Delete watcher on IO thread in MessagePumpLibeventTest.QuitWatcher.

Delete watcher on IO thread in MessagePumpLibeventTest.QuitWatcher.

If StartWatching is run before the WaitableEvent is signaled,
the StopWatching should be called (on the IO thread). Otherwise,
a DCHECK will be called in the destructor of WaitableEventWatcher.

BUG=internal b/21093761
BUG=481290
Committed: https://crrev.com/2b94f7bf55796a7112bd8dc5fafb0797e8bc9949
Cr-Commit-Position: refs/heads/master@{#335451}

[derekjchow1, 2015-06-16 00:15:45]

derekjchow@chromium.org changed reviewers:
+ thakis@chromium.org

[derekjchow1, 2015-06-16 00:15:45]

Hi,

PTAL. The Chromecast team has been running into the following errors on our m43
based branch:

[ RUN      ] MessagePumpLibeventTest.QuitWatcher
[1859:1859:0615/170638:9002378412:FATAL:message_loop.cc(271)] Check failed: this
== current() (0xb8a06c30 vs. 0xbeb665b0)
[1/1] MessagePumpLibeventTest.QuitWatcher (CRASHED)

It seems to me there's a race condition on whether StartWatching gets called
first or if Signal is called first. Moving ownership of |watcher| to the io
thread seems to fix the problem. WDYT?

[derekjchow1, 2015-06-19 16:48:13]

Ping on this issue.

[Nico, 2015-06-19 22:10:57]

thakis@chromium.org changed reviewers:
+ chirantan@chromium.org, thestig@chromium.org

[Nico, 2015-06-19 22:10:58]

This sounds related to https://codereview.chromium.org/1059353003 /
https://code.google.com/p/chromium/issues/detail?id=481290

[Chirantan Ekbote, 2015-06-20 00:09:37]

Thanks for fixing this!  It's definitely related to crbug.com/481290 so could
you please also include this bug in the description?  Otherwise, lgtm with one
question.

https://codereview.chromium.org/1184093007/diff/1/base/message_loop/message_p...
File base/message_loop/message_pump_libevent_unittest.cc (right):

https://codereview.chromium.org/1184093007/diff/1/base/message_loop/message_p...
base/message_loop/message_pump_libevent_unittest.cc:271:
Unretained(watcher.get())));
Why not just use base::Passed here?

[derekjchow1, 2015-06-20 00:34:07]

https://codereview.chromium.org/1184093007/diff/1/base/message_loop/message_p...
File base/message_loop/message_pump_libevent_unittest.cc (right):

https://codereview.chromium.org/1184093007/diff/1/base/message_loop/message_p...
base/message_loop/message_pump_libevent_unittest.cc:271:
Unretained(watcher.get())));
On 2015/06/20 00:09:37, Chirantan Ekbote wrote:
> Why not just use base::Passed here?

I don't think we can Pass a scoped_ptr as the "this" argument to a class method.
That being said, we can probably use base::Owned.

https://code.google.com/p/chromium/codesearch#chromium/src/base/callback.h&q=...

[derekjchow1, 2015-06-20 00:38:35]

The CQ bit was checked by derekjchow@chromium.org

[derekjchow1, 2015-06-20 00:38:35]

The patchset sent to the CQ was uploaded after l-g-t-m from
chirantan@chromium.org
Link to the patchset: https://codereview.chromium.org/1184093007/#ps20001
(title: "Use base::Owned")

[commit-bot: I haz the power, 2015-06-20 00:39:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1184093007/20001

[commit-bot: I haz the power, 2015-06-20 00:47:17]

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

[commit-bot: I haz the power, 2015-06-20 00:47:18]

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

[Nico, 2015-06-21 05:00:19]

rs-lgtm

[Nico, 2015-06-21 05:00:27]

The CQ bit was checked by thakis@chromium.org

[commit-bot: I haz the power, 2015-06-21 05:00:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1184093007/20001

[commit-bot: I haz the power, 2015-06-21 05:14:45]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-06-21 05:15:38]

Patchset 2 (id:??) landed as
https://crrev.com/2b94f7bf55796a7112bd8dc5fafb0797e8bc9949
Cr-Commit-Position: refs/heads/master@{#335451}