Change the gyp_chromium to build chromium in a symbolic link

Change the gyp_chromium to build chromium in a symbolic link

If the path of chromium source is a symbolic link,
os.getcwd() returns the real path but __file__ will include
the symbolic link. It causes gyp failures due to the inconsistency.

BUG=
TEST=


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=146172

[tony, 2012-07-09 19:29:58]

You have to use the Publish+Mail Comments link on the left side to notify
reviewers of your patch.

This is probably OK.

In your change description, can you be explicit about what directory is a
symlink?

[M-A Ruel, 2012-07-09 19:32:14]

http://codereview.chromium.org/10692086/diff/1/build/gyp_chromium
File build/gyp_chromium (right):

http://codereview.chromium.org/10692086/diff/1/build/gyp_chromium#newcode16
build/gyp_chromium:16: script_dir = os.path.realpath(os.path.dirname(__file__))
The correct way of doing it is:
script_dir = os.path.dirname(os.path.abspath(__file__))

[yongsheng, 2012-07-11 01:29:47]

On 2012/07/09 19:29:58, tony wrote:
> You have to use the Publish+Mail Comments link on the left side to notify
> reviewers of your patch.
> 
> This is probably OK.
> 
> In your change description, can you be explicit about what directory is a
> symlink?
ok, I'll.

[yongsheng, 2012-07-11 01:32:46]

On 2012/07/09 19:32:14, Marc-Antoine Ruel wrote:
> http://codereview.chromium.org/10692086/diff/1/build/gyp_chromium
> File build/gyp_chromium (right):
> 
> http://codereview.chromium.org/10692086/diff/1/build/gyp_chromium#newcode16
> build/gyp_chromium:16: script_dir =
os.path.realpath(os.path.dirname(__file__))
> The correct way of doing it is:
> script_dir = os.path.dirname(os.path.abspath(__file__))
thanks. I'll refine it.
Actually, this issue is found when I built chrome for android. I downloaded the
chromium source code and android source code. In the next, I created a symbolic
link in ${android-sourcecode}/external/chrome, which linked to real directory
where chromium source code was placed. Then it got failures.

[yongsheng, 2012-07-11 02:21:25]

On 2012/07/09 19:32:14, Marc-Antoine Ruel wrote:
> http://codereview.chromium.org/10692086/diff/1/build/gyp_chromium
> File build/gyp_chromium (right):
> 
> http://codereview.chromium.org/10692086/diff/1/build/gyp_chromium#newcode16
> build/gyp_chromium:16: script_dir =
os.path.realpath(os.path.dirname(__file__))
> The correct way of doing it is:
> script_dir = os.path.dirname(os.path.abspath(__file__))
sorry, this doesn't work for some cases.
If you run gyp_chromium like this ./build/gyp_chromium, then __file__ is a
relative path and os.path.abspath can generate the real path(not symbolic path).
If you run gyp_chromium like this ${CHROMIUM_SRC}/build/gyp_chromium( absolute
path, android run it in this way), then __file__ is an absolute path and
os.path.abspath does nothing so it's still a symbolic path. so it fails.

[tony, 2012-07-11 16:34:07]

LGTM.  Marc-Antoine, is it OK with you too?

[M-A Ruel, 2012-07-11 16:35:06]

On 2012/07/11 16:34:07, tony wrote:
> LGTM.  Marc-Antoine, is it OK with you too?

lgtm, since os.path.realpath() is a superset of os.path.abspath().

[commit-bot: I haz the power, 2012-07-11 18:05:49]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/yongsheng.zhu@intel.com/10692086/7001

[commit-bot: I haz the power, 2012-07-11 19:34:24]

Change committed as 146172

[yongsheng, 2012-07-12 01:18:26]

On 2012/07/11 16:35:06, Marc-Antoine Ruel wrote:
> On 2012/07/11 16:34:07, tony wrote:
> > LGTM.  Marc-Antoine, is it OK with you too?
> 
> lgtm, since os.path.realpath() is a superset of os.path.abspath().

thanks very much for you guys.