Check to see if a dependency was skipped in custom_deps, and if so, skip it in

Check to see if a dependency was skipped in custom_deps, and if so, skip it in
recurse.

R=maruel@chromium.org
BUG=

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=212373

[M-A Ruel, 2013-07-18 15:11:43]

https://codereview.chromium.org/19670012/diff/1/gclient.py
File gclient.py (right):

https://codereview.chromium.org/19670012/diff/1/gclient.py#newcode710
gclient.py:710: elif os.path.isdir(cwd):
Shouldn't this check be sufficient? I mean if the directory doesn't exist, the
dependency wasn't fetched.

[iannucci, 2013-07-18 18:29:14]

On 2013/07/18 15:11:43, Marc-Antoine Ruel wrote:
> https://codereview.chromium.org/19670012/diff/1/gclient.py
> File gclient.py (right):
> 
> https://codereview.chromium.org/19670012/diff/1/gclient.py#newcode710
> gclient.py:710: elif os.path.isdir(cwd):
> Shouldn't this check be sufficient? I mean if the directory doesn't exist, the
> dependency wasn't fetched.

It's insufficient when something else mkdir's the directory. I think this is
happening in gyp to make some of it's scripting assumptions easier. In any
event, we should be using DEPS as truth, not relying on the state of the
filesystem to record the fact that we probably didn't fetch something...

[M-A Ruel, 2013-07-18 18:33:17]

On 2013/07/18 18:29:14, iannucci wrote:
> It's insufficient when something else mkdir's the directory. I think this is
> happening in gyp to make some of it's scripting assumptions easier. In any
> event, we should be using DEPS as truth, not relying on the state of the
> filesystem to record the fact that we probably didn't fetch something...

ok, lgtm :)

[commit-bot: I haz the power, 2013-07-18 18:35:31]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/iannucci@chromium.org/19670012/1

[commit-bot: I haz the power, 2013-07-18 18:37:51]

Change committed as 212373