Cleanup handling of GL versions, fix function names

[bsalomon, 2011-09-22 13:46:09]

On 2011/09/21 19:11:15, bsalomon wrote:

ping

[reed1, 2011-09-22 13:57:39]

LGTM

[junov1, 2011-09-22 14:19:20]

Just one minor comment. Otherwise LGTM

http://codereview.appspot.com/5097044/diff/5001/gpu/include/GrGLInterface.h
File gpu/include/GrGLInterface.h (right):

http://codereview.appspot.com/5097044/diff/5001/gpu/include/GrGLInterface.h#n...
gpu/include/GrGLInterface.h:29: #define GR_GL_VER(major, minor) (((int)(major)
<< 16) | ((int)(minor)))
Typing as GrFixed is confusing since the interpretation of the fractional part
is x/64k. For example this macro would convert a version of 3.3 to the number
3.0000457763671875 in fixed point representation. In general I think your
approach is sound, but I would typedef it to uint32_t rather than GrFixed to
avoid misinterpretation.

[reed1, 2011-09-22 14:21:15]

+1 good point. We're really packing two numbers into one int.

On 2011/09/22 14:19:20, junov1 wrote:
> Just one minor comment. Otherwise LGTM
> 
> http://codereview.appspot.com/5097044/diff/5001/gpu/include/GrGLInterface.h
> File gpu/include/GrGLInterface.h (right):
> 
>
http://codereview.appspot.com/5097044/diff/5001/gpu/include/GrGLInterface.h#n...
> gpu/include/GrGLInterface.h:29: #define GR_GL_VER(major, minor) (((int)(major)
> << 16) | ((int)(minor)))
> Typing as GrFixed is confusing since the interpretation of the fractional part
> is x/64k. For example this macro would convert a version of 3.3 to the number
> 3.0000457763671875 in fixed point representation. In general I think your
> approach is sound, but I would typedef it to uint32_t rather than GrFixed to
> avoid misinterpretation.

[bsalomon, 2011-09-22 14:34:21]

Closed with 2308,9. Changed to uint32_t.