Split WarningService from ExtensionSystem.

Split WarningService from ExtensionSystem.

This patch removes both the WarningService accessors from
ExtensionSystem and also the ExtensionWarningBadgeService instance
that it created as this can be built automatically for every
BrowserContext simply by declaring so in its new factory.

BUG=

Committed: https://crrev.com/22114a38f8e580253332a3fbef0a1892aee1e982
Cr-Commit-Position: refs/heads/master@{#302438}

[Reilly Grant (use Gerrit), 2014-10-31 18:09:41]

reillyg@chromium.org changed reviewers:
+ mek@chromium.org

[Reilly Grant (use Gerrit), 2014-10-31 18:10:50]

reillyg@chromium.org changed reviewers:
+ xiyuan@chromium.org

[Reilly Grant (use Gerrit), 2014-10-31 18:28:47]

xiyuan@, please take a look at
chrome/browser/ui/webui/extensions/extension_settings_handler.cc.

mek@, please take a look at everything else.

[xiyuan, 2014-10-31 18:39:54]

extension_settings_handler.cc LGTM

[Marijn Kruisselbrink, 2014-10-31 20:38:10]

lgtm if you fix the virtual/override issues

https://codereview.chromium.org/698623002/diff/1/chrome/browser/extensions/ex...
File chrome/browser/extensions/extension_warning_badge_service.cc (right):

https://codereview.chromium.org/698623002/diff/1/chrome/browser/extensions/ex...
chrome/browser/extensions/extension_warning_badge_service.cc:93:
warning_service_observer_.Add(WarningService::Get(profile_));
The old code removed this observer in the Shutdown stage of KeyedService, while
your new code does this in the destructor stage. I'm not sure why the old code
did that though, as I can't think of any case where the two stage shutdown would
be necessary for this particular dependency, so it's probably okay to keep it
like this.
(actually, I can see why the old code needed it; the order of those data members
in ExtensionSystemImpl::Shared would have caused them to be destroyed in the
wrong order)

https://codereview.chromium.org/698623002/diff/1/chrome/browser/extensions/ex...
File chrome/browser/extensions/extension_warning_badge_service_factory.h
(right):

https://codereview.chromium.org/698623002/diff/1/chrome/browser/extensions/ex...
chrome/browser/extensions/extension_warning_badge_service_factory.h:26: virtual
~ExtensionWarningBadgeServiceFactory();
override instead of virtual

https://codereview.chromium.org/698623002/diff/1/chrome/browser/extensions/ex...
chrome/browser/extensions/extension_warning_badge_service_factory.h:29: virtual
KeyedService* BuildServiceInstanceFor(
The current styleguide says you should only use one of [virtual, final,
override]. So get rid of the virtual.

https://codereview.chromium.org/698623002/diff/1/extensions/browser/warning_s...
File extensions/browser/warning_service_factory.h (right):

https://codereview.chromium.org/698623002/diff/1/extensions/browser/warning_s...
extensions/browser/warning_service_factory.h:24: virtual
~WarningServiceFactory();
override instead of virtual

https://codereview.chromium.org/698623002/diff/1/extensions/browser/warning_s...
extensions/browser/warning_service_factory.h:27: virtual KeyedService*
BuildServiceInstanceFor(
no virtual when you already have override

[Reilly Grant (use Gerrit), 2014-11-03 15:56:47]

The CQ bit was checked by reillyg@chromium.org

[commit-bot: I haz the power, 2014-11-03 15:57:55]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/698623002/20001

[commit-bot: I haz the power, 2014-11-03 16:48:06]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2014-11-03 16:48:55]

Patchset 2 (id:??) landed as
https://crrev.com/22114a38f8e580253332a3fbef0a1892aee1e982
Cr-Commit-Position: refs/heads/master@{#302438}