Fix LifeCycleWatcher in ScopedVectorTest to not be referenced after being destructed.

Fix LifeCycleWatcher in ScopedVectorTest to not be referenced after being destructed.

If the LifeCycleObject owned by LifeCycleWatcher still exists at the
time LifeCycleWatcher is destroyed, the LifeCycleObject's destructor
will call LifeCycleWatcher::OnLifeCycleDestroy() after
LifeCycleWatcher's destructor has finished running.

BUG=535321
Committed: https://crrev.com/8723a2c2a90e060006d1b45a6ce58bacf17bb9d8
Cr-Commit-Position: refs/heads/master@{#350683}

[Anand Mistry (off Chromium), 2015-09-24 05:35:23]

amistry@chromium.org changed reviewers:
+ danakj@chromium.org

[danakj, 2015-09-24 17:42:19]

LGTM

[danakj, 2015-09-24 17:43:52]

https://codereview.chromium.org/1365893002/diff/1/base/memory/scoped_vector_u...
File base/memory/scoped_vector_unittest.cc (right):

https://codereview.chromium.org/1365893002/diff/1/base/memory/scoped_vector_u...
base/memory/scoped_vector_unittest.cc:71: // Stop watching the watched object.
Without this, the object's destructor
Alternatively, you could just keep the watcher alive in the test this is causing
a problem for? Which test(s) are wrong here?

[Anand Mistry (off Chromium), 2015-09-24 21:56:27]

https://codereview.chromium.org/1365893002/diff/1/base/memory/scoped_vector_u...
File base/memory/scoped_vector_unittest.cc (right):

https://codereview.chromium.org/1365893002/diff/1/base/memory/scoped_vector_u...
base/memory/scoped_vector_unittest.cc:71: // Stop watching the watched object.
Without this, the object's destructor
On 2015/09/24 17:43:51, danakj wrote:
> Alternatively, you could just keep the watcher alive in the test this is
causing
> a problem for? Which test(s) are wrong here?

The two tests that fail are WeakClear and InsertRange. Both create watched
object thats aren't deleted by ScopedVector, so it's up to LifeCycleWatcher to
clean up at the end of the test.

[danakj, 2015-09-24 22:00:37]

I see. We could clear the vector, but this seems fine and the branch doesn't
look particularly dangerous to breaking (or hiding failures), so ya LGTM still.

[Anand Mistry (off Chromium), 2015-09-24 22:12:28]

The CQ bit was checked by amistry@chromium.org

[commit-bot: I haz the power, 2015-09-24 22:12:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1365893002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1365893002/1

[commit-bot: I haz the power, 2015-09-24 22:28:39]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-09-24 22:29:50]

Patchset 1 (id:??) landed as
https://crrev.com/8723a2c2a90e060006d1b45a6ce58bacf17bb9d8
Cr-Commit-Position: refs/heads/master@{#350683}