Fixed flaky StunRequestTests which depended on the wall clock

Fixed flaky StunRequestTests which depended on the wall clock

StunRequestTests were using the real time clock to measure fairly large
retransmit intervals (up to several seconds). This was making the tests
slow and flaky when the system was heavily loaded.

See https://build.chromium.org/p/client.webrtc/builders/Win64%20Release/builds/9274/steps/rtc_unittests/logs/stdio
for an example of a recent failure.

This change makes the tests use a simulated clock instead. They are now
very quick, precise and reliable.

R=honghaiz@webrtc.org, zhihuang@webrtc.org

Committed: https://chromium.googlesource.com/external/webrtc/+/b9d8d10d429150dfac0128cb20355b3abc86d406

[skvlad, 2016-09-02 21:36:18]

Description was changed from

==========
Fixed flaky StunRequestTests which depended on the wall clock

StunRequestTests were using the real time clock to measure fairly large
retransmit intervals (up to several seconds). This was making the tests
slow and flaky when the system was heavily loaded.

See
https://build.chromium.org/p/client.webrtc/builders/Win64%20Release/builds/92...
for an example of a recent failure.

This change makes the tests use a simulated clock instead. They are now
very quick, precise and reliable.
==========

to

==========
Fixed flaky StunRequestTests which depended on the wall clock

StunRequestTests were using the real time clock to measure fairly large
retransmit intervals (up to several seconds). This was making the tests
slow and flaky when the system was heavily loaded.

See
https://build.chromium.org/p/client.webrtc/builders/Win64%20Release/builds/92...
for an example of a recent failure.

This change makes the tests use a simulated clock instead. They are now
very quick, precise and reliable.
==========

[skvlad, 2016-09-02 21:36:18]

skvlad@webrtc.org changed reviewers:
+ mflodman@webrtc.org

[Zhi Huang, 2016-09-06 18:34:40]

zhihuang@webrtc.org changed reviewers:
+ zhihuang@webrtc.org

[Zhi Huang, 2016-09-06 18:34:40]

lgtm

[honghaiz3, 2016-09-06 21:26:26]

honghaiz@webrtc.org changed reviewers:
+ honghaiz@webrtc.org

[honghaiz3, 2016-09-06 21:26:27]

https://codereview.webrtc.org/2300143005/diff/20001/webrtc/p2p/base/stunreque...
File webrtc/p2p/base/stunrequest_unittest.cc (right):

https://codereview.webrtc.org/2300143005/diff/20001/webrtc/p2p/base/stunreque...
webrtc/p2p/base/stunrequest_unittest.cc:158: }
I think the while loop can be replaced with 
SIMULATED_WAIT(request_count_ != i, A_LONG_WAIT_TIME, fake_clock);

[skvlad, 2016-09-06 21:50:29]

https://codereview.webrtc.org/2300143005/diff/20001/webrtc/p2p/base/stunreque...
File webrtc/p2p/base/stunrequest_unittest.cc (right):

https://codereview.webrtc.org/2300143005/diff/20001/webrtc/p2p/base/stunreque...
webrtc/p2p/base/stunrequest_unittest.cc:158: }
On 2016/09/06 21:26:27, honghaiz3 wrote:
> I think the while loop can be replaced with 
> SIMULATED_WAIT(request_count_ != i, A_LONG_WAIT_TIME, fake_clock);

Thanks for the suggestion! Didn't know about it. I've changed it to use
SIMULATED_WAIT.

[honghaiz3, 2016-09-06 22:00:00]

lgtm

https://codereview.webrtc.org/2300143005/diff/40001/webrtc/p2p/base/stunreque...
File webrtc/p2p/base/stunrequest_unittest.cc (right):

https://codereview.webrtc.org/2300143005/diff/40001/webrtc/p2p/base/stunreque...
webrtc/p2p/base/stunrequest_unittest.cc:179: SIMULATED_WAIT(false, 9500,
fake_clock);
In case that this takes too long to run, it may be OK to use AdvanceTime by
100ms every time, although this is closer to the original test.

[skvlad, 2016-09-06 22:16:35]

The CQ bit was checked by skvlad@webrtc.org

[skvlad, 2016-09-06 22:16:36]

The patchset sent to the CQ was uploaded after l-g-t-m from zhihuang@webrtc.org
Link to the patchset: https://codereview.webrtc.org/2300143005/#ps40001 (title:
"Used SIMULATED_WAIT as advised by honghaiz@")

[commit-bot: I haz the power, 2016-09-06 22:16:46]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.webrtc.org/...

[commit-bot: I haz the power, 2016-09-06 23:50:08]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-06 23:50:09]

Try jobs failed on following builders:
  linux_baremetal on master.tryserver.webrtc (JOB_TIMED_OUT, no build URL)

[skvlad, 2016-09-07 00:19:02]

Description was changed from

==========
Fixed flaky StunRequestTests which depended on the wall clock

StunRequestTests were using the real time clock to measure fairly large
retransmit intervals (up to several seconds). This was making the tests
slow and flaky when the system was heavily loaded.

See
https://build.chromium.org/p/client.webrtc/builders/Win64%20Release/builds/92...
for an example of a recent failure.

This change makes the tests use a simulated clock instead. They are now
very quick, precise and reliable.
==========

to

==========
Fixed flaky StunRequestTests which depended on the wall clock

StunRequestTests were using the real time clock to measure fairly large
retransmit intervals (up to several seconds). This was making the tests
slow and flaky when the system was heavily loaded.

See
https://build.chromium.org/p/client.webrtc/builders/Win64%20Release/builds/92...
for an example of a recent failure.

This change makes the tests use a simulated clock instead. They are now
very quick, precise and reliable.

R=honghaiz@webrtc.org, zhihuang@webrtc.org

Committed:
https://chromium.googlesource.com/external/webrtc/+/b9d8d10d429150dfac0128cb2...
==========

[commit-bot: I haz the power, 2016-09-07 00:19:03]

Description was changed from

==========
Fixed flaky StunRequestTests which depended on the wall clock

StunRequestTests were using the real time clock to measure fairly large
retransmit intervals (up to several seconds). This was making the tests
slow and flaky when the system was heavily loaded.

See
https://build.chromium.org/p/client.webrtc/builders/Win64%20Release/builds/92...
for an example of a recent failure.

This change makes the tests use a simulated clock instead. They are now
very quick, precise and reliable.

R=honghaiz@webrtc.org, zhihuang@webrtc.org

Committed:
https://chromium.googlesource.com/external/webrtc/+/b9d8d10d429150dfac0128cb2...
==========

to

==========
Fixed flaky StunRequestTests which depended on the wall clock

StunRequestTests were using the real time clock to measure fairly large
retransmit intervals (up to several seconds). This was making the tests
slow and flaky when the system was heavily loaded.

See
https://build.chromium.org/p/client.webrtc/builders/Win64%20Release/builds/92...
for an example of a recent failure.

This change makes the tests use a simulated clock instead. They are now
very quick, precise and reliable.

R=honghaiz@webrtc.org, zhihuang@webrtc.org

Committed: https://crrev.com/b9d8d10d429150dfac0128cb20355b3abc86d406
Cr-Commit-Position: refs/heads/master@{#14097}
==========

[skvlad, 2016-09-07 00:19:03]

Committed patchset #4 (id:60001) manually as
b9d8d10d429150dfac0128cb20355b3abc86d406 (presubmit successful).