Add basic browser test for EmbeddedWorker/ServiceWorker

Add basic browser test for EmbeddedWorker/ServiceWorker

This test does:
- kick registerServiceWorker (primarily for installing ServiceWorkerProviderHost)
- manually associate ServiceWorkerProviderHost to an EmbeddedWorker (eventually this should be done via ServiceWorkerVersion)
- Start and stop the EmbeddedWorker, and observe OnStarted/OnStopped msgs are sent from the child process

BUG=313530, 285976
TEST=ServiceWorkerBrowserTest.EmbeddedWorkerBasic

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=244960

[kinuko, 2014-01-08 05:46:29]

A small test that connects what we have now. PTL

[alecflett, 2014-01-08 23:02:14]

lgtm with one minor nit/question about the test itself.

https://codereview.chromium.org/127573002/diff/550001/content/browser/service...
File content/browser/service_worker/service_worker_browsertest.cc (right):

https://codereview.chromium.org/127573002/diff/550001/content/browser/service...
content/browser/service_worker/service_worker_browsertest.cc:127:
start_run_loop.Run();
is it worth putting a check here that the worker is actually running in between,
at this point in the test function?
 (not sure if you can call worker_->status() on the main thread, but you could
put a boolean that is safe-ish to access from both threads)

I only ask because from looking at this, this test could pass simply by neither
of those two functions never running in the first place.. (i.e. say for instance
the test somehow manages to complete and tear down before the thread actually
starts. That would be a failure of RunLoop, but I've seen silent failures like
that before!

[kinuko, 2014-01-09 09:08:54]

https://codereview.chromium.org/127573002/diff/550001/content/browser/service...
File content/browser/service_worker/service_worker_browsertest.cc (right):

https://codereview.chromium.org/127573002/diff/550001/content/browser/service...
content/browser/service_worker/service_worker_browsertest.cc:127:
start_run_loop.Run();
On 2014/01/08 23:02:15, alecflett wrote:
> is it worth putting a check here that the worker is actually running in
between,
> at this point in the test function?
>  (not sure if you can call worker_->status() on the main thread, but you could
> put a boolean that is safe-ish to access from both threads)
> 
> I only ask because from looking at this, this test could pass simply by
neither
> of those two functions never running in the first place.. (i.e. say for
instance
> the test somehow manages to complete and tear down before the thread actually
> starts. That would be a failure of RunLoop, but I've seen silent failures like
> that before!

Makes sense, I didn't include many cautious guards in the initial patch.

Added some more checks to reduce unexpected long-running/timeout possibility.

[kinuko, 2014-01-09 09:12:25]

Avi: mind reviewing content/{public,}/browser/storage_partition*.h changes?

It's to expose an accessor method for ServiceWorkerContext in the content/public
API (so that tests can call it).

[kinuko, 2014-01-14 08:33:19]

avi: ping. (Or I can pick another reviewer)

[kinuko, 2014-01-15 00:42:32]

jochen: could you do owner review for content/ stuff?

[jochen (gone - plz use gerrit), 2014-01-15 10:09:40]

lgtm

[commit-bot: I haz the power, 2014-01-15 14:03:33]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/kinuko@chromium.org/127573002/740001

[commit-bot: I haz the power, 2014-01-15 16:43:40]

Failed to trigger a try job on linux_rel
HTTP Error 400: Bad Request

[commit-bot: I haz the power, 2014-01-15 16:45:18]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/kinuko@chromium.org/127573002/1080001

[commit-bot: I haz the power, 2014-01-15 19:31:29]

Change committed as 244960