Fix a predictor browsertest for the PreconnectMore experiment

Fix some predictor browser tests

Some of the official builds are breaking, so this is a
speculative fix for bots that do not have the
"PreconnectMore" experimental config.

BUG=612106
Committed: https://crrev.com/9f7e92f1a88efa3a6a1c64c9df94bd8beeb0fda2
Cr-Commit-Position: refs/heads/master@{#393903}

[Charlie Harrison, 2016-05-16 18:56:22]

Description was changed from

==========
Fix a flaky predictor browser test

BUG=
==========

to

==========
Fix some predictor browser tests

Some of the official builds are breaking, so this is a 
speculative fix for bots that do not have the
"PreconnectMore" experimental config.

Additionally, DrMemory builds break on the
Predictor TwoNavigations2 test, so this change includes a
small change to try to reduce a potential race.

BUG=612106
==========

[Charlie Harrison, 2016-05-16 18:56:22]

csharrison@chromium.org changed reviewers:
+ mmenke@chromium.org

[Charlie Harrison, 2016-05-16 19:03:30]

Description was changed from

==========
Fix some predictor browser tests

Some of the official builds are breaking, so this is a 
speculative fix for bots that do not have the
"PreconnectMore" experimental config.

Additionally, DrMemory builds break on the
Predictor TwoNavigations2 test, so this change includes a
small change to try to reduce a potential race.

BUG=612106
==========

to

==========
Fix some predictor browser tests

Some of the official builds are breaking, so this is a 
speculative fix for bots that do not have the
"PreconnectMore" experimental config.

BUG=612106
==========

[Charlie Harrison, 2016-05-16 19:04:35]

Matt, can you take a quick look at this?

crbug.com/612106 shows that the new renderer initated tests are failing with
timeouts. This indicates to me that they are not using the finch test configs.
I've added the enable_features=PreconnectMore command line flag so the
experiment is always running for the browser tests regardless of finch config.

Additionally, I've updated a PostTask to a PostTaskAndReply to reduce possible
races.

[mmenke, 2016-05-16 19:10:08]

LGTM

https://codereview.chromium.org/1984583002/diff/40001/chrome/browser/net/pred...
File chrome/browser/net/predictor_browsertest.cc (right):

https://codereview.chromium.org/1984583002/diff/40001/chrome/browser/net/pred...
chrome/browser/net/predictor_browsertest.cc:599: base::RunLoop run_loop;
These shouldn't be needed - anything that does any network activity on the IO
thread triggered after this point will necessarily happen after this event.  The
only concern would be if we tore down the test fixture before the task runs. 
Since we always wait on some IO thread task, shouldn't be a problem there.  So
this part of the CL is fine, but I don't think it fixes anything.

[Charlie Harrison, 2016-05-16 19:13:39]

https://codereview.chromium.org/1984583002/diff/40001/chrome/browser/net/pred...
File chrome/browser/net/predictor_browsertest.cc (right):

https://codereview.chromium.org/1984583002/diff/40001/chrome/browser/net/pred...
chrome/browser/net/predictor_browsertest.cc:599: base::RunLoop run_loop;
On 2016/05/16 19:10:07, mmenke wrote:
> These shouldn't be needed - anything that does any network activity on the IO
> thread triggered after this point will necessarily happen after this event. 
The
> only concern would be if we tore down the test fixture before the task runs. 
> Since we always wait on some IO thread task, shouldn't be a problem there.  So
> this part of the CL is fine, but I don't think it fixes anything.

Ack. Thanks for the clarification. I'll keep it for test simplicity (readers
don't have to know that implementation detail).

[Charlie Harrison, 2016-05-16 19:13:44]

The CQ bit was checked by csharrison@chromium.org

[commit-bot: I haz the power, 2016-05-16 19:14:19]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1984583002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1984583002/40001

[commit-bot: I haz the power, 2016-05-16 19:42:38]

Description was changed from

==========
Fix some predictor browser tests

Some of the official builds are breaking, so this is a 
speculative fix for bots that do not have the
"PreconnectMore" experimental config.

BUG=612106
==========

to

==========
Fix some predictor browser tests

Some of the official builds are breaking, so this is a 
speculative fix for bots that do not have the
"PreconnectMore" experimental config.

BUG=612106
==========

[commit-bot: I haz the power, 2016-05-16 19:42:39]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-05-16 19:44:22]

Description was changed from

==========
Fix some predictor browser tests

Some of the official builds are breaking, so this is a 
speculative fix for bots that do not have the
"PreconnectMore" experimental config.

BUG=612106
==========

to

==========
Fix some predictor browser tests

Some of the official builds are breaking, so this is a
speculative fix for bots that do not have the
"PreconnectMore" experimental config.

BUG=612106

Committed: https://crrev.com/9f7e92f1a88efa3a6a1c64c9df94bd8beeb0fda2
Cr-Commit-Position: refs/heads/master@{#393903}
==========

[commit-bot: I haz the power, 2016-05-16 19:44:23]

Patchset 3 (id:??) landed as
https://crrev.com/9f7e92f1a88efa3a6a1c64c9df94bd8beeb0fda2
Cr-Commit-Position: refs/heads/master@{#393903}