gpu/GrOvalRenderer.cpp#...
src/gpu/GrOvalRenderer.cpp:1642: verts->fOuterRadius = -args.fInnerRadius;
does this need to be normalized? If so just use -innerRadius?

[jvanverth1, 2016-08-26 14:35:20]

https://codereview.chromium.org/2275423002/diff/1/src/gpu/GrOvalRenderer.cpp
File src/gpu/GrOvalRenderer.cpp (right):

https://codereview.chromium.org/2275423002/diff/1/src/gpu/GrOvalRenderer.cpp#...
src/gpu/GrOvalRenderer.cpp:1642: verts->fOuterRadius = -args.fInnerRadius;
On 2016/08/25 20:24:27, egdaniel wrote:
> does this need to be normalized? If so just use -innerRadius?

No, the math is a little weird. We pass in normalized coordinates for the
offset, subtract the length of that from 1, and then multiply by whatever's sent
for the outerRadius to put the difference into device space.  So t = radius*(1 -
distance/radius), where distance is in device space, so d/r is the normalized
distance.

That said, it used to be just t = radius - distance, and I'm not sure why I
changed it. I have a note to go back and figure out why I made it more
complicated.

[jvanverth1, 2016-08-26 14:45:58]

The CQ bit was checked by jvanverth@google.com

[commit-bot: I haz the power, 2016-08-26 14:46:04]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-08-26 15:15:35]

Description was changed from

==========
Fix for fat stroked roundrects.

The original method did not correctly compute the distance value along
the inner edge, so the antialiasing was not quite correct. This fixes
that by adding additional vertices and setting a different outerRadius.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2275423002
==========

to

==========
Fix for fat stroked roundrects.

The original method did not correctly compute the distance value along
the inner edge, so the antialiasing was not quite correct. This fixes
that by adding additional vertices and setting a different outerRadius.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2275423002

Committed:
https://skia.googlesource.com/skia/+/6a3976114d8a6aaa564c3c7571cfe7d1727a3a6e
==========

[commit-bot: I haz the power, 2016-08-26 15:15:36]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/skia/+/6a3976114d8a6aaa564c3c7571cfe7d1727a3a6e