sync: temp fix for ShowErrorUI

sync: temp fix for ShowErrorUI

If there's an auth error, and the user clicks a signin link, show the login dialog straight away.
Before this, the code would make a generic "ShowSetup" call with "NONFATAL_ERROR", resulting
in infinite spin.

BUG=91286, 92265
TEST=change gaia password for active sync account that is ASP enabled. observe error in UI, sign back in. dialog should dismiss on success.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=96459

[tim (not reviewing), 2011-08-11 01:14:04]

ping!

[tim (not reviewing), 2011-08-11 18:01:04]

On 2011/08/11 01:14:04, timsteele wrote:
> ping!

note: this is a pretty tiny fix I'd like to merge to m14.

[Andrew T Wilson (Slow), 2011-08-11 18:13:04]

On 2011/08/11 18:01:04, timsteele wrote:
> On 2011/08/11 01:14:04, timsteele wrote:
> > ping!
> 
> note: this is a pretty tiny fix I'd like to merge to m14.

LGTM. I think we need a more extensive fix for m15, but this is good for m14.

[binji, 2011-08-11 18:15:38]

Sorry Tim, I was out sick yesterday and today.

LGTM as long as it doesn't break the passphrase error flow. (though I don't have
commit bit yet)...

http://codereview.chromium.org/7590018/diff/1/chrome/browser/sync/profile_syn...
File chrome/browser/sync/profile_sync_service.cc (right):

http://codereview.chromium.org/7590018/diff/1/chrome/browser/sync/profile_syn...
chrome/browser/sync/profile_sync_service.cc:815: if (last_auth_error_.state() !=
AuthError::NONE)
Will this branch be taken if there is a passphrase error?

[tim (not reviewing), 2011-08-11 18:18:35]

Thanks guys!  Agree we should tighten some screws for m15... I'll look at bug
91065 and try to summarize thoughts...

http://codereview.chromium.org/7590018/diff/1/chrome/browser/sync/profile_syn...
File chrome/browser/sync/profile_sync_service.cc (right):

http://codereview.chromium.org/7590018/diff/1/chrome/browser/sync/profile_syn...
chrome/browser/sync/profile_sync_service.cc:815: if (last_auth_error_.state() !=
AuthError::NONE)
On 2011/08/11 18:15:39, binji wrote:
> Will this branch be taken if there is a passphrase error?

Only if there is also an authentication error, in which case I think it's okay
to show the auth error first.