Fix default ARCHS value for iOS project.

Fix default ARCHS value for iOS project.

Xcode 5.0.0 only default to x86_64 when building Mac project, not
iOS projects (since default target of iOS 5.0 does not supports
64-bits binaries).

Patch by Sylvain Defresne <sdefresne@chromium.org>;

BUG=chromium:322075
R=thakis@chromium.org

Committed: https://code.google.com/p/gyp/source/detail?r=1798

[sdefresne, 2013-11-21 17:19:02]

Please take a look.

[Nico, 2013-11-22 01:40:07]

Can you add a test for this?

[sdefresne, 2013-11-25 14:38:06]

I've tested more, and in fact, Xcode does not default to x86_64
when building iOS projects but still uses i386 (for the simulator
target).

I've changed my CL to be much simpler and just uses the old
default value in _GetDefaultArch when generating iOS project.
Could you take another look?

[Nico, 2013-11-25 19:39:07]

lgtm

Sounds like this makes things better than before.

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/TestA...
File test/ios/app-bundle/TestApp/only-compile-in-32-bits.m (right):

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/TestA...
test/ios/app-bundle/TestApp/only-compile-in-32-bits.m:7: void
OnlyCompileIn32Bits()
This is a bit subtle. How about

  #if !defined(__LP64__)
  #error 32bit build
  #endif

or

  #if sizeof(long) == 4)
  #error 32bit build
  #endif?

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/test-...
File test/ios/app-bundle/test-archs.gyp (right):

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/test-...
test/ios/app-bundle/test-archs.gyp:1: # Copyright (c) 2011 Google Inc. All
rights reserved.
2013

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/test-...
test/ios/app-bundle/test-archs.gyp:31: '-fobjc-abi-version=2',
(do you need this?)

https://codereview.chromium.org/81213003/diff/90001/test/ios/gyptest-archs.py
File test/ios/gyptest-archs.py (right):

https://codereview.chromium.org/81213003/diff/90001/test/ios/gyptest-archs.py...
test/ios/gyptest-archs.py:27: 
2 newlines between toplevel things

[Nico, 2013-11-25 19:39:53]

(since this is a gyp CL, say "BUG=chromium:12345", to make clear this isn't for
a gyp bug but a chromium one)

[sdefresne, 2013-11-26 09:29:15]

I've addressed all the comments. I don't have commit access,
so could you commit the CL once you're happy with it?

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/TestA...
File test/ios/app-bundle/TestApp/only-compile-in-32-bits.m (right):

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/TestA...
test/ios/app-bundle/TestApp/only-compile-in-32-bits.m:7: void
OnlyCompileIn32Bits()
On 2013/11/25 19:39:07, Nico wrote:
> This is a bit subtle. How about
> 
>   #if !defined(__LP64__)
>   #error 32bit build
>   #endif
> 
> or
> 
>   #if sizeof(long) == 4)
>   #error 32bit build
>   #endif?

Done.

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/test-...
File test/ios/app-bundle/test-archs.gyp (right):

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/test-...
test/ios/app-bundle/test-archs.gyp:1: # Copyright (c) 2011 Google Inc. All
rights reserved.
On 2013/11/25 19:39:07, Nico wrote:
> 2013

Done everywhere.

https://codereview.chromium.org/81213003/diff/90001/test/ios/app-bundle/test-...
test/ios/app-bundle/test-archs.gyp:31: '-fobjc-abi-version=2',
On 2013/11/25 19:39:07, Nico wrote:
> (do you need this?)

Removed (came from copying file).

https://codereview.chromium.org/81213003/diff/90001/test/ios/gyptest-archs.py
File test/ios/gyptest-archs.py (right):

https://codereview.chromium.org/81213003/diff/90001/test/ios/gyptest-archs.py...
test/ios/gyptest-archs.py:27: 
On 2013/11/25 19:39:07, Nico wrote:
> 2 newlines between toplevel things

Done.

[justincohen, 2013-11-26 14:19:44]

Committed patchset #5 manually as r1798 (presubmit successful).

[scottmg, 2013-11-26 20:31:32]

On 2013/11/26 14:19:44, justincohen wrote:
> Committed patchset #5 manually as r1798 (presubmit successful).

This broke:
http://build.chromium.org/p/tryserver.nacl/builders/gyp-mac/builds/1417

"git try --no_search" to run tryjobs, for future reference.

[justincohen, 2013-11-26 20:33:13]

Apologies.  I can revert, but there's a CL in review here to disable the tests
(the bots are running a very old Xcode)

https://codereview.chromium.org/88813002/

Should I revert or wait for this?

[scottmg, 2013-11-26 20:40:01]

On 2013/11/26 20:33:13, justincohen wrote:
> Apologies.  I can revert, but there's a CL in review here to disable the tests
> (the bots are running a very old Xcode)
> 
> https://codereview.chromium.org/88813002/
> 
> Should I revert or wait for this?

Oh that's fine, I didn't see the other CL.

(Also, I didn't mean for my comment to sound so passive-aggressive, I never seem
to remember that "git cl try" doesn't work, and that the magic --no_search is
required. :)

[sdefresne, 2013-11-26 22:49:22]

On 2013/11/26 20:40:01, scottmg wrote:
> On 2013/11/26 20:33:13, justincohen wrote:
> > Apologies.  I can revert, but there's a CL in review here to disable the
tests
> > (the bots are running a very old Xcode)
> > 
> > https://codereview.chromium.org/88813002/
> > 
> > Should I revert or wait for this?
> 
> Oh that's fine, I didn't see the other CL.
> 
> (Also, I didn't mean for my comment to sound so passive-aggressive, I never
seem
> to remember that "git cl try" doesn't work, and that the magic --no_search is
> required. :)

Sorry, it's my fault. I only have the most recent version of Xcode installed,
and I did use "git cl try" (as strongly encouraged by the message [1] that I got
when I tried to run "git try") to run the bots, without realizing that it does
not
work.

Maybe this message should be made conditional on the fact that the repository
is chromium. Or "git cl try" should be fixed to work on gyp repository too.

[1]: https://codereview.chromium.org/74273002/patch/1/10001