initial changes to add local matrix to primitive processor

initial changes to add local matrix to primitive processor

BUG=skia:

Committed: https://skia.googlesource.com/skia/+/290c09b8bbd8d221d363150e2ce87158f4668df0

[joshualitt, 2014-12-18 22:10:56]

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

[joshualitt, 2014-12-18 22:10:57]

The shader with bitmap font thing is strange.  Should I use a different shader
aside from radial gradient?

[joshualitt, 2014-12-18 22:52:29]

On 2014/12/18 22:10:57, joshualitt wrote:
> The shader with bitmap font thing is strange.  Should I use a different shader
> aside from radial gradient?

I have also added equality testing to pathproc and GP, and added a comment on
get_transform_matrix in GLPrograms to explain the strange use of the identity
matrix in that function

[bsalomon, 2014-12-19 14:15:54]

https://codereview.chromium.org/820523002/diff/40001/src/gpu/GrGeometryProces...
File src/gpu/GrGeometryProcessor.h (right):

https://codereview.chromium.org/820523002/diff/40001/src/gpu/GrGeometryProces...
src/gpu/GrGeometryProcessor.h:93: virtual const SkMatrix& localMatrix() const =
0;
Why not just have the base class own a protected matrix init'ed to I and have a
non-virtual getter? Subclasses can set it, but don't have to worry about it if
they don't use it. Looks like you kinda did that on GP anyway. I don't think it
hurts the pathprim case much to hold a matrix but save a vcall.

https://codereview.chromium.org/820523002/diff/40001/src/gpu/GrGeometryProces...
src/gpu/GrGeometryProcessor.h:318: SkMatrix fLocalMatrix;
move to primproc?

https://codereview.chromium.org/820523002/diff/40001/src/gpu/gl/GrGLProgram.cpp
File src/gpu/gl/GrGLProgram.cpp (right):

https://codereview.chromium.org/820523002/diff/40001/src/gpu/gl/GrGLProgram.c...
src/gpu/gl/GrGLProgram.cpp:45: if (localMatrix.cheapEqualTo(SkMatrix::I())) {
There is a .isIdentity() method on SkMatrix that will use the type flags to know
if the matrix is identity, definitely cheaper when the matrix is usually
identity.

[joshualitt, 2014-12-19 16:00:36]

On 2014/12/19 14:15:54, bsalomon wrote:
>
https://codereview.chromium.org/820523002/diff/40001/src/gpu/GrGeometryProces...
> File src/gpu/GrGeometryProcessor.h (right):
> 
>
https://codereview.chromium.org/820523002/diff/40001/src/gpu/GrGeometryProces...
> src/gpu/GrGeometryProcessor.h:93: virtual const SkMatrix& localMatrix() const
=
> 0;
> Why not just have the base class own a protected matrix init'ed to I and have
a
> non-virtual getter? Subclasses can set it, but don't have to worry about it if
> they don't use it. Looks like you kinda did that on GP anyway. I don't think
it
> hurts the pathprim case much to hold a matrix but save a vcall.
> 
>
https://codereview.chromium.org/820523002/diff/40001/src/gpu/GrGeometryProces...
> src/gpu/GrGeometryProcessor.h:318: SkMatrix fLocalMatrix;
> move to primproc?
> 
>
https://codereview.chromium.org/820523002/diff/40001/src/gpu/gl/GrGLProgram.cpp
> File src/gpu/gl/GrGLProgram.cpp (right):
> 
>
https://codereview.chromium.org/820523002/diff/40001/src/gpu/gl/GrGLProgram.c...
> src/gpu/gl/GrGLProgram.cpp:45: if (localMatrix.cheapEqualTo(SkMatrix::I())) {
> There is a .isIdentity() method on SkMatrix that will use the type flags to
know
> if the matrix is identity, definitely cheaper when the matrix is usually
> identity.

feedback incorporated

[bsalomon, 2014-12-19 16:22:23]

lgtm if we remove the virtual

Should we have a BT base class with fUsesLocalCoords?

https://codereview.chromium.org/820523002/diff/80001/src/gpu/GrGeometryProces...
File src/gpu/GrGeometryProcessor.h (right):

https://codereview.chromium.org/820523002/diff/80001/src/gpu/GrGeometryProces...
src/gpu/GrGeometryProcessor.h:95: virtual const SkMatrix& localMatrix() const {
return fLocalMatrix; }
non-virtual?

[joshualitt, 2014-12-19 21:35:15]

The CQ bit was checked by joshualitt@google.com

[commit-bot: I haz the power, 2014-12-19 21:36:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/820523002/140001

[commit-bot: I haz the power, 2014-12-19 21:45:25]

Committed patchset #8 (id:140001) as
https://skia.googlesource.com/skia/+/290c09b8bbd8d221d363150e2ce87158f4668df0