Fix a bug in long variable name mapping.

Fix a bug in long variable name mapping.

Originally we only match varying variables in vertex/fragment shaders.  However, we should also match uniform names.  For example, if a uniform variable name is used in both vertex and fragment shader, and after mapping the names are different, then the name mapping table is no longer 1:1.

BUG=106638 (crbug)
TEST=http://seanseefried.com/chrome-bug/simple-broken.html shows a magenta square
Committed: http://code.google.com/p/angleproject/source/detail?r=939

[zmo, 2012-01-06 01:20:51]

Please review.  Tested locally with the test case provided by bug reporter.

Ken, should we add a test case in webgl conformance suite for this?  or it's too
late to add more?

[kbr1, 2012-01-06 02:11:20]

On 2012/01/06 01:20:51, zmo wrote:
> Please review.  Tested locally with the test case provided by bug reporter.

Please update this CL after you've successfully run the WebGL conformance suite
and I'll LGTM it.

> Ken, should we add a test case in webgl conformance suite for this?  or it's
too
> late to add more?

I think we should add a test case for this now to
glsl/misc/glsl-long-variable-names.html.

[zmo, 2012-01-06 02:15:21]

On 2012/01/06 02:11:20, kbr1 wrote:
> On 2012/01/06 01:20:51, zmo wrote:
> > Please review.  Tested locally with the test case provided by bug reporter.
> 
> Please update this CL after you've successfully run the WebGL conformance
suite
> and I'll LGTM it.

Tested, no more failures than without this CL.

> 
> > Ken, should we add a test case in webgl conformance suite for this?  or it's
> too
> > late to add more?

Will do.

> 
> I think we should add a test case for this now to
> glsl/misc/glsl-long-variable-names.html.

[kbr1, 2012-01-06 02:16:22]

On 2012/01/06 02:15:21, zmo wrote:
> Tested, no more failures than without this CL.

LGTM then.