Fix ServiceWorkerCache::Keys() unittest memory leak.

Fix ServiceWorkerCache::Keys() unittest memory leak.

I thought that this was fixed (Linux Valgrind had passed the content test) but it turns out that the failure is flaky so I need to run it a few times to be sure.  Fix again and run the tests a few times to make sure they work.

If after submitting it's still not fixed I'll suppress it and fix it instead of reverting.

Reverted CL: https://codereview.chromium.org/543093003

BUG=392621
Committed: https://crrev.com/8b58d10c279f70f8ee31b6fd25be2d6830fc12b4
Cr-Commit-Position: refs/heads/master@{#294313}

[jkarlin, 2014-09-10 18:01:15]

jkarlin@chromium.org changed reviewers:
+ michaeln@chromium.org

[jkarlin, 2014-09-10 18:01:16]

michaeln: This has passed linux valgrind 3 times in a row now.  If it passes
chromeos too and you lgtm then I'll submit.  If it fails again I'll suppress
(since it's just a unittest teardown issue) and fix later.

[jkarlin, 2014-09-10 18:44:05]

On 2014/09/10 18:01:16, jkarlin wrote:
> michaeln: This has passed linux valgrind 3 times in a row now.  If it passes
> chromeos too and you lgtm then I'll submit.  If it fails again I'll suppress
> (since it's just a unittest teardown issue) and fix later.

Sorry, that message wasn't clear.  If (after submitting) any valgrind error
occurs I'll suppress.  I didn't mean to imply that I would submit if the
chromeos test failed.

[michaeln, 2014-09-10 19:38:22]

oh, simplecache creates a private sequencedworkerpool composed of up to 50
threads under the covers (that threadcount seems high). i thought the
implementation did all its work on the taskrunner provided to it by the caller.

Maybe a call to flush those tasks should be dropped into
BrowserThreadImpl::FlushThreadPoolHelper() or at the few callsites to that
method there should also be a call to
disk_cache::SimpleBackendImpl::FlushWorkerPoolForTesting(). Our test harnesses
are setup to use the BrowserThreadImpl method between test fixtures. If that
works maybe it could fix this sort of problem for all tests that involve
simplecache in some way. See
https://code.google.com/p/chromium/issues/detail?id=168415 for some related
history.

[jkarlin, 2014-09-10 20:23:19]

On 2014/09/10 19:38:22, michaeln wrote:
> oh, simplecache creates a private sequencedworkerpool composed of up to 50
> threads under the covers (that threadcount seems high). i thought the
> implementation did all its work on the taskrunner provided to it by the
caller.
> 
> Maybe a call to flush those tasks should be dropped into
> BrowserThreadImpl::FlushThreadPoolHelper() or at the few callsites to that
> method there should also be a call to
> disk_cache::SimpleBackendImpl::FlushWorkerPoolForTesting(). Our test harnesses
> are setup to use the BrowserThreadImpl method between test fixtures. If that
> works maybe it could fix this sort of problem for all tests that involve
> simplecache in some way. See
> https://code.google.com/p/chromium/issues/detail?id=168415 for some related
> history.

Looks as if it passed chromeos now too.  Given that it took me *all* day to get
the chromeos trybot to run to this point, is it okay if I submit this and look
into your suggestion in a follow-up CL?  That way I can land the next 4 patches
that are and ready to go.

[michaeln1, 2014-09-10 21:05:42]

lgtm

> Looks as if it passed chromeos now too.  Given that it took me *all* day to
get
> the chromeos trybot to run to this point, is it okay if I submit this and look
> into your suggestion in a follow-up CL?  That way I can land the next 4
patches
> that are and ready to go.
sgtm

[jkarlin, 2014-09-10 22:18:53]

The CQ bit was checked by jkarlin@chromium.org

[commit-bot: I haz the power, 2014-09-10 22:20:55]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/jkarlin@chromium.org/559003002/20001

[commit-bot: I haz the power, 2014-09-10 23:56:57]

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

[commit-bot: I haz the power, 2014-09-10 23:56:58]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux
(http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[jkarlin, 2014-09-11 00:14:38]

On 2014/09/10 23:56:58, I haz the power (commit-bot) wrote:
> Try jobs failed on following builders:
>   chromium_presubmit on tryserver.chromium.linux
>
(http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

michalen: Can you l-g-t-m from your other account?  Thanks!

[michaeln, 2014-09-11 00:45:17]

sry... lgtm 2

[jkarlin, 2014-09-11 00:51:22]

The CQ bit was checked by jkarlin@chromium.org

[commit-bot: I haz the power, 2014-09-11 00:55:31]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patchset/559003002/20001

[commit-bot: I haz the power, 2014-09-11 01:18:35]

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

[commit-bot: I haz the power, 2014-09-11 01:18:37]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux
(http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[jkarlin, 2014-09-11 01:32:27]

The CQ bit was checked by jkarlin@chromium.org

[commit-bot: I haz the power, 2014-09-11 01:39:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patchset/559003002/40001

[commit-bot: I haz the power, 2014-09-11 03:39:27]

Committed patchset #3 (id:40001) as 48fbcb0e5f0a01431fbdcac9015dbe477c0580c0

[commit-bot: I haz the power, 2014-09-11 03:40:22]

Patchset 3 (id:??) landed as
https://crrev.com/8b58d10c279f70f8ee31b6fd25be2d6830fc12b4
Cr-Commit-Position: refs/heads/master@{#294313}