Revert of Revert "gclient: Change the .gclient URL mismatch warning into an error."

Revert of Revert "gclient: Change the .gclient URL mismatch warning into an error." (https://codereview.chromium.org/213483003/)

Reason for revert:
Re-landing this after https://codereview.chromium.org/214813002/

Original issue's description:
> Revert "gclient: Change the .gclient URL mismatch warning into an error."
> 
> This reverts commit c01bff751856165467bd67abb3757a701deaa375.
> 
> BUG=356881
> TBR=iannucci
> 
> Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=259727

TBR=iannucci@chromium.org,smut@google.com
NOTREECHECKS=true
NOTRY=true
BUG=356881
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=261158

[borenet, 2014-04-01 20:37:13]

Created Revert of Revert "gclient: Change the .gclient URL mismatch warning into
an error."

[borenet, 2014-04-01 20:37:17]

The CQ bit was unchecked by borenet@google.com

[smut, 2014-04-01 20:58:30]

I think the warning has stopped appearing on our bots using, so lgtm (but I'm
not a committer).

[iannucci, 2014-04-02 17:34:47]

take 2: lgtm

[borenet, 2014-04-02 19:10:22]

The CQ bit was checked by borenet@google.com

[commit-bot: I haz the power, 2014-04-02 19:10:23]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/borenet@google.com/221423004/1

[commit-bot: I haz the power, 2014-04-02 19:12:14]

Change committed as 261158

[v.putkinen, 2014-04-08 09:26:50]

This change breaks the cache_dir mechanism of gclient. If cache_dir is used in
.gclient, the URLs will always mismatch since the checkout's remote URL points
to cache_dir. There probably should be a special case for cache_dir in
GClient._CheckConfig. 

Or perhaps SCMWrapper.GetActualRemoteURL should be aware of cache_dir and return
the actual URL. For example, it could follow .git/objects/info/alternates and
return the remote URL from there.

Btw, is this a proper place for this discussion? If not, please redirect me to
the correct place. :)

[borenet, 2014-04-08 15:30:25]

On 2014/04/08 09:26:50, v.putkinen wrote:
> This change breaks the cache_dir mechanism of gclient. If cache_dir is used in
> .gclient, the URLs will always mismatch since the checkout's remote URL points
> to cache_dir. There probably should be a special case for cache_dir in
> GClient._CheckConfig. 
> 
> Or perhaps SCMWrapper.GetActualRemoteURL should be aware of cache_dir and
return
> the actual URL. For example, it could follow .git/objects/info/alternates and
> return the remote URL from there.
> 
> Btw, is this a proper place for this discussion? If not, please redirect me to
> the correct place. :)

I filed https://code.google.com/p/chromium/issues/detail?id=361155 and uploaded
https://codereview.chromium.org/228793002 to implement your second suggestion.

[borenet, 2014-04-08 20:03:32]

On 2014/04/08 15:30:25, borenet wrote:
> On 2014/04/08 09:26:50, v.putkinen wrote:
> > This change breaks the cache_dir mechanism of gclient. If cache_dir is used
in
> > .gclient, the URLs will always mismatch since the checkout's remote URL
points
> > to cache_dir. There probably should be a special case for cache_dir in
> > GClient._CheckConfig. 
> > 
> > Or perhaps SCMWrapper.GetActualRemoteURL should be aware of cache_dir and
> return
> > the actual URL. For example, it could follow .git/objects/info/alternates
and
> > return the remote URL from there.
> > 
> > Btw, is this a proper place for this discussion? If not, please redirect me
to
> > the correct place. :)
> 
> I filed https://code.google.com/p/chromium/issues/detail?id=361155 and
uploaded
> https://codereview.chromium.org/228793002 to implement your second suggestion.

I think this is fixed now, can you verify?