Fix missing variable MSVS expansion of copies arguments with Ninja.

Fix missing variable MSVS expansion of copies arguments with Ninja.

The Visual Studio emulation does not expand the special
variables like 'VSInstallDir' into the 'copies' actions.

Here is a small broken example:

{
  'targets': [
    {
      'target_name': 'Copy_Target',
      'type': 'none',
      'copies': [
        {
          'destination': '<(PRODUCT_DIR)',
          'files': [
            '$(VSInstallDir)\\bin\\cl.exe',
          ],
        },
      ],
    },
  ],
}

Patch from etienneb@chromium.org.

R=scottmg@chromium.org

Committed: https://code.google.com/p/gyp/source/detail?r=1870

[etienneb, 2014-03-12 04:05:20]

PTAL.

I didn't try to see if mac/linux are broken.
I launched a git cl try, but nothing seems to happen.

[scottmg, 2014-03-12 05:41:03]

On 2014/03/12 04:05:20, etienneb wrote:
> PTAL.
> 
> I didn't try to see if mac/linux are broken.
> I launched a git cl try, but nothing seems to happen.

I didn't look yet, but you can run tryjobs with "git try --no_search" (note, not
"git cl try").

[etienneb, 2014-03-12 11:43:24]

On 2014/03/12 05:41:03, scottmg wrote:
> On 2014/03/12 04:05:20, etienneb wrote:
> > PTAL.
> > 
> > I didn't try to see if mac/linux are broken.
> > I launched a git cl try, but nothing seems to happen.
> 
> I didn't look yet, but you can run tryjobs with "git try --no_search" (note,
not
> "git cl try").

I hope it's flaky, but I received:
Sorry, Tryserver is not available.

[etienneb, 2014-03-12 14:08:50]

On 2014/03/12 11:43:24, etienneb wrote:
> On 2014/03/12 05:41:03, scottmg wrote:
> > On 2014/03/12 04:05:20, etienneb wrote:
> > > PTAL.
> > > 
> > > I didn't try to see if mac/linux are broken.
> > > I launched a git cl try, but nothing seems to happen.
> > 
> > I didn't look yet, but you can run tryjobs with "git try --no_search" (note,
> not
> > "git cl try").
> 
> I hope it's flaky, but I received:
> Sorry, Tryserver is not available.

I'm not sure I can launch a try job, I'm not a chromium committer.

[scottmg, 2014-03-12 18:13:27]

On 2014/03/12 14:08:50, etienneb wrote:
> On 2014/03/12 11:43:24, etienneb wrote:
> > On 2014/03/12 05:41:03, scottmg wrote:
> > > On 2014/03/12 04:05:20, etienneb wrote:
> > > > PTAL.
> > > > 
> > > > I didn't try to see if mac/linux are broken.
> > > > I launched a git cl try, but nothing seems to happen.
> > > 
> > > I didn't look yet, but you can run tryjobs with "git try --no_search"
(note,
> > not
> > > "git cl try").
> > 
> > I hope it's flaky, but I received:
> > Sorry, Tryserver is not available.
> 
> I'm not sure I can launch a try job, I'm not a chromium committer.

Oh right, sorry. I forgot that way commits to svn as an intermediate. I just
launched them for the current patch.

[scottmg, 2014-03-12 18:19:17]

https://codereview.chromium.org/194883010/diff/20001/pylib/gyp/generator/ninj...
File pylib/gyp/generator/ninja.py (right):

https://codereview.chromium.org/194883010/diff/20001/pylib/gyp/generator/ninj...
pylib/gyp/generator/ninja.py:853: env = self.GetSortedXcodeEnv()
collapse this one too?

https://codereview.chromium.org/194883010/diff/20001/pylib/gyp/generator/ninj...
pylib/gyp/generator/ninja.py:1204: assert false, "Invalid architecture."
I feel like this will probably be called for Linux and is expected to return
None. Or maybe GetSortedXcodeEnv returns an empty object or something? Hmm.

https://codereview.chromium.org/194883010/diff/20001/pylib/gyp/generator/ninj...
pylib/gyp/generator/ninja.py:1206: def GetSortedMsvsEnv(self,
additional_settings=None):
The environment returned here isn't sorted, so I don't love these names.

Could you make "GetSortedToolchainEnv" just "GetToolchainEnv"?

[etienneb, 2014-03-12 20:07:52]

scott, could you launch try bots please.

https://codereview.chromium.org/194883010/diff/20001/pylib/gyp/generator/ninj...
File pylib/gyp/generator/ninja.py (right):

https://codereview.chromium.org/194883010/diff/20001/pylib/gyp/generator/ninj...
pylib/gyp/generator/ninja.py:853: env = self.GetSortedXcodeEnv()
On 2014/03/12 18:19:17, scottmg wrote:
> collapse this one too?

Done.

https://codereview.chromium.org/194883010/diff/20001/pylib/gyp/generator/ninj...
pylib/gyp/generator/ninja.py:1204: assert false, "Invalid architecture."
I changed it to have the same semantic as before.

On 2014/03/12 18:19:17, scottmg wrote:
> I feel like this will probably be called for Linux and is expected to return
> None. Or maybe GetSortedXcodeEnv returns an empty object or something? Hmm.

https://codereview.chromium.org/194883010/diff/20001/pylib/gyp/generator/ninj...
pylib/gyp/generator/ninja.py:1206: def GetSortedMsvsEnv(self,
additional_settings=None):
On 2014/03/12 18:19:17, scottmg wrote:
> The environment returned here isn't sorted, so I don't love these names.
> 
> Could you make "GetSortedToolchainEnv" just "GetToolchainEnv"?

Done.

[scottmg, 2014-03-12 20:20:10]

otherwise LGTM once trybots are green.

https://codereview.chromium.org/194883010/diff/40001/test/ninja/normalize-pat...
File test/ninja/normalize-paths-win/gyptest-normalize-paths.py (right):

https://codereview.chromium.org/194883010/diff/40001/test/ninja/normalize-pat...
test/ninja/normalize-paths-win/gyptest-normalize-paths.py:35: print copytarget
remove print

[etienneb, 2014-03-12 20:25:10]

On 2014/03/12 20:20:10, scottmg wrote:
> otherwise LGTM once trybots are green.
> 
>
https://codereview.chromium.org/194883010/diff/40001/test/ninja/normalize-pat...
> File test/ninja/normalize-paths-win/gyptest-normalize-paths.py (right):
> 
>
https://codereview.chromium.org/194883010/diff/40001/test/ninja/normalize-pat...
> test/ninja/normalize-paths-win/gyptest-normalize-paths.py:35: print copytarget
> remove print

I don't think I can land it.
I'm not a commiter (and CQ doesn't seems to work on this repo).
If this is right, you can land it when the bot are green.

[scottmg, 2014-03-12 21:58:52]

Committed patchset #4 manually as r1870 (presubmit successful).