Revert of Ensure BrowserThread::CurrentlyOn is correct through MessageLoop teardown

Revert of Ensure BrowserThread::CurrentlyOn is correct through MessageLoop teardown (patchset #8 id:160001 of https://codereview.chromium.org/2180253003/ )

Reason for revert:
Test failures in the changes to the ComponentCloudPolicyTest. Doh.

Original issue's description:
> Ensure BrowserThread::CurrentlyOn is correct through MessageLoop teardown
>
> Changes BrowserThread::CurrentlyOn (and thus
> BrowserThreadTaskRunner::RunsTasksOnCurrentThread()) to correctly report
> a BrowserThread's association during MessageLoop destruction notification.
>
> Also adds an explicit Start() to BrowserThreadImpl as there are tests which
> call it and which incorrectly assumed base::Thread's implementation would
> call BrowserThreadImpl::StartWithOptions (which it wouldn't since the latter
> is not a virtual function.)
>
> This change provokes many tests to delete ExtensionFunction instances that
> were otherwise being leaked because UIThreadExtensionFunction::Destroy was
> incorrectly deferring destruction when run on the UI thread during shutdown.
> This in turn revealed a few small bugs which have also been fixed here.
>
> BUG=631093
> R=jam@chromium.org
>
> Committed: https://crrev.com/b02da29fb9116d1a1fb4fd0476628f333ff6bd1a
> Cr-Commit-Position: refs/heads/master@{#408295}

TBR=jam@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=631093
Committed: https://crrev.com/dd05854bb54363ce7d57ad1bedfe8df33ad76191
Cr-Commit-Position: refs/heads/master@{#408311}

[Ken Rockot(use gerrit already), 2016-07-28 02:03:30]

The CQ bit was checked by rockot@chromium.org

[Ken Rockot(use gerrit already), 2016-07-28 02:03:30]

Created Revert of Ensure BrowserThread::CurrentlyOn is correct through
MessageLoop teardown

[commit-bot: I haz the power, 2016-07-28 02:03:49]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-07-28 02:04:40]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-07-28 02:06:39]

Description was changed from

==========
Revert of Ensure BrowserThread::CurrentlyOn is correct through MessageLoop
teardown (patchset #8 id:160001 of https://codereview.chromium.org/2180253003/ )

Reason for revert:
Test failures in the changes to the ComponentCloudPolicyTest. Doh.

Original issue's description:
> Ensure BrowserThread::CurrentlyOn is correct through MessageLoop teardown
> 
> Changes BrowserThread::CurrentlyOn (and thus
> BrowserThreadTaskRunner::RunsTasksOnCurrentThread()) to correctly report
> a BrowserThread's association during MessageLoop destruction notification.
> 
> Also adds an explicit Start() to BrowserThreadImpl as there are tests which
> call it and which incorrectly assumed base::Thread's implementation would
> call BrowserThreadImpl::StartWithOptions (which it wouldn't since the latter
> is not a virtual function.)
> 
> This change provokes many tests to delete ExtensionFunction instances that
> were otherwise being leaked because UIThreadExtensionFunction::Destroy was
> incorrectly deferring destruction when run on the UI thread during shutdown.
> This in turn revealed a few small bugs which have also been fixed here.
> 
> BUG=631093
> R=jam@chromium.org
> 
> Committed: https://crrev.com/b02da29fb9116d1a1fb4fd0476628f333ff6bd1a
> Cr-Commit-Position: refs/heads/master@{#408295}

TBR=jam@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=631093
==========

to

==========
Revert of Ensure BrowserThread::CurrentlyOn is correct through MessageLoop
teardown (patchset #8 id:160001 of https://codereview.chromium.org/2180253003/ )

Reason for revert:
Test failures in the changes to the ComponentCloudPolicyTest. Doh.

Original issue's description:
> Ensure BrowserThread::CurrentlyOn is correct through MessageLoop teardown
>
> Changes BrowserThread::CurrentlyOn (and thus
> BrowserThreadTaskRunner::RunsTasksOnCurrentThread()) to correctly report
> a BrowserThread's association during MessageLoop destruction notification.
>
> Also adds an explicit Start() to BrowserThreadImpl as there are tests which
> call it and which incorrectly assumed base::Thread's implementation would
> call BrowserThreadImpl::StartWithOptions (which it wouldn't since the latter
> is not a virtual function.)
>
> This change provokes many tests to delete ExtensionFunction instances that
> were otherwise being leaked because UIThreadExtensionFunction::Destroy was
> incorrectly deferring destruction when run on the UI thread during shutdown.
> This in turn revealed a few small bugs which have also been fixed here.
>
> BUG=631093
> R=jam@chromium.org
>
> Committed: https://crrev.com/b02da29fb9116d1a1fb4fd0476628f333ff6bd1a
> Cr-Commit-Position: refs/heads/master@{#408295}

TBR=jam@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=631093

Committed: https://crrev.com/dd05854bb54363ce7d57ad1bedfe8df33ad76191
Cr-Commit-Position: refs/heads/master@{#408311}
==========

[commit-bot: I haz the power, 2016-07-28 02:06:39]

Patchset 1 (id:??) landed as
https://crrev.com/dd05854bb54363ce7d57ad1bedfe8df33ad76191
Cr-Commit-Position: refs/heads/master@{#408311}