Get rid of NonCachedProgramInfoManager.

Get rid of NonCachedProgramInfoManager.

It is not used at all.

Note that I didn't change any logic at all except
1) Get rid of NonCachedProgramInfoManager.
2) Merge CachedProgramInfoManager into ProgramInfoManager.
3) Get rid of the Create() function.
4) Add ~VertexAttrib() ~UniformInfo() ~Program() as required by compiler.

BUG=453127
TEST=no change at all
R=piman@chromium.org

Committed: https://crrev.com/730fb0d7dcd9fe50368a22a60478efa656819e21
Cr-Commit-Position: refs/heads/master@{#314027}

[Zhenyao Mo, 2015-01-30 21:26:44]

piman: please take a look.

[Zhenyao Mo, 2015-01-30 21:28:28]

Others, FYI (feel free to review)

[bajones, 2015-01-30 21:40:21]

bajones@chromium.org changed reviewers:
+ bajones@chromium.org

[bajones, 2015-01-30 21:40:22]

Generally LG here, but I'll let piman give final approval. A couple of comments
re: the trybot failures.

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
File gpu/command_buffer/client/program_info_manager.h (right):

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager.h:56: ~UniformInfo() {
Pretty sure just moving the closing curly brace onto this line would be enough
to avoid the compile errors that Linux is throwing.

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager.h:73: ~VertexAttrib() {
Ditto

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager.h:86: ~Program() {
Ditto

[piman, 2015-01-30 21:55:39]

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
File gpu/command_buffer/client/program_info_manager.h (right):

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager.h:56: ~UniformInfo() {
On 2015/01/30 21:40:22, bajones wrote:
> Pretty sure just moving the closing curly brace onto this line would be enough
> to avoid the compile errors that Linux is throwing.

It actually needs to be out-of-line because of non-POD fields.

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
File gpu/command_buffer/client/program_info_manager_program.cc (right):

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager_program.cc:1: // Copyright (c)
2015 The Chromium Authors. All rights reserved.
Why separate this into another file?

[Zhenyao Mo, 2015-01-30 22:11:38]

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
File gpu/command_buffer/client/program_info_manager.h (right):

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager.h:56: ~UniformInfo() {
Thank you both for the tips.  However, why my local Linux bot compiler did not
catch such failures?

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
File gpu/command_buffer/client/program_info_manager_program.cc (right):

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager_program.cc:1: // Copyright (c)
2015 The Chromium Authors. All rights reserved.
On 2015/01/30 21:55:39, piman (Very slow to review) wrote:
> Why separate this into another file?

Because I am planning to add UniformBlock related catching here, almost double
the file size.  So to separate the Program class files into its own file.  I
feel it's easier to read and maintain this way.

[piman, 2015-01-30 22:15:45]

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
File gpu/command_buffer/client/program_info_manager.h (right):

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager.h:56: ~UniformInfo() {
On 2015/01/30 22:11:38, Zhenyao Mo wrote:
> Thank you both for the tips.  However, why my local Linux bot compiler did not
> catch such failures?

It's a clang plugin.

[piman, 2015-01-30 22:18:23]

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
File gpu/command_buffer/client/program_info_manager_program.cc (right):

https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
gpu/command_buffer/client/program_info_manager_program.cc:1: // Copyright (c)
2015 The Chromium Authors. All rights reserved.
On 2015/01/30 22:11:38, Zhenyao Mo wrote:
> On 2015/01/30 21:55:39, piman (Very slow to review) wrote:
> > Why separate this into another file?
> 
> Because I am planning to add UniformBlock related catching here, almost double
> the file size.  So to separate the Program class files into its own file.  I
> feel it's easier to read and maintain this way.

I disagree :) If a class is defined in foo.h, the natural place to look for the
implementation is foo.cc. If it's not there you have to look for it. For
something named "Program", it is difficult to search for it in the codebase.

Another option is to not make it an inner class (which are discouraged anyway).

[Zhenyao Mo, 2015-01-30 22:39:37]

On 2015/01/30 22:18:23, piman (Very slow to review) wrote:
>
https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
> File gpu/command_buffer/client/program_info_manager_program.cc (right):
> 
>
https://codereview.chromium.org/893713002/diff/1/gpu/command_buffer/client/pr...
> gpu/command_buffer/client/program_info_manager_program.cc:1: // Copyright (c)
> 2015 The Chromium Authors. All rights reserved.
> On 2015/01/30 22:11:38, Zhenyao Mo wrote:
> > On 2015/01/30 21:55:39, piman (Very slow to review) wrote:
> > > Why separate this into another file?
> > 
> > Because I am planning to add UniformBlock related catching here, almost
double
> > the file size.  So to separate the Program class files into its own file.  I
> > feel it's easier to read and maintain this way.
> 
> I disagree :) If a class is defined in foo.h, the natural place to look for
the
> implementation is foo.cc. If it's not there you have to look for it. For
> something named "Program", it is difficult to search for it in the codebase.
> 
> Another option is to not make it an inner class (which are discouraged
anyway).

OK, moved it back to the original .cc file.

[piman, 2015-01-30 22:42:07]

LGTM, thanks

[Zhenyao Mo, 2015-01-30 22:42:45]

The CQ bit was checked by zmo@chromium.org

[commit-bot: I haz the power, 2015-01-30 22:43:22]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/893713002/40001

[commit-bot: I haz the power, 2015-01-30 23:34:43]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-01-30 23:36:17]

Patchset 3 (id:??) landed as
https://crrev.com/730fb0d7dcd9fe50368a22a60478efa656819e21
Cr-Commit-Position: refs/heads/master@{#314027}