Move VertexBuilder to a GrGLFullShaderBuilder subclass

Move VertexBuilder to a GrGLFullShaderBuilder subclass

Removes the VertexBuilder nested class from GrGLShaderBuilder in favor
of a new GrGLFullShaderBuilder subclass, and adds an optional emitCode
overload to GrGLEffect that takes a GrGLFullShaderBuilder. Makes
setData virtual in GrGLEffectArray and adds a GrGLVertexEffectArray
subclass that gets built using a GrGLFullShaderBuilder. Also adds a
new GrGLVertexEffect subclass that makes the GrGLFullShaderBuilder
overload required for emitCode, and updates GrGLEffects to inherit
from GrGLVertexEffect where needed.

Committed: http://code.google.com/p/skia/source/detail?r=11612

[bsalomon, 2013-10-03 14:44:52]

https://codereview.chromium.org/25474006/diff/1/src/effects/gradients/SkTwoPo...
File src/effects/gradients/SkTwoPointConicalGradient.cpp (right):

https://codereview.chromium.org/25474006/diff/1/src/effects/gradients/SkTwoPo...
src/effects/gradients/SkTwoPointConicalGradient.cpp:345: virtual void
emitCode(GrGLFullShaderBuilder*,
Eek.. do we need both virtuals for emitting code on the same class? Or is this
something that goes away when we use the transform mechanism for vB?

https://codereview.chromium.org/25474006/diff/1/src/gpu/GrAARectRenderer.cpp
File src/gpu/GrAARectRenderer.cpp (right):

https://codereview.chromium.org/25474006/diff/1/src/gpu/GrAARectRenderer.cpp#...
src/gpu/GrAARectRenderer.cpp:176: &vsRectEdgeName, &fsRectEdgeName);
can you realign the params?

https://codereview.chromium.org/25474006/diff/1/src/gpu/gl/GrGLEffect.h
File src/gpu/gl/GrGLEffect.h (right):

https://codereview.chromium.org/25474006/diff/1/src/gpu/gl/GrGLEffect.h#newco...
src/gpu/gl/GrGLEffect.h:126: virtual void emitCode(GrGLShaderBuilder* builder,
Is this a case where it would make sense to redeclare the function as
non-virtual? (Want final or sealed!)

[Chris Dalton, 2013-10-03 21:43:22]

https://codereview.chromium.org/25474006/diff/1/src/effects/gradients/SkTwoPo...
File src/effects/gradients/SkTwoPointConicalGradient.cpp (right):

https://codereview.chromium.org/25474006/diff/1/src/effects/gradients/SkTwoPo...
src/effects/gradients/SkTwoPointConicalGradient.cpp:345: virtual void
emitCode(GrGLFullShaderBuilder*,
On 2013/10/03 14:44:52, bsalomon wrote:
> Eek.. do we need both virtuals for emitting code on the same class? Or is this
> something that goes away when we use the transform mechanism for vB?

Done.

https://codereview.chromium.org/25474006/diff/1/src/gpu/GrAARectRenderer.cpp
File src/gpu/GrAARectRenderer.cpp (right):

https://codereview.chromium.org/25474006/diff/1/src/gpu/GrAARectRenderer.cpp#...
src/gpu/GrAARectRenderer.cpp:176: &vsRectEdgeName, &fsRectEdgeName);
On 2013/10/03 14:44:52, bsalomon wrote:
> can you realign the params?

Done.

https://codereview.chromium.org/25474006/diff/1/src/gpu/gl/GrGLEffect.h
File src/gpu/gl/GrGLEffect.h (right):

https://codereview.chromium.org/25474006/diff/1/src/gpu/gl/GrGLEffect.h#newco...
src/gpu/gl/GrGLEffect.h:126: virtual void emitCode(GrGLShaderBuilder* builder,
On 2013/10/03 14:44:52, bsalomon wrote:
> Is this a case where it would make sense to redeclare the function as
> non-virtual? (Want final or sealed!)

Done. (It will be enough protection to hide the GrGLFullShaderBuilder* variant
from the base class, and make the new GrGLFullShader* version pure virtual in
the derived class.)

[bsalomon, 2013-10-04 15:28:00]

On 2013/10/03 21:43:22, Chris Dalton wrote:
>
https://codereview.chromium.org/25474006/diff/1/src/effects/gradients/SkTwoPo...
> File src/effects/gradients/SkTwoPointConicalGradient.cpp (right):
> 
>
https://codereview.chromium.org/25474006/diff/1/src/effects/gradients/SkTwoPo...
> src/effects/gradients/SkTwoPointConicalGradient.cpp:345: virtual void
> emitCode(GrGLFullShaderBuilder*,
> On 2013/10/03 14:44:52, bsalomon wrote:
> > Eek.. do we need both virtuals for emitting code on the same class? Or is
this
> > something that goes away when we use the transform mechanism for vB?
> 
> Done.
> 
> https://codereview.chromium.org/25474006/diff/1/src/gpu/GrAARectRenderer.cpp
> File src/gpu/GrAARectRenderer.cpp (right):
> 
>
https://codereview.chromium.org/25474006/diff/1/src/gpu/GrAARectRenderer.cpp#...
> src/gpu/GrAARectRenderer.cpp:176: &vsRectEdgeName, &fsRectEdgeName);
> On 2013/10/03 14:44:52, bsalomon wrote:
> > can you realign the params?
> 
> Done.
> 
> https://codereview.chromium.org/25474006/diff/1/src/gpu/gl/GrGLEffect.h
> File src/gpu/gl/GrGLEffect.h (right):
> 
>
https://codereview.chromium.org/25474006/diff/1/src/gpu/gl/GrGLEffect.h#newco...
> src/gpu/gl/GrGLEffect.h:126: virtual void emitCode(GrGLShaderBuilder* builder,
> On 2013/10/03 14:44:52, bsalomon wrote:
> > Is this a case where it would make sense to redeclare the function as
> > non-virtual? (Want final or sealed!)
> 
> Done. (It will be enough protection to hide the GrGLFullShaderBuilder* variant
> from the base class, and make the new GrGLFullShader* version pure virtual in
> the derived class.)

lgtm

[commit-bot: I haz the power, 2013-10-04 15:28:35]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/cdalton@nvidia.com/25474006/8001

[commit-bot: I haz the power, 2013-10-04 15:42:59]

Change committed as 11612