Make WebGraphics3DCommandBufferImpl use GLES2Implemenation directly.

Make WebGraphics3DCommandBufferImpl use GLES2Implemenation directly.

Note: I can do this 2 ways. I can either make it call it truly
direclty as in

  gles2_implementation_->BindBuffer(args)

or I can do as I have using CALL_GL_API in which case it's

  CALL_GL_API(BindBuffer)(args)

The only advantage to the second format is a macro changes
it back to the way it was. On the other hand the class is
called WebGraphics3DCommandBufferImpl so maybe I should
change it to the first style.

TEST=try-bots and ran chrome myself
BUG=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=86325

[greggman, 2011-05-21 00:49:02]

Happy to change this if you object the style

[Ken Russell (switch to Gerrit), 2011-05-21 01:09:21]

These are just my thoughts. I don't have a strong opinion. Adding Al to the
reviewer list as he has probably given it more thought.

http://codereview.chromium.org/7056020/diff/1/content/renderer/webgraphicscon...
File content/renderer/webgraphicscontext3d_command_buffer_impl.cc (right):

http://codereview.chromium.org/7056020/diff/1/content/renderer/webgraphicscon...
content/renderer/webgraphicscontext3d_command_buffer_impl.cc:221:
localMakeContextCurrent();
I thought the reason for this change was to make the implementation more robust
by eliminating the need for making context_ current, at least in the majority of
the entry points (I don't know about CreateParentTexture, DeleteParentTexture,
etc.). If that's the case then we should just get rid of the macro, go through
the GLES2Implementation, and delete the calls to makeContextCurrent (or at least
all that can be deleted) rather than adding the new localMakeContextCurrent.

http://codereview.chromium.org/7056020/diff/1/content/renderer/webgraphicscon...
File content/renderer/webgraphicscontext3d_command_buffer_impl.h (right):

http://codereview.chromium.org/7056020/diff/1/content/renderer/webgraphicscon...
content/renderer/webgraphicscontext3d_command_buffer_impl.h:442: // The
GLES2Implementation we use for OpenGL R
Is the "R" at the end a typo?

[greggman, 2011-05-21 01:22:49]

http://codereview.chromium.org/7056020/diff/1/content/renderer/webgraphicscon...
File content/renderer/webgraphicscontext3d_command_buffer_impl.cc (right):

http://codereview.chromium.org/7056020/diff/1/content/renderer/webgraphicscon...
content/renderer/webgraphicscontext3d_command_buffer_impl.cc:221:
localMakeContextCurrent();
You're right.  I changed it to the simpler way.

http://codereview.chromium.org/7056020/diff/1/content/renderer/webgraphicscon...
File content/renderer/webgraphicscontext3d_command_buffer_impl.h (right):

http://codereview.chromium.org/7056020/diff/1/content/renderer/webgraphicscon...
content/renderer/webgraphicscontext3d_command_buffer_impl.h:442: // The
GLES2Implementation we use for OpenGL R
On 2011/05/21 01:09:21, kbr wrote:
> Is the "R" at the end a typo?

Done.

[apatrick_chromium, 2011-05-23 18:15:50]

I think this is the better way too. LGTM.

[Ken Russell (switch to Gerrit), 2011-05-23 18:20:28]

LGTM

[greggman, 2011-05-23 19:03:36]

hey James, can you please review this (or stamp approval) since you are the
OWNER of content/renderer

[jamesr, 2011-05-23 19:17:17]

LGTM!