Remove Pipeline::IsInitialized() and replace stop_pending_ with checks for stop_cb_.

Remove Pipeline::IsInitialized() and replace stop_pending_ with checks for stop_cb_.

Unfortunately until bug 110228 is fixed I have to keep IsInitializedForTesting() around but atleast it enforces that state_ is accessed on the same thread instead of locking.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=150899

[scherkus (not reviewing), 2012-08-09 19:27:10]

http://codereview.chromium.org/10823261/diff/6/media/base/pipeline.h
File media/base/pipeline.h (right):

http://codereview.chromium.org/10823261/diff/6/media/base/pipeline.h#newcode206
media/base/pipeline.h:206: bool IsInitializedForTesting();
FYI this gets removed in follow-up CL http://codereview.chromium.org/10823261/

[Ami GONE FROM CHROMIUM, 2012-08-09 20:46:23]

lgtm

http://codereview.chromium.org/10823261/diff/6/media/base/pipeline.h
File media/base/pipeline.h (right):

http://codereview.chromium.org/10823261/diff/6/media/base/pipeline.h#newcode108
media/base/pipeline.h:108: // Stop() has completed:
I thought error during Stop() doesn't fire error_cb.  IOW s/completed/has been
called/ above?

[scherkus (not reviewing), 2012-08-09 21:02:59]

http://codereview.chromium.org/10823261/diff/6/media/base/pipeline.h
File media/base/pipeline.h (right):

http://codereview.chromium.org/10823261/diff/6/media/base/pipeline.h#newcode108
media/base/pipeline.h:108: // Stop() has completed:
On 2012/08/09 20:46:23, Ami Fischman wrote:
> I thought error during Stop() doesn't fire error_cb.  IOW s/completed/has been
> called/ above?

You are correct but as discussed offline this is as accurate of a guarantee
we're willing to make (i.e., error/ended _could_ fire after Stop() is called
since Stop() has to queue a task on the pipeline msg loop)