Remove the clang hacks for the CrOS GN toolchains.

Remove the clang hacks for the CrOS GN toolchains.

They're no longer needed and are interfering w/ the CrOS
teams' attempts to actually use clang.

R=llozano@chromium.org, stevenjb@chromium.org, brettw@chromium.org
BUG=624901

[Dirk Pranke, 2016-06-30 20:11:49]

Patchset #1 (id:1) has been deleted

[Dirk Pranke, 2016-06-30 20:12:24]

Description was changed from

==========
Remove the clang hacks for the CrOS GN toolchains.

They're no longer needed and are interfering w/ the CrOS
teams' attempts to actually use clang.

R=llozano@chromium.org, stevenjb@chromium.org, bretw@chromium.org
BUG=624901
==========

to

==========
Remove the clang hacks for the CrOS GN toolchains.

They're no longer needed and are interfering w/ the CrOS
teams' attempts to actually use clang.

R=llozano@chromium.org, stevenjb@chromium.org, brettw@chromium.org
BUG=624901
==========

[Dirk Pranke, 2016-06-30 20:12:24]

dpranke@chromium.org changed reviewers:
+ brettw@chromium.org
- bretw@chromium.org

[stevenjb, 2016-06-30 20:46:53]

Since we have is_clang=false currently, this doesn't look like it can break
anything... right? In which case lgtm.

[Dirk Pranke, 2016-06-30 20:55:46]

On 2016/06/30 20:46:53, stevenjb wrote:
> Since we have is_clang=false currently, this doesn't look like it can break
> anything... right? In which case lgtm.

Correct, I don't believe this will break anything. The greatest risk is for
linux desktop CrOS builds, but the tryjobs will cover that.

[Dirk Pranke, 2016-06-30 20:57:42]

On 2016/06/30 20:55:46, Dirk Pranke wrote:
> On 2016/06/30 20:46:53, stevenjb wrote:
> > Since we have is_clang=false currently, this doesn't look like it can break
> > anything... right? In which case lgtm.
> 
> Correct, I don't believe this will break anything. The greatest risk is for
> linux desktop CrOS builds, but the tryjobs will cover that.

And, I just realized that this'll break those horribly. So, another patch is
coming up :).

[Dirk Pranke, 2016-06-30 22:21:36]

Abandoning, as this is the wrong approach (see linked bug for more).