Split GrDrawState and GrOptDrawState into separate classes and remove base class.

Split GrDrawState and GrOptDrawState into separate classes and remove base class.

Besides splitting the two classes, there are no logical changes here and mostly moving code around.

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/89af44a0f1323dc136b238c00ccf143e15abb243

[egdaniel, 2014-09-24 19:20:53]

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

[bsalomon, 2014-09-25 14:29:28]

lgtm w/ a couple nits and Qs

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h
File src/gpu/GrDrawState.h (right):

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h#ne...
src/gpu/GrDrawState.h:97: bool validateVertexAttribs() const;
can this be private? (also what about getFFVAttribIndices?

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h#ne...
src/gpu/GrDrawState.h:259: const GrFragmentStage& getCoverageStage(int stageIdx)
const { return fCoverageStages[stageIdx]; }
line wrap, also must these be public?

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h#ne...
src/gpu/GrDrawState.h:438: // TODO: determine whether we really need to leave
matrix unmodified
Let's remove this comment... It's a common pattern in Skia to not modify out
params when a function fails.

[bsalomon, 2014-09-25 14:29:29]

lgtm w/ a couple nits and Qs

[egdaniel, 2014-09-25 14:48:31]

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h
File src/gpu/GrDrawState.h (right):

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h#ne...
src/gpu/GrDrawState.h:97: bool validateVertexAttribs() const;
The validateVertexAttribs() only use is a call from GrDrawTarget so must be
public.

getFFVAI is called from ODS when creating itself. Theoretically I suppose it is
possible to recreated the FFVAI array in ODS from just the VAArray but this
makes it much simpler.

On 2014/09/25 14:29:28, bsalomon wrote:
> can this be private? (also what about getFFVAttribIndices?

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h#ne...
src/gpu/GrDrawState.h:259: const GrFragmentStage& getCoverageStage(int stageIdx)
const { return fCoverageStages[stageIdx]; }
yes they must be public. They are our only way to get any of the stages from
DrawState

On 2014/09/25 14:29:28, bsalomon wrote:
> line wrap, also must these be public?

[bsalomon, 2014-09-25 15:10:17]

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h
File src/gpu/GrDrawState.h (right):

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h#ne...
src/gpu/GrDrawState.h:259: const GrFragmentStage& getCoverageStage(int stageIdx)
const { return fCoverageStages[stageIdx]; }
On 2014/09/25 14:48:30, egdaniel wrote:
> yes they must be public. They are our only way to get any of the stages from
> DrawState
> 
> On 2014/09/25 14:29:28, bsalomon wrote:
> > line wrap, also must these be public?
> 

I suppose that is in order to make GrODS? At some point we should figure out a
mechanism to have privileged access for GrODS. Maybe friend, or maybe something
narrower.

[egdaniel, 2014-09-25 15:18:30]

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h
File src/gpu/GrDrawState.h (right):

https://codereview.chromium.org/597323002/diff/20001/src/gpu/GrDrawState.h#ne...
src/gpu/GrDrawState.h:259: const GrFragmentStage& getCoverageStage(int stageIdx)
const { return fCoverageStages[stageIdx]; }
Excluding all internal uses, one use is to make the ODS and the other is
DrawTarget uses them in checkDraw() (but only in debug). The methods on ODS
probably need to stay public in the end since that is how gl grabs the stages.

On 2014/09/25 15:10:17, bsalomon wrote:
> On 2014/09/25 14:48:30, egdaniel wrote:
> > yes they must be public. They are our only way to get any of the stages from
> > DrawState
> > 
> > On 2014/09/25 14:29:28, bsalomon wrote:
> > > line wrap, also must these be public?
> > 
> 
> I suppose that is in order to make GrODS? At some point we should figure out a
> mechanism to have privileged access for GrODS. Maybe friend, or maybe
something
> narrower.

[egdaniel, 2014-09-25 19:05:22]

The CQ bit was checked by egdaniel@google.com

[commit-bot: I haz the power, 2014-09-25 19:05:53]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/597323002/60001

[commit-bot: I haz the power, 2014-09-25 19:09:23]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-09-25 19:09:23]

Try jobs failed on following builders:
  Build-Win-VS2013-x86-Debug-Trybot on tryserver.skia
(http://108.170.220.76:10117/builders/Build-Win-VS2013-x86-Debug-Trybot/builds...)

[egdaniel, 2014-09-26 13:09:37]

The CQ bit was checked by egdaniel@google.com

[commit-bot: I haz the power, 2014-09-26 13:09:53]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/597323002/60001

[commit-bot: I haz the power, 2014-09-26 13:15:08]

Committed patchset #4 (id:60001) as 89af44a0f1323dc136b238c00ccf143e15abb243