Remove getter of writable GrEffectStage from GrDrawState.

Remove getter of writable GrEffectStage from GrDrawState.

Upcoming changes will require GrDrawState to know things about the set of installed effects. Thus all setting of effects must go through a GrDrawState function (setEffect()). This change accomplishes that.
Committed: https://code.google.com/p/skia/source/detail?r=7411

[bsalomon, 2013-01-25 16:25:34]

All of the GrEffectStage::reset() calls were no longer necessary. They were
there from a time when lots of per-stage state was not in GrEffect (and really
from before there was GrEffect at all).

[JimVV, 2013-01-25 17:39:51]

I don't know enough to say it looks good, but it makes sense.

What replaces all of the drawState->stage(x)->reset() calls? 

On 2013/01/25 16:25:34, bsalomon wrote:
> All of the GrEffectStage::reset() calls were no longer necessary. They were
> there from a time when lots of per-stage state was not in GrEffect (and really
> from before there was GrEffect at all).

[bsalomon, 2013-01-25 17:48:23]

On 2013/01/25 17:39:51, JimVV wrote:
> I don't know enough to say it looks good, but it makes sense.
> 
> What replaces all of the drawState->stage(x)->reset() calls? 


Nothing :) We used to have a bunch of state in the class that evolved into
today's GrEffectStage (a matrix, a texture, etc). This was before such things
were migrated to GrEffect subclasses. So it used to be important to reset a
stage before you configured it in case you didn't set all the fields. Now that
stuff is gone and reset() simply sets the effect to NULL. All the callers
removed in the CL are calling reset() just before installing an effect. So the
calls are unnecessary.

There is a function on GrDrawState called disableStage() which will set the
effect to NULL. After this CL it will internally call this->setEffect(i, NULL)
so that all (or most) of our effect state tracking in GrDrawState can live in
setEffect.

[robertphillips, 2013-01-25 18:14:48]

lgtm + nits

https://codereview.appspot.com/7214045/diff/1/src/gpu/GrContext.cpp
File src/gpu/GrContext.cpp (right):

https://codereview.appspot.com/7214045/diff/1/src/gpu/GrContext.cpp#newcode1391
src/gpu/GrContext.cpp:1391: effect.reset(this->createPMToUPMEffect(src,
swapRAndB, textureMatrix));
NULL != effect?

https://codereview.appspot.com/7214045/diff/1/src/gpu/GrContext.cpp#newcode1397
src/gpu/GrContext.cpp:1397: // there is no longer any point to using the
scratch.
NULL != effect?

https://codereview.appspot.com/7214045/diff/1/src/gpu/GrContext.cpp#newcode1597
src/gpu/GrContext.cpp:1597: effect.reset(this->createUPMToPMEffect(texture,
swapRAndB, textureMatrix));
NULL == effect?

[bsalomon, 2013-01-28 14:26:16]

https://codereview.appspot.com/7214045/diff/1/src/gpu/GrContext.cpp
File src/gpu/GrContext.cpp (right):

https://codereview.appspot.com/7214045/diff/1/src/gpu/GrContext.cpp#newcode1391
src/gpu/GrContext.cpp:1391: effect.reset(this->createPMToUPMEffect(src,
swapRAndB, textureMatrix));
On 2013/01/25 18:14:48, robertphillips wrote:
> NULL != effect?

Done.

https://codereview.appspot.com/7214045/diff/1/src/gpu/GrContext.cpp#newcode1397
src/gpu/GrContext.cpp:1397: // there is no longer any point to using the
scratch.
On 2013/01/25 18:14:48, robertphillips wrote:
> NULL != effect?

Done.

https://codereview.appspot.com/7214045/diff/1/src/gpu/GrContext.cpp#newcode1597
src/gpu/GrContext.cpp:1597: effect.reset(this->createUPMToPMEffect(texture,
swapRAndB, textureMatrix));
On 2013/01/25 18:14:48, robertphillips wrote:
> NULL == effect?

Done.