Remove GrClipMaskManager's path renderer chain (in favor of GrContext's)

Remove GrClipMaskManager's path renderer chain (in favor of GrContext's)

[robertphillips, 2012-05-22 20:04:59]

First refactoring to simplify SW-only clip mask generation test logic (i.e.,
useSWOnlyPath)

http://codereview.appspot.com/6211078/diff/1/include/gpu/GrContext.h
File include/gpu/GrContext.h (right):

http://codereview.appspot.com/6211078/diff/1/include/gpu/GrContext.h#newcode686
include/gpu/GrContext.h:686: bool antiAlias);
Brian - we could also befriend GrClipMaskManager?

[bsalomon, 2012-05-22 20:11:13]

On 2012/05/22 20:04:59, robertphillips wrote:
> First refactoring to simplify SW-only clip mask generation test logic (i.e.,
> useSWOnlyPath)
> 
> http://codereview.appspot.com/6211078/diff/1/include/gpu/GrContext.h
> File include/gpu/GrContext.h (right):
> 
>
http://codereview.appspot.com/6211078/diff/1/include/gpu/GrContext.h#newcode686
> include/gpu/GrContext.h:686: bool antiAlias);
> Brian - we could also befriend GrClipMaskManager?

LGTM

[bsalomon, 2012-05-22 20:23:43]

Meant to publish this comment earlier

http://codereview.appspot.com/6211078/diff/1/include/gpu/GrContext.h
File include/gpu/GrContext.h (right):

http://codereview.appspot.com/6211078/diff/1/include/gpu/GrContext.h#newcode686
include/gpu/GrContext.h:686: bool antiAlias);
On 2012/05/22 20:04:59, robertphillips wrote:
> Brian - we could also befriend GrClipMaskManager?

I think having this public is OK. GrPathRenderer.h is in src/gpu so someone
would have to be looking where there not supposed to use this. Also, we will
eventually getting around to creating a GrContextInternal that hides all the
junk that shouldn't be public from the public interface.

[robertphillips, 2012-05-22 20:29:29]

committed as r4032