Added simple vertex shader demo....

Added simple vertex shader demo. I had to make the following modifications to third-party code:
- Changed client-side vertex array to VBO
- Used GL_UNSIGNED_SHORT instead of GL_UNSIGNED_INT in glDrawElements because GL_UNSIGNED_INT is not supported for by OpenGL ES 2.0 for indices
BUG=26099
TEST=Run simple_vertex_shader executable, you should see a rotating cube.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=36127

[alokp, 2010-01-11 21:28:51]

Hi Gregg,
Could you quickly check if I am missing something. The same demo is working fine
with AMD's gles2 emulator but not with command buffer. Is there anything in the
demo not implemented by the command buffer. The only change I made was to
convert client side arrays to vbo.

Thanks,
-Alok

[greggman, 2010-01-11 22:54:03]

On Mon, Jan 11, 2010 at 1:28 PM, <alokp@chromium.org> wrote:

> Reviewers: greggman,
>
> Message:
> Hi Gregg,
> Could you quickly check if I am missing something. The same demo is working
> fine
> with AMD's gles2 emulator but not with command buffer. Is there anything in
> the
> demo not implemented by the command buffer. The only change I made was to
> convert client side arrays to vbo.
>

glDrawElements ( GL_TRIANGLES, userData->numIndices, GL_UNSIGNED_INT, 0 );

GL_UNSIGNED_INT for indicies is not supported by OpenGL ES 2.0 at all.  You
have to use GL_UNSIGNED_BYTE or GL_UNSIGNED_SHORT




> Thanks,
> -Alok
>
> Description:
> Added simple vertex shader demo.
> BUG=26099
> TEST=Run simple_vertex_shader executable, you should see a rotating cube.
>
> Please review this at http://codereview.chromium.org/543015
>
> SVN Base: svn://chrome-svn/chrome/trunk/src/
>
> Affected files:
>  M     gpu/demos/demos.gyp
>  A     gpu/demos/simple_vertex_shader/main.cc
>  M
> third_party/gles2_book/Chapter_8/Simple_VertexShader/Simple_VertexShader.h
>  M
> third_party/gles2_book/Chapter_8/Simple_VertexShader/Simple_VertexShader.c
>
>
>

[greggman, 2010-01-11 23:14:17]

PS: This bug is listed in the errata page for the GL ES 2.0 book

http://opengles-book.com/errata.html


>    - Several of the examples in the sample code pass GL_UNSIGNED_INT in to
>    glDrawElements for the parameter "type", but this is only supported if the
>    OES_element_index_uint extension is supported.  The example code on the
>    website for iPhone were update to use GL_UNSIGNED_SHORT (reported by
>    Vidyasagar Gopireddy)
>
>
On Mon, Jan 11, 2010 at 2:53 PM, Gregg Tavares <gman@chromium.org> wrote:

>
>
> On Mon, Jan 11, 2010 at 1:28 PM, <alokp@chromium.org> wrote:
>
>> Reviewers: greggman,
>>
>> Message:
>> Hi Gregg,
>> Could you quickly check if I am missing something. The same demo is
>> working fine
>> with AMD's gles2 emulator but not with command buffer. Is there anything
>> in the
>> demo not implemented by the command buffer. The only change I made was to
>> convert client side arrays to vbo.
>>
>
> glDrawElements ( GL_TRIANGLES, userData->numIndices, GL_UNSIGNED_INT, 0 );
>
> GL_UNSIGNED_INT for indicies is not supported by OpenGL ES 2.0 at all.  You
> have to use GL_UNSIGNED_BYTE or GL_UNSIGNED_SHORT
>
>
>
>
>> Thanks,
>> -Alok
>>
>> Description:
>> Added simple vertex shader demo.
>> BUG=26099
>> TEST=Run simple_vertex_shader executable, you should see a rotating cube.
>>
>> Please review this at http://codereview.chromium.org/543015
>>
>> SVN Base: svn://chrome-svn/chrome/trunk/src/
>>
>> Affected files:
>>  M     gpu/demos/demos.gyp
>>  A     gpu/demos/simple_vertex_shader/main.cc
>>  M
>> third_party/gles2_book/Chapter_8/Simple_VertexShader/Simple_VertexShader.h
>>  M
>> third_party/gles2_book/Chapter_8/Simple_VertexShader/Simple_VertexShader.c
>>
>>
>>
>

[alokp, 2010-01-12 17:14:05]

Hi Gregg,
Changing int to short did the trick. Thanks for catching it. Should we report a
GL error in this case?

-Alok

[greggman, 2010-01-12 17:34:35]

There was a bug I fixed yesterday that was preventing errors from being
reported correctly. I'll try to have that checked in today.

-gregg


On Tue, Jan 12, 2010 at 9:14 AM, <alokp@chromium.org> wrote:

> Hi Gregg,
> Changing int to short did the trick. Thanks for catching it. Should we
> report a
> GL error in this case?
>
> -Alok
>
>
> http://codereview.chromium.org/543015
>

[alokp, 2010-01-12 17:35:49]

Please re-review. I still need to update mvp matrix every frame, but I will do
it in another CL. That will require adding a timer and changing the function
signature of Application::Draw().

[alokp, 2010-01-13 02:55:45]

Hi Gregg, Could you please review this. I am kinda blocked. Thanks.

[greggman, 2010-01-13 03:10:43]

LGTM

http://codereview.chromium.org/543015/diff/9001/9005
File third_party/gles2_book/Chapter_8/Simple_VertexShader/Simple_VertexShader.c
(right):

http://codereview.chromium.org/543015/diff/9001/9005#newcode128
third_party/gles2_book/Chapter_8/Simple_VertexShader/Simple_VertexShader.c:128:
glEnableVertexAttribArray ( userData->positionLoc );
Any particular reason you switched the order of these two lines?

[alokp, 2010-01-13 03:52:59]

http://codereview.chromium.org/543015/diff/9001/9005
File third_party/gles2_book/Chapter_8/Simple_VertexShader/Simple_VertexShader.c
(right):

http://codereview.chromium.org/543015/diff/9001/9005#newcode128
third_party/gles2_book/Chapter_8/Simple_VertexShader/Simple_VertexShader.c:128:
glEnableVertexAttribArray ( userData->positionLoc );
no particular reason. does it matter?