Adding a test to cover XCode UID collisions.

Adding a test to cover XCode UID collisions.

CL https://codereview.appspot.com/6476070/ fixed the issue, this
test verifies that the collision is no longer happening.

r1504

[noyau (Ping after 24h), 2012-09-13 11:57:04]

As requested by @thakis in the related CL, this adds a test to cover the corner
case that was causing the issue.

[Nico, 2012-09-13 12:07:23]

Thanks!

https://chromiumcodereview.appspot.com/10911280/diff/1/test/mac/gyptest-dupli...
File test/mac/gyptest-duplicate-names.py (right):

https://chromiumcodereview.appspot.com/10911280/diff/1/test/mac/gyptest-dupli...
test/mac/gyptest-duplicate-names.py:16: if sys.platform == 'darwin':
The test didn't look mac-specific to me. Can you remove this line? (…and send
try jobs?)

[noyau (Ping after 24h), 2012-09-13 13:08:28]

On 2012/09/13 12:07:23, Nico wrote:
> Thanks!
> 
>
https://chromiumcodereview.appspot.com/10911280/diff/1/test/mac/gyptest-dupli...
> File test/mac/gyptest-duplicate-names.py (right):
> 
>
https://chromiumcodereview.appspot.com/10911280/diff/1/test/mac/gyptest-dupli...
> test/mac/gyptest-duplicate-names.py:16: if sys.platform == 'darwin':
> The test didn't look mac-specific to me. Can you remove this line? (…and send
> try jobs?)

This is testing a mac specific issue with xcode generation, hence the location
in mac/ and the darwin test. If I remove the darwin test should I move the test
up in a non mac specific directory (and if yes, where?)  ?

[Nico, 2012-09-14 03:16:40]

On 2012/09/13 13:08:28, noyau wrote:
> On 2012/09/13 12:07:23, Nico wrote:
> > Thanks!
> > 
> >
>
https://chromiumcodereview.appspot.com/10911280/diff/1/test/mac/gyptest-dupli...
> > File test/mac/gyptest-duplicate-names.py (right):
> > 
> >
>
https://chromiumcodereview.appspot.com/10911280/diff/1/test/mac/gyptest-dupli...
> > test/mac/gyptest-duplicate-names.py:16: if sys.platform == 'darwin':
> > The test didn't look mac-specific to me. Can you remove this line? (…and
send
> > try jobs?)
> 
> This is testing a mac specific issue with xcode generation, hence the location
> in mac/ and the darwin test.

Right, but this should work with all other generators too.

> If I remove the darwin test should I move the test
> up in a non mac specific directory (and if yes, where?)  ?

I think this makes sense. I don't know where it fits in. Look around a bit (find
test -path '*duplicate*' maybe), or pick something yourself if you don't find
anything related.

[noyau (Ping after 24h), 2012-09-14 12:00:35]

PTAL: moved the test and performed a try-bot run as requested.

https://chromiumcodereview.appspot.com/10911280/diff/1/test/mac/gyptest-dupli...
File test/mac/gyptest-duplicate-names.py (right):

https://chromiumcodereview.appspot.com/10911280/diff/1/test/mac/gyptest-dupli...
test/mac/gyptest-duplicate-names.py:16: if sys.platform == 'darwin':
On 2012/09/13 12:07:23, Nico wrote:
> The test didn't look mac-specific to me. Can you remove this line? (…and send
> try jobs?)

Moved the code out of mac/ into generator-output/ and removed that line.

https://chromiumcodereview.appspot.com/10911280/diff/10/test/generator-output...
File test/generator-output/gyptest-duplicate-names.py (right):

https://chromiumcodereview.appspot.com/10911280/diff/10/test/generator-output...
test/generator-output/gyptest-duplicate-names.py:16: test =
TestGyp.TestGyp(formats=['xcode'])
Removed the if making it a darwin specific clause and limited the test to xcode
only (as the collision was xcode generation specific)

[noyau (Ping after 24h), 2012-09-14 14:21:25]

PTAL: Now runs with all generators.

[Nico, 2012-09-14 14:34:53]

lgtm. Thanks again for adding this!

The generator-output folder contains tests for the -Ggenerator_output flag,
which this test doesn't have much to do with. Maybe move it some place else.

[noyau (Ping after 24h), 2012-09-20 12:39:24]

PTAL:

Moved the test again and enabled an existing one for xcode as it was testing the
same issue.

[Nico, 2012-09-20 12:52:02]

lgtm

https://chromiumcodereview.appspot.com/10911280/diff/23001/test/same-gyp-name...
File test/same-gyp-name/gyptest-all.py (right):

https://chromiumcodereview.appspot.com/10911280/diff/23001/test/same-gyp-name...
test/same-gyp-name/gyptest-all.py:33: chdir1 = chdir2 = 'relocate/src'
Do you understand why this is necessary? Other tests (test/subdir) don't seem to
need it, and I thought the test runner should hide differences like this…

(If not, no problem. Maybe add a "# TODO: Not clear why this is needed")

[noyau (Ping after 24h), 2012-09-20 13:27:30]

On 2012/09/20 12:52:02, Nico wrote:
> lgtm
> 
>
https://chromiumcodereview.appspot.com/10911280/diff/23001/test/same-gyp-name...
> File test/same-gyp-name/gyptest-all.py (right):
> 
>
https://chromiumcodereview.appspot.com/10911280/diff/23001/test/same-gyp-name...
> test/same-gyp-name/gyptest-all.py:33: chdir1 = chdir2 = 'relocate/src'
> Do you understand why this is necessary? Other tests (test/subdir) don't seem
to
> need it, and I thought the test runner should hide differences like this…
> 
> (If not, no problem. Maybe add a "# TODO: Not clear why this is needed")

subdir is doing exactly the same : checking the test format and changing chdir
to match. See test/subdirectory/gyptest-top-all.py:35. Now do I understand why
this is necessary, I'm not so sure.

[Nico, 2012-09-20 13:31:54]

> subdir is doing exactly the same : checking the test format and changing chdir
> to match. See test/subdirectory/gyptest-top-all.py:35. Now do I understand why
> this is necessary, I'm not so sure.

Ah ok. Thanks for checking, and thanks for your patience. Do you have commit
rights, or do you want me to land this?

[noyau (Ping after 24h), 2012-09-20 13:41:44]

On 2012/09/20 13:31:54, Nico wrote:
> > subdir is doing exactly the same : checking the test format and changing
chdir
> > to match. See test/subdirectory/gyptest-top-all.py:35. Now do I understand
why
> > this is necessary, I'm not so sure.
> 
> Ah ok. Thanks for checking, and thanks for your patience. Do you have commit
> rights, or do you want me to land this?

I do not have commit rights on gyp. If the commit checkbox is not connected to a
CQ, then yes, I need you to land it.

[Nico, 2012-09-21 04:55:12]

r1504. Thanks again!