Restart browser between smoke tests, fixes failing webview bypass test.

Restart browser between smoke tests, fixes failing webview bypass test.

The bypass smoke test uses an existing browser session, where the favicon requewst has already been made and completed through the data saver proxy.  When the favicon is requested again during the bypass test, it will still appear to come through the proxy causing the case to fail.  If instead we use a fresh browser session, bypass works as expected for the favicon.

In general, it's good practice to make tests as atomic as possible, so this should be better design for the smoke tests (although it adds ~12 seconds to the total execution time).

BUG=452161
Committed: https://crrev.com/52dc983774bcd3262472f025d0429e10799c3f80
Cr-Commit-Position: refs/heads/master@{#313153}

[bustamante, 2015-01-26 20:17:12]

bustamante@chromium.org changed reviewers:
+ bengr@chromium.org, sclittle@chromium.org

[sclittle, 2015-01-26 21:08:17]

https://codereview.chromium.org/862803004/diff/20001/tools/chrome_proxy/integ...
File tools/chrome_proxy/integration_tests/chrome_proxy_pagesets/smoke.py
(right):

https://codereview.chromium.org/862803004/diff/20001/tools/chrome_proxy/integ...
tools/chrome_proxy/integration_tests/chrome_proxy_pagesets/smoke.py:14:
self.restart_after = True
Could you specify this in chrome_proxy_measurements.py instead, by having
ChromeProxySmoke.__init__ set restart_after_each_page=True? That way it'll be
more consistent with the other tests.

[bustamante, 2015-01-26 21:08:21]

This should fix the failing smoke tests on Webview waterfall, please take a look
when you get a chance.

[bustamante, 2015-01-26 21:26:43]

On 2015/01/26 21:08:17, sclittle wrote:
>
https://codereview.chromium.org/862803004/diff/20001/tools/chrome_proxy/integ...
> File tools/chrome_proxy/integration_tests/chrome_proxy_pagesets/smoke.py
> (right):
> 
>
https://codereview.chromium.org/862803004/diff/20001/tools/chrome_proxy/integ...
> tools/chrome_proxy/integration_tests/chrome_proxy_pagesets/smoke.py:14:
> self.restart_after = True
> Could you specify this in chrome_proxy_measurements.py instead, by having
> ChromeProxySmoke.__init__ set restart_after_each_page=True? That way it'll be
> more consistent with the other tests.

Yeah, Done.

[sclittle, 2015-01-26 21:29:14]

LGTM

[bengr, 2015-01-26 21:40:23]

lgtm

https://codereview.chromium.org/862803004/diff/20001/tools/chrome_proxy/integ...
File tools/chrome_proxy/integration_tests/chrome_proxy_pagesets/smoke.py
(right):

https://codereview.chromium.org/862803004/diff/20001/tools/chrome_proxy/integ...
tools/chrome_proxy/integration_tests/chrome_proxy_pagesets/smoke.py:14:
self.restart_after = True
On 2015/01/26 21:08:17, sclittle wrote:
> Could you specify this in chrome_proxy_measurements.py instead, by having
> ChromeProxySmoke.__init__ set restart_after_each_page=True? That way it'll be
> more consistent with the other tests.

+1

[bustamante, 2015-01-26 21:41:57]

The CQ bit was checked by bustamante@chromium.org

[commit-bot: I haz the power, 2015-01-26 21:42:53]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/862803004/60001

[commit-bot: I haz the power, 2015-01-26 22:31:35]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2015-01-26 22:32:15]

Patchset 4 (id:??) landed as
https://crrev.com/52dc983774bcd3262472f025d0429e10799c3f80
Cr-Commit-Position: refs/heads/master@{#313153}