Added preprocessor support for GL_FRAGMENT_PRECISION_HIGH.

Added preprocessor support for GL_FRAGMENT_PRECISION_HIGH.
I will send the tests in a separate patch because I need to refactor the way compiler_tests are setup.
Committed: https://code.google.com/p/angleproject/source/detail?r=1984

[Alok Priyadarshi, 2013-02-22 22:50:24]

I will figure out why flex is dropping support for 64 bit before committing.

[kbr1, 2013-02-23 00:12:56]

LGTM with one change.

https://codereview.appspot.com/7402051/diff/2001/include/GLSLANG/ShaderLang.h
File include/GLSLANG/ShaderLang.h (right):

https://codereview.appspot.com/7402051/diff/2001/include/GLSLANG/ShaderLang.h...
include/GLSLANG/ShaderLang.h:40: #define ANGLE_SH_VERSION 108
I think you should update this to 109 in case we need to add #ifdefs in WebKit
and Chromium (before rolling to this version of ANGLE) to handle this update.
Otherwise we may find ourselves in a state where we accidentally disable highp
support on some platforms for a period of time, and break shaders.

[Alok Priyadarshi, 2013-03-04 19:44:56]

https://codereview.appspot.com/7402051/diff/2001/include/GLSLANG/ShaderLang.h
File include/GLSLANG/ShaderLang.h (right):

https://codereview.appspot.com/7402051/diff/2001/include/GLSLANG/ShaderLang.h...
include/GLSLANG/ShaderLang.h:40: #define ANGLE_SH_VERSION 108
On 2013/02/23 00:12:56, kbr1 wrote:
> I think you should update this to 109 in case we need to add #ifdefs in WebKit
> and Chromium (before rolling to this version of ANGLE) to handle this update.
> Otherwise we may find ourselves in a state where we accidentally disable highp
> support on some platforms for a period of time, and break shaders.

Done.

[nicolas, 2013-03-04 19:52:58]

Even though this has been withdrawn as an extension
(http://www.khronos.org/registry/gles/extensions/OES/OES_fragment_precision_hi...),
I think it can still be treated internally as an extension. It's just not
reported by glGetString(GL_EXTENSIONS). Treating it as an extension makes it
very straightforward to support it. See my patch on ANGLE issue 413.

Your patch set looks good to me too if you think treating it as an extension is
too confusing.

[Alok Priyadarshi, 2013-03-04 20:02:31]

Ok. I am going to commit this patch. In a separate path, I will enable
fragment-precision-high unconditionally for ANGLE.

[Alok Priyadarshi, 2013-03-04 20:03:15]

Committed patchset #3 manually as r1984 (presubmit successful).

[thakis, 2015-03-08 01:49:15]

https://codereview.appspot.com/7402051/diff/7001/src/compiler/ParseHelper.h
File src/compiler/ParseHelper.h (right):

https://codereview.appspot.com/7402051/diff/7001/src/compiler/ParseHelper.h#n...
src/compiler/ParseHelper.h:42: checksPrecisionErrors(checksPrecErrors),
Hi Shannon,

I noticed that this constructor happens to initialize all fields except the new
field fragmentPrecisionHigh you added in this CL. Is that intentional, or was
that an oversight?

Thanks,
Nico

[thakis, 2015-03-08 01:49:48]

…or Alok. For some reason, `git show` claims this line is from
shannon.woods%transgaming.com@gtempaccount.com

[Jamie Madill, 2015-03-09 14:28:27]

On 2015/03/08 01:49:48, thakis wrote:
> …or Alok. For some reason, `git show` claims this line is from
> mailto:shannon.woods%transgaming.com@gtempaccount.com

Nico, I believe this is an oversight. We should initialize everything in the
constructor. Is this giving a warning in Clang? Feel free to upload a CL to fix
it.

[thakis, 2015-03-09 14:35:31]

On Mon, Mar 9, 2015 at 7:28 AM, <jmadill@chromium.org> wrote:

> On 2015/03/08 01:49:48, thakis wrote:
>
>> …or Alok. For some reason, `git show` claims this line is from
>> mailto:shannon.woods%transgaming.com@gtempaccount.com
>>
>
> Nico, I believe this is an oversight. We should initialize everything in
> the constructor. Is this giving a warning in Clang?
>

A warning prototype I wrote flagged this. However, it looks like the
warning has too many false positives, so I won't merge it into clang.

[Alok Priyadarshi, 2015-03-09 19:42:56]

https://codereview.appspot.com/7402051/diff/7001/src/compiler/ParseHelper.h
File src/compiler/ParseHelper.h (right):

https://codereview.appspot.com/7402051/diff/7001/src/compiler/ParseHelper.h#n...
src/compiler/ParseHelper.h:42: checksPrecisionErrors(checksPrecErrors),
On 2015/03/08 01:49:15, thakis wrote:
> Hi Shannon,
> 
> I noticed that this constructor happens to initialize all fields except the
new
> field fragmentPrecisionHigh you added in this CL. Is that intentional, or was
> that an oversight?
> 
> Thanks,
> Nico

Even if it was intentional I do not remember after 2 years :)
Looks like an oversight to me.