Add dependencies to webkit pieces.

Add dependencies to webkit pieces.

libcompositor (.so) gets linked against libwebkit_fileapi (.a) and libwebkit_quota (.a), each of which is built with googleurl, but does not include it as a dependency, causing libcompositor (.so) to not include all of the symbols it requires.

R=kinuko@chromium.org
BUG=112626
TEST=Compiled chromeos with shared library on a buildbot like setup.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=132201

[Peter Mayo, 2012-04-12 01:37:47]

This uses the same style as I found googleurl referenced in other gypi files in
webkit.

kinuko added references to goggleurl in at least 76858 (
http://codereview.chromium.org/6609009 ) and happens to be an OWNER of both
directories.

I would request this review not be shallow, since I am not completely
comfortable with/sure of this. (especially where the effects outside of the
new/broken bots is concerned)

[Peter Mayo, 2012-04-13 01:42:52]

Note:  Previous try failures were based on troubles with the client/checkout
from which I issued them, not the change itself.

[kinuko, 2012-04-13 02:51:58]

On 2012/04/13 01:42:52, Peter Mayo wrote:
> Note:  Previous try failures were based on troubles with the client/checkout
> from which I issued them, not the change itself.

Sorry for the delay, I needed to look for other usage of temp_gyp.
LGTM thanks!

[Peter Mayo, 2012-04-13 04:09:18]

the linux chromeos failures look stable.

Does the linkage change the sharedness? Does having the same .a linked into 2
(or more) .so cause a conflict?  Not sure.

Going to try a clobber.

[kinuko, 2012-04-13 11:42:04]

Aren't they just flaky?  The last two try result doesn't look like a linkage
issue to me.

On 2012/04/13 04:09:18, Peter Mayo wrote:
> the linux chromeos failures look stable.
> 
> Does the linkage change the sharedness? Does having the same .a linked into 2
> (or more) .so cause a conflict?  Not sure.
> 
> Going to try a clobber.

[Peter Mayo, 2012-04-13 14:52:26]

On 2012/04/13 11:42:04, kinuko wrote:
> Aren't they just flaky?  The last two try result doesn't look like a linkage
> issue to me.
> 
> On 2012/04/13 04:09:18, Peter Mayo wrote:
> > the linux chromeos failures look stable.
> > 
> > Does the linkage change the sharedness? Does having the same .a linked into
2
> > (or more) .so cause a conflict?  Not sure.
> > 
> > Going to try a clobber.

http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_chro...
makes it look like flake, the other ones were all dying on the same test
(InstantTest.PreloadsInstant failing on being hidden when asked for paint), that
the same builders were passing for other try jobs, based on the local state, so
I was worried about singleton/static initializer issues.

[commit-bot: I haz the power, 2012-04-13 15:41:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/petermayo@chromium.org/9980022/4002

[commit-bot: I haz the power, 2012-04-13 17:08:59]

Change committed as 132201