Remove coordchanges from drawstate

Remove coordchanges from drawstate

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/d27f73ef27ff65a6a0a5d00aa8e5b784b1a0b47e

[joshualitt, 2014-12-23 15:49:03]

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

[joshualitt, 2014-12-23 15:50:59]

Brian, I'm going to clean up the constructor to the GP to move localMatrix
before opaqueVertexColors, but otherwise this is ready for review.  I am going
to start cleaning up after this, ie removing the fragmentStage, and then taking
the VM off drawState and moving it to GP.

[joshualitt, 2014-12-23 18:08:09]

On 2014/12/23 15:50:59, joshualitt wrote:
> Brian, I'm going to clean up the constructor to the GP to move localMatrix
> before opaqueVertexColors, but otherwise this is ready for review.  I am going
> to start cleaning up after this, ie removing the fragmentStage, and then
taking
> the VM off drawState and moving it to GP.

Disregard this comment, I'll shuffle the agrguments when I add the viewmatrix. 
I'm thinking the arguments to the GP base class should look like (color,
viewMatrix, localMatrix)

[bsalomon, 2014-12-23 20:18:40]

overall looks right. A few small comments inline. Everywhere where we can omit
the LM as a param I think that is preferable.

https://codereview.chromium.org/818233002/diff/40001/src/gpu/GrAAHairLinePath...
File src/gpu/GrAAHairLinePathRenderer.cpp (right):

https://codereview.chromium.org/818233002/diff/40001/src/gpu/GrAAHairLinePath...
src/gpu/GrAAHairLinePathRenderer.cpp:917: SkMatrix invert = SkMatrix::I();
Why don't we just do this once outside the if (lineCnt) and if (quadCnt ||
conicCnt) blocks?  I bet the fact that we have both is just a copy/paste
accident.

https://codereview.chromium.org/818233002/diff/40001/src/gpu/GrInOrderDrawBuf...
File src/gpu/GrInOrderDrawBuffer.cpp (right):

https://codereview.chromium.org/818233002/diff/40001/src/gpu/GrInOrderDrawBuf...
src/gpu/GrInOrderDrawBuffer.cpp:128: bool hasExplicitLocalCoords =
SkToBool(localRect);
I don't think this comment is quite right... if we have a lr we still use lm,
but we apply it directly the lr to compute an lc vertex attribute.

[joshualitt, 2014-12-23 20:44:18]

On 2014/12/23 20:18:40, bsalomon wrote:
> overall looks right. A few small comments inline. Everywhere where we can omit
> the LM as a param I think that is preferable.
> 
>
https://codereview.chromium.org/818233002/diff/40001/src/gpu/GrAAHairLinePath...
> File src/gpu/GrAAHairLinePathRenderer.cpp (right):
> 
>
https://codereview.chromium.org/818233002/diff/40001/src/gpu/GrAAHairLinePath...
> src/gpu/GrAAHairLinePathRenderer.cpp:917: SkMatrix invert = SkMatrix::I();
> Why don't we just do this once outside the if (lineCnt) and if (quadCnt ||
> conicCnt) blocks?  I bet the fact that we have both is just a copy/paste
> accident.

done
>
https://codereview.chromium.org/818233002/diff/40001/src/gpu/GrInOrderDrawBuf...
> File src/gpu/GrInOrderDrawBuffer.cpp (right):
> 
>
https://codereview.chromium.org/818233002/diff/40001/src/gpu/GrInOrderDrawBuf...
> src/gpu/GrInOrderDrawBuffer.cpp:128: bool hasExplicitLocalCoords =
> SkToBool(localRect);
> I don't think this comment is quite right... if we have a lr we still use lm,
> but we apply it directly the lr to compute an lc vertex attribute.

done

[bsalomon, 2014-12-29 14:20:58]

lgtm

[joshualitt, 2014-12-29 15:33:45]

The CQ bit was checked by joshualitt@google.com

[commit-bot: I haz the power, 2014-12-29 15:34:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/818233002/80001

[commit-bot: I haz the power, 2014-12-29 15:43:39]

Committed patchset #5 (id:80001) as
https://skia.googlesource.com/skia/+/d27f73ef27ff65a6a0a5d00aa8e5b784b1a0b47e