Move LoginPromptTest.TestCancelAuth to a browsertest.

Move LoginPromptTest.TestCancelAuth to a browsertest.

BUG=90198
TEST=browser_tests --gtest_filter=LoginPromptBrowserTest.TestCancelAuth.


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=106544

[cbentzel, 2011-10-19 19:03:12]

Thanks for fixing one of the top flaky tests.

http://codereview.chromium.org/8316022/diff/7001/chrome/browser/ui/login/logi...
File chrome/browser/ui/login/login_prompt_browsertest.cc (right):

http://codereview.chromium.org/8316022/diff/7001/chrome/browser/ui/login/logi...
chrome/browser/ui/login/login_prompt_browsertest.cc:168:
remaining_notification_count_(notification_count) {}
Do you need notification count? Is the sequence of LOAD_STOPs consistent enough
that you could just use two of the old-style WindowedLoadStopObserver's? Or is
this simpler?

[asanka, 2011-10-19 19:21:17]

http://codereview.chromium.org/8316022/diff/7001/chrome/browser/ui/login/logi...
File chrome/browser/ui/login/login_prompt_browsertest.cc (right):

http://codereview.chromium.org/8316022/diff/7001/chrome/browser/ui/login/logi...
chrome/browser/ui/login/login_prompt_browsertest.cc:168:
remaining_notification_count_(notification_count) {}
On 2011/10/19 19:03:12, cbentzel wrote:
> Do you need notification count? Is the sequence of LOAD_STOPs consistent
enough
> that you could just use two of the old-style WindowedLoadStopObserver's? Or is
> this simpler?

In the TestCancelAuth test, when we initiate a navigation to an unauthenticated
page while an auth prompt is being displayed, we will see AUTH_CANCELLED,
LOAD_STOP (for the previous page), LOAD_START (for the new page) and a LOAD_STOP
(for the new page).  We don't have a good synchronization point between the
LOAD_STOP events to register for the next LOAD_STOP.  So I elected to wait for
both together.

[cbentzel, 2011-10-19 19:30:29]

On 2011/10/19 19:21:17, asanka wrote:
>
http://codereview.chromium.org/8316022/diff/7001/chrome/browser/ui/login/logi...
> File chrome/browser/ui/login/login_prompt_browsertest.cc (right):
> 
>
http://codereview.chromium.org/8316022/diff/7001/chrome/browser/ui/login/logi...
> chrome/browser/ui/login/login_prompt_browsertest.cc:168:
> remaining_notification_count_(notification_count) {}
> On 2011/10/19 19:03:12, cbentzel wrote:
> > Do you need notification count? Is the sequence of LOAD_STOPs consistent
> enough
> > that you could just use two of the old-style WindowedLoadStopObserver's? Or
is
> > this simpler?
> 
> In the TestCancelAuth test, when we initiate a navigation to an
unauthenticated
> page while an auth prompt is being displayed, we will see AUTH_CANCELLED,
> LOAD_STOP (for the previous page), LOAD_START (for the new page) and a
LOAD_STOP
> (for the new page).  We don't have a good synchronization point between the
> LOAD_STOP events to register for the next LOAD_STOP.  So I elected to wait for
> both together.

Makes sense.

LGTM

[commit-bot: I haz the power, 2011-10-20 17:09:21]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/asanka@chromium.org/8316022/7001

[commit-bot: I haz the power, 2011-10-20 19:37:16]

Change committed as 106544

[Paweł Hajdan Jr., 2011-10-21 08:49:04]

http://codereview.chromium.org/8316022/diff/7001/chrome/browser/ui/login/logi...
File chrome/browser/ui/login/login_prompt_browsertest.cc (right):

http://codereview.chromium.org/8316022/diff/7001/chrome/browser/ui/login/logi...
chrome/browser/ui/login/login_prompt_browsertest.cc:162: class
WindowedLoadStopObserver
Could you please generalize it to chrome/test?