Use vertex buffer objects instead of client side arrays in fixed-function codepaths

Use vertex buffer objects instead of client side arrays in fixed-function codepaths

Pass vertices to the fixed-function shaders in vertex buffer object
stored at attribute 0. This is specified in OpenGL Compatibility
Profile Section 10.8.2 (Transferring Vertices With Vertex Attribute
Zero).

This makes future integration to Chromium command buffer better, since
the command buffer already has code for vertex buffer objects but not
vertex arrays.

Delete unused function pointers related to fixed-function codepaths.

Committed: http://code.google.com/p/skia/source/detail?r=13557

[Kimmo Kinnunen, 2014-02-14 14:30:09]

Brian, would this direction be acceptable in order to enable NVPR in Chromium?
Eg. replacing client-side vertex arrays with vertex buffer objects that are
bound to hardcoded attribute locations?

[bsalomon, 2014-02-14 14:40:20]

I'm not really sure that I understand the proposal. Is there some info about
GL_CHROMIUM_NV_path_rendering? Also, since you're removing some of the code
paths used for GL_NV_path_rendering, I'm wondering if there are changes to that
spec as well?

[Kimmo Kinnunen, 2014-02-18 14:58:33]

On 2014/02/14 14:40:20, bsalomon wrote:
> I'm not really sure that I understand the proposal. Is there some info about
> GL_CHROMIUM_NV_path_rendering? Also, since you're removing some of the code
> paths used for GL_NV_path_rendering, I'm wondering if there are changes to
that
> spec as well?

Sorry, I was maybe jumping ahead too much.

I updated the patch and the description to be more explicit:

What I'm trying to do here is to replace glVertexPointer calls with vertex
buffer object usage (using glVertexAttribPointer + glEnableVertexAttrib(0)).

(The Chromium related patch was split to
https://codereview.chromium.org/163913007)

[bsalomon, 2014-02-20 15:44:28]

On 2014/02/18 14:58:33, kkinnunen wrote:
> On 2014/02/14 14:40:20, bsalomon wrote:
> > I'm not really sure that I understand the proposal. Is there some info about
> > GL_CHROMIUM_NV_path_rendering? Also, since you're removing some of the code
> > paths used for GL_NV_path_rendering, I'm wondering if there are changes to
> that
> > spec as well?
> 
> Sorry, I was maybe jumping ahead too much.
> 
> I updated the patch and the description to be more explicit:
> 
> What I'm trying to do here is to replace glVertexPointer calls with vertex
> buffer object usage (using glVertexAttribPointer + glEnableVertexAttrib(0)).
> 
> (The Chromium related patch was split to
> https://codereview.chromium.org/163913007)

I see... so would the regular driver outside of Chromium also expose the new
functionality of the pseudo-extension? I think we wouldn't want to remove this
code until we have drivers that can work with the new code.

[Kimmo Kinnunen, 2014-02-21 06:27:12]

On 2014/02/20 15:44:28, bsalomon wrote:
> On 2014/02/18 14:58:33, kkinnunen wrote:
> > I updated the patch and the description to be more explicit:
> > 
> > What I'm trying to do here is to replace glVertexPointer calls with vertex
> > buffer object usage (using glVertexAttribPointer + glEnableVertexAttrib(0)).
> I see... so would the regular driver outside of Chromium also expose the new
> functionality of the pseudo-extension? I think we wouldn't want to remove this
> code until we have drivers that can work with the new code.

Nono, the patch here is intended to work for any driver today, NVIDIA or ATI or
whatever.

It does not need any driver work or any pseudo extensions. There's no new
functionality.

This intends to use fixed-function vertex processing as it is defined in OGL
spec.

[bsalomon, 2014-02-21 14:05:21]

Ah... I see, I took another look at the patch and get it now :) lgtm.

[Kimmo Kinnunen, 2014-02-24 11:51:55]

The CQ bit was checked by kkinnunen@nvidia.com

[commit-bot: I haz the power, 2014-02-24 11:51:58]

CQ is trying da patch. Follow status at
https://skia-tree-status.appspot.com/cq/kkinnunen@nvidia.com/130423013/470001

[commit-bot: I haz the power, 2014-02-24 12:05:16]

Change committed as 13557