First pass at stencil clip mask refactoring

This is just the first pass. TODO:

I have left the code in place so that the diff can work at all - I will move it in the next delivery

The handling of the scissor rect is still a bit rough but I want to leave it be for now. I had hoped to eliminate the use of the stack to store the scissor rect but it proved problematic.

There are some rough edges w.r.t. the invalidation of the HW scissoring bound (I would like to invalidate it more often) and the Apple-specific scissoring handling (I would like to unify it will the enable/disable-Scissoring system). I would like to defer both of these changes until this first pass makes it in.

[bsalomon, 2012-04-16 14:31:45]

Overall I like the direction of capturing the clip implementation in its own
class.

http://codereview.appspot.com/6031043/diff/1/src/gpu/GrGpu.h
File src/gpu/GrGpu.h (right):

http://codereview.appspot.com/6031043/diff/1/src/gpu/GrGpu.h#newcode75
src/gpu/GrGpu.h:75: class GrClipMaskCreator {
Would it make sense to make this a private/protected nested class of GrGpu? It
would get private access so we'd avoid exposing some GrGpu stuff in its public
interface and the class it self wouldn't be exposed to includers of GrGpu.h. We
could also forward declare it in this h file and declare the full class in the
cpp (GrGpu.cpp initiallly but eventually its own cpp file).

Since this guy is also doing some tracking of whether the clip is represented in
the stencil buffer he seems to be doing a little bit more than Creating. Should
we call it a Manager or something else?

http://codereview.appspot.com/6031043/diff/1/src/gpu/gl/GrGpuGL.cpp
File src/gpu/gl/GrGpuGL.cpp (right):

http://codereview.appspot.com/6031043/diff/1/src/gpu/gl/GrGpuGL.cpp#newcode1748
src/gpu/gl/GrGpuGL.cpp:1748: this->enableScissoring(dirtyRect);
Does this not currently work?

[robertphillips, 2012-04-16 15:14:12]

http://codereview.appspot.com/6031043/diff/1/src/gpu/GrGpu.h
File src/gpu/GrGpu.h (right):

http://codereview.appspot.com/6031043/diff/1/src/gpu/GrGpu.h#newcode75
src/gpu/GrGpu.h:75: class GrClipMaskCreator {
Can we talk about the header/private issue whenever you are ready?

I have renamed the class to be a Manager.

http://codereview.appspot.com/6031043/diff/1/src/gpu/gl/GrGpuGL.cpp
File src/gpu/gl/GrGpuGL.cpp (right):

http://codereview.appspot.com/6031043/diff/1/src/gpu/gl/GrGpuGL.cpp#newcode1748
src/gpu/gl/GrGpuGL.cpp:1748: this->enableScissoring(dirtyRect);
It may ... but I ran into enough difficultly just getting the primary
refactoring working that I wanted to wait on this. One way or another the #if 1
block will disappear on the next delivery.

[bsalomon, 2012-04-16 16:23:16]

On 2012/04/16 15:14:12, robertphillips wrote:
> http://codereview.appspot.com/6031043/diff/1/src/gpu/GrGpu.h
> File src/gpu/GrGpu.h (right):
> 
> http://codereview.appspot.com/6031043/diff/1/src/gpu/GrGpu.h#newcode75
> src/gpu/GrGpu.h:75: class GrClipMaskCreator {
> Can we talk about the header/private issue whenever you are ready?
> 
> I have renamed the class to be a Manager.
> 
> http://codereview.appspot.com/6031043/diff/1/src/gpu/gl/GrGpuGL.cpp
> File src/gpu/gl/GrGpuGL.cpp (right):
> 
>
http://codereview.appspot.com/6031043/diff/1/src/gpu/gl/GrGpuGL.cpp#newcode1748
> src/gpu/gl/GrGpuGL.cpp:1748: this->enableScissoring(dirtyRect);
> It may ... but I ran into enough difficultly just getting the primary
> refactoring working that I wanted to wait on this. One way or another the #if
1
> block will disappear on the next delivery.

LGTM (discussed where class should live in person)

[robertphillips, 2012-04-16 16:33:59]

committed as r3692