Retry Chromoting browser-test execution if test times-out without launching browser.

Retry Chromoting browser-test execution if test times-out without launching browser.

BUG=480025
Committed: https://crrev.com/4a934c5a447493e153d68e0efbf191421a11c3e7
Cr-Commit-Position: refs/heads/master@{#333646}

[anandc, 2015-06-09 23:39:04]

anandc@chromium.org changed reviewers:
+ joedow@chromium.org

[joedow, 2015-06-09 23:59:32]

https://codereview.chromium.org/1172183003/diff/1/testing/chromoting/browser_...
File testing/chromoting/browser_tests_launcher.py (right):

https://codereview.chromium.org/1172183003/diff/1/testing/chromoting/browser_...
testing/chromoting/browser_tests_launcher.py:44: 
I think you should move the TestCaseSetup() call from line 217 up here.  It
would be cleaner if you have the original run and retry logic in a loop which
terminates when:
1.) Test passes
2.) Test fails for unexpected reason
3.) Test has failed with the time out issue the maximum number of times

This would clean up the flow and also allow you to tweak the re-run conditions
later on if needed.

https://codereview.chromium.org/1172183003/diff/1/testing/chromoting/browser_...
testing/chromoting/browser_tests_launcher.py:52: # eventually passes, we'll also
look for "(TIMED OUT)", before retrying.
Just curious, is there an error code we can key off of or is string compare the
only way to tell?

[anandc, 2015-06-10 01:02:12]

Thanks Joe. PTAL at patch-set #2.

https://codereview.chromium.org/1172183003/diff/1/testing/chromoting/browser_...
File testing/chromoting/browser_tests_launcher.py (right):

https://codereview.chromium.org/1172183003/diff/1/testing/chromoting/browser_...
testing/chromoting/browser_tests_launcher.py:44: 
On 2015/06/09 23:59:32, joedow wrote:
> I think you should move the TestCaseSetup() call from line 217 up here.  It
> would be cleaner if you have the original run and retry logic in a loop which
> terminates when:
> 1.) Test passes
> 2.) Test fails for unexpected reason
> 3.) Test has failed with the time out issue the maximum number of times
> 
> This would clean up the flow and also allow you to tweak the re-run conditions
> later on if needed.

Done.

https://codereview.chromium.org/1172183003/diff/1/testing/chromoting/browser_...
testing/chromoting/browser_tests_launcher.py:52: # eventually passes, we'll also
look for "(TIMED OUT)", before retrying.
On 2015/06/09 23:59:32, joedow wrote:
> Just curious, is there an error code we can key off of or is string compare
the
> only way to tell?

There doesn't seem to be a special error code for a time-out.
See here for results when a time-out occurs (search for "Return code = "):
https://chromium-swarm.appspot.com/user/task/27f3b5c2605bbd10

And here for results when a test fails for a reason other than a time-out:
https://chromium-swarm.appspot.com/user/task/27f3ba00601c8f10
Both return an error code of 1.

[joedow, 2015-06-10 01:32:00]

lgtm

[anandc, 2015-06-10 01:49:10]

The CQ bit was checked by anandc@chromium.org

[commit-bot: I haz the power, 2015-06-10 01:51:29]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1172183003/20001

[commit-bot: I haz the power, 2015-06-10 02:01:29]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-06-10 02:02:14]

Patchset 2 (id:??) landed as
https://crrev.com/4a934c5a447493e153d68e0efbf191421a11c3e7
Cr-Commit-Position: refs/heads/master@{#333646}