Implement the ability to layout test Web Worker-based Notifications.

Implement the ability to layout test Web Worker-based Notifications.

Layout tests currently utilize a mocked NotificationPresenter implemented
in the test runner, but this won't be sufficient for Web Worker-based
notifications given that these won't be tied to a WebFrame.

Instead, pull up the layout test functionality to the browser process in
content_shell (only available when running layout tests!), similar to the
ability to request Notification permissions, so that we can allow displaying
and interacting with "displayed" notifications in layout tests as well.

The existing code path for Document-based Notifications is kept intact, but
will be removed once both Blink and Chromium switch over.

BUG=392187
Committed: https://crrev.com/a18fd1f1c85223d078d1b52456898da58708b2ed
Cr-Commit-Position: refs/heads/master@{#301084}

[Peter Beverloo, 2014-10-23 15:21:28]

PTAL. This won't do anything until we switch over the Blink implementation,
since we won't reach LayoutTestNotificationManager::Show(). I'll send out an
Intent to Implement momentarily.

Full patches which I'm splitting up can be seen here:
  Blink: https://codereview.chromium.org/624033003/
  Chromium: https://codereview.chromium.org/626973002/

[Peter Beverloo, 2014-10-23 16:35:54]

peter@chromium.org changed reviewers:
+ mkwst@chromium.org

[Peter Beverloo, 2014-10-23 16:35:54]

(Let's actually add Mike as a reviewer :-).)

[Mike West, 2014-10-23 19:05:35]

LGTM % comments.

https://codereview.chromium.org/659043003/diff/1/content/shell/browser/layout...
File content/shell/browser/layout_test/layout_test_content_browser_client.cc
(right):

https://codereview.chromium.org/659043003/diff/1/content/shell/browser/layout...
content/shell/browser/layout_test/layout_test_content_browser_client.cc:39:
layout_test_notification_manager_.reset(
This drops the thread assertion you used to have. Is that intentional?

https://codereview.chromium.org/659043003/diff/1/content/shell/browser/layout...
content/shell/browser/layout_test/layout_test_content_browser_client.cc:103:
LayoutTestNotificationManager* manager = GetLayoutTestNotificationManager();
Nit: This probably fits on one line as `if (LayoutTestNotificationManager ...)`

https://codereview.chromium.org/659043003/diff/1/content/shell/common/shell_m...
File content/shell/common/shell_messages.h (right):

https://codereview.chromium.org/659043003/diff/1/content/shell/common/shell_m...
content/shell/common/shell_messages.h:109: std::string /* title */)
At some point, it would be good to move our layout-test-only messages out to a
different message file. Not in this CL.

[Peter Beverloo, 2014-10-24 11:32:06]

Thanks Mike!

https://codereview.chromium.org/659043003/diff/1/content/shell/browser/layout...
File content/shell/browser/layout_test/layout_test_content_browser_client.cc
(right):

https://codereview.chromium.org/659043003/diff/1/content/shell/browser/layout...
content/shell/browser/layout_test/layout_test_content_browser_client.cc:39:
layout_test_notification_manager_.reset(
On 2014/10/23 19:05:35, Mike West wrote:
> This drops the thread assertion you used to have. Is that intentional?

Yes. The LayoutTestNotificationManager can now be accessed on the IO thread and
the UI thread depending on the interaction (to match the notification system
living in //chrome). Per-method asserts have been added there.

https://codereview.chromium.org/659043003/diff/1/content/shell/browser/layout...
content/shell/browser/layout_test/layout_test_content_browser_client.cc:103:
LayoutTestNotificationManager* manager = GetLayoutTestNotificationManager();
On 2014/10/23 19:05:35, Mike West wrote:
> Nit: This probably fits on one line as `if (LayoutTestNotificationManager
...)`

No, it doesn't, but it does when using auto \o/.

https://codereview.chromium.org/659043003/diff/1/content/shell/common/shell_m...
File content/shell/common/shell_messages.h (right):

https://codereview.chromium.org/659043003/diff/1/content/shell/common/shell_m...
content/shell/common/shell_messages.h:109: std::string /* title */)
On 2014/10/23 19:05:35, Mike West wrote:
> At some point, it would be good to move our layout-test-only messages out to a
> different message file. Not in this CL.

Agreed. Let me follow up with that :-).

[Peter Beverloo, 2014-10-24 11:32:24]

The CQ bit was checked by peter@chromium.org

[commit-bot: I haz the power, 2014-10-24 11:33:35]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/659043003/20001

[commit-bot: I haz the power, 2014-10-24 12:16:03]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2014-10-24 12:17:11]

Patchset 2 (id:??) landed as
https://crrev.com/a18fd1f1c85223d078d1b52456898da58708b2ed
Cr-Commit-Position: refs/heads/master@{#301084}