Make the login WebUI listen NOTIFICATION_AUTH_SUPPLIED and update the system request context.

Make the login WebUI listen NOTIFICATION_AUTH_SUPPLIED and update the system request context.

Possibly the user has authenticated against a proxy server and we might
need the credentials for enrollment and other system requests from the
main |g_browser_process| request context (see bug http://crosbug.com/24861)
So we transfer any credentials to the global request context here.

BUG=chromium-os:24861
TEST=Manual as described in the bug. (Requires infrastructure setup.)

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=117630

[pastarmovj, 2012-01-12 17:59:48]

Hey Ivan,
I have another small CL and will be grateful if you can review it. This time
fixing an issue with proxy credentials not being transferred to the system
request context used by the enterprise enrollment and policy fetching code.
Nikita suggested you were the right to review that change as you have also
worked on proxy related code recently.
Thanks,
Julian
P.S. This is rather a short term solution as we have to solve a much bigger
problem with proxies requiring authentication not working for many other things
but we have to solve this more or less ASAP to be able to ship Chromebooks for
enterprises which will have such setups quite often.

[Ivan Korotkov, 2012-01-12 19:37:15]

LGTM with nits:

http://codereview.chromium.org/9186039/diff/1/chrome/browser/chromeos/login/e...
File chrome/browser/chromeos/login/existing_user_controller.cc (right):

http://codereview.chromium.org/9186039/diff/1/chrome/browser/chromeos/login/e...
chrome/browser/chromeos/login/existing_user_controller.cc:104: LOG(INFO) <<
"Main request context populated with authentication data.";
Should it really be an INFO log? Probably more appropriate to replace with VLOG
or even DVLOG.

http://codereview.chromium.org/9186039/diff/1/chrome/browser/chromeos/login/e...
chrome/browser/chromeos/login/existing_user_controller.cc:224:
base::Unretained(default_profile)),
I've been told that it's deeply wrong to pass Profile to IO thread because it is
a purely UI-thread object. Also I'm not sure if it can't go away in 2 secs.
Please pass make_scoped_refptr(X->GetRequestContext()) for both arguments
instead.

http://codereview.chromium.org/9186039/diff/1/chrome/browser/chromeos/login/e...
chrome/browser/chromeos/login/existing_user_controller.cc:225: 2000);
Define a const for 2000, please :)

[pastarmovj, 2012-01-13 10:15:44]

Addressed the issues you suggested.

http://codereview.chromium.org/9186039/diff/1/chrome/browser/chromeos/login/e...
File chrome/browser/chromeos/login/existing_user_controller.cc (right):

http://codereview.chromium.org/9186039/diff/1/chrome/browser/chromeos/login/e...
chrome/browser/chromeos/login/existing_user_controller.cc:104: LOG(INFO) <<
"Main request context populated with authentication data.";
On 2012/01/12 19:37:16, Ivan Korotkov wrote:
> Should it really be an INFO log? Probably more appropriate to replace with
VLOG
> or even DVLOG.

Yes you are right I only wanted at least one piece of information printed as
INFO I changed it to VLOG here (it is still useful for release builds too to
track timing compared to other possible messages.

http://codereview.chromium.org/9186039/diff/1/chrome/browser/chromeos/login/e...
chrome/browser/chromeos/login/existing_user_controller.cc:224:
base::Unretained(default_profile)),
On 2012/01/12 19:37:16, Ivan Korotkov wrote:
> I've been told that it's deeply wrong to pass Profile to IO thread because it
is
> a purely UI-thread object. Also I'm not sure if it can't go away in 2 secs.
> Please pass make_scoped_refptr(X->GetRequestContext()) for both arguments
> instead. 

Good advice, thanks! Done as suggested and now I pass a RefCounted thread safe
object so we should be good even two seconds later.

http://codereview.chromium.org/9186039/diff/1/chrome/browser/chromeos/login/e...
chrome/browser/chromeos/login/existing_user_controller.cc:225: 2000);
On 2012/01/12 19:37:16, Ivan Korotkov wrote:
> Define a const for 2000, please :)

Done.

[Ivan Korotkov, 2012-01-13 10:20:27]

Final LGTM with a small naming nit:

http://codereview.chromium.org/9186039/diff/6001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/existing_user_controller.cc (right):

http://codereview.chromium.org/9186039/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/existing_user_controller.cc:84: const long int
kAuthCacheTransferDelay = 2000;
kAuthCacheTransferDelayMs