Fix repeated point quads/cubics in convex pr and update convexpaths GM

Fix repeated point quads/cubics in convex pr and update convexpaths GM

Committed: http://code.google.com/p/skia/source/detail?r=10744

[bsalomon, 2013-08-13 15:19:31]

Remove stale comment

[bsalomon, 2013-08-13 15:21:21]

This addresses http://crbug.com/124357

https://codereview.chromium.org/23034003/diff/1/gm/convexpaths.cpp
File gm/convexpaths.cpp (left):

https://codereview.chromium.org/23034003/diff/1/gm/convexpaths.cpp#oldcode51
gm/convexpaths.cpp:51: // CW
When the convex PR was first written it was useful to have all these CW and CCW
paths. After this change there are still paths that wind each way, just not
paths that are identical other than their winding.

https://codereview.chromium.org/23034003/diff/1/gm/convexpaths.cpp#oldcode147
gm/convexpaths.cpp:147: //It turns out arcTos are not automatically marked as
convex and they
no longer true

[robertphillips, 2013-08-13 16:29:55]

lgtm + nits & questions

https://codereview.chromium.org/23034003/diff/2001/gm/convexpaths.cpp
File gm/convexpaths.cpp (right):

https://codereview.chromium.org/23034003/diff/2001/gm/convexpaths.cpp#newcode189
gm/convexpaths.cpp:189: 
2x?

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
File src/gpu/GrAAConvexPathRenderer.cpp (right):

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:222: SegmentArray* segments,
rm degenerateTestData param?

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:230: SegmentArray* segments,
rm degenerateTestData param?

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:231: DegenerateTestData* degenerateTestData)
{
Do we care if pts[0] == pts[2]?

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:233: SkPoint linePts[2] = {pts[0], pts[2]};
Couldn't we just pass &pts[1]?

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:245: SegmentArray* segments,
rm degenerateTestData?

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:284: case SkPath::kLine_Verb: {
Do we really need to map pts[0] here?

[bsalomon, 2013-08-13 16:55:29]

Update for Rob's comments

[bsalomon, 2013-08-13 16:55:56]

https://codereview.chromium.org/23034003/diff/2001/gm/convexpaths.cpp
File gm/convexpaths.cpp (right):

https://codereview.chromium.org/23034003/diff/2001/gm/convexpaths.cpp#newcode189
gm/convexpaths.cpp:189: 
On 2013/08/13 16:29:55, robertphillips wrote:
> 2x?

Done.

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
File src/gpu/GrAAConvexPathRenderer.cpp (right):

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:222: SegmentArray* segments,
On 2013/08/13 16:29:55, robertphillips wrote:
> rm degenerateTestData param?

right, thanks. I was originally doing the deg. check in here but that didn't
work.

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:230: SegmentArray* segments,
On 2013/08/13 16:29:55, robertphillips wrote:
> rm degenerateTestData param?

Done.

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:231: DegenerateTestData* degenerateTestData)
{
On 2013/08/13 16:29:55, robertphillips wrote:
> Do we care if pts[0] == pts[2]?

Done. (made it skip if so and added test case to the GM)

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:233: SkPoint linePts[2] = {pts[0], pts[2]};
On 2013/08/13 16:29:55, robertphillips wrote:
> Couldn't we just pass &pts[1]?

Yes. I originally had it so that we didn't transform or pass the first pt for
lines but changed it for consistency. I put it back.

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:245: SegmentArray* segments,
On 2013/08/13 16:29:55, robertphillips wrote:
> rm degenerateTestData?

Done.

https://codereview.chromium.org/23034003/diff/2001/src/gpu/GrAAConvexPathRend...
src/gpu/GrAAConvexPathRenderer.cpp:284: case SkPath::kLine_Verb: {
On 2013/08/13 16:29:55, robertphillips wrote:
> Do we really need to map pts[0] here?

Done.

[jvanverth1, 2013-08-13 18:53:04]

lgtm

[commit-bot: I haz the power, 2013-08-15 14:13:28]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/bsalomon@google.com/23034003/8001

[commit-bot: I haz the power, 2013-08-15 18:16:35]

Change committed as 10744