media: Add ServiceFactory mojo interface and implementation.

media: Add ServiceFactory mojo interface and implementation.

ServiceFactory is an interface used to create media services. Services created
in one factory instance are isolated from services created in another factory
instance. This is used in Chromium to isolate media services for different
RenderFrames.

BUG=510653
TEST=Plays the same EME test page in 2 tabs.

Committed: https://crrev.com/c0299ae3c8a93f9d19d2367e4cc7568231f708bb
Cr-Commit-Position: refs/heads/master@{#339780}

[xhwang, 2015-07-17 23:56:06]

xhwang@chromium.org changed reviewers:
+ jam@chromium.org, rockot@chromium.org

[xhwang, 2015-07-17 23:56:07]

As discussed, adding a factory interface for media to isolate media services.
This solves the service/application identity issue. I still need to fix
HTMLViewer and tests, but I'd like to get early feedback to make sure I am on
the right track before I go too far.

PTAL

[xhwang, 2015-07-18 00:03:55]

https://codereview.chromium.org/1230313010/diff/1/media/mojo/services/service...
File media/mojo/services/service_factory_impl.cc (right):

https://codereview.chromium.org/1230313010/diff/1/media/mojo/services/service...
media/mojo/services/service_factory_impl.cc:42: new
MojoCdmService(cdm_service_context_.GetWeakPtr(), service_provider_,
Previously |cdm_service_context_| is owned by the app, which is assumed to
always outlive the services, so we pass the raw pointer to cdm_service_context_
here.

Now that |cdm_service_context_| is owned by ServiceFactoryImpl, which itself is
a service. Since service can go at any time (using StrongBind), I am using weak
pointers to prevent access violation.

Maybe I should not use StrongBind for ServiceFactoryImpl so that
ServiceFactoryImpl will always outlive the services it created. But I'd like to
defer this to the next CL...

[xhwang, 2015-07-18 00:42:27]

Now html_viewer and tests are fixed.

[jam, 2015-07-21 16:18:58]

On 2015/07/18 00:42:27, xhwang wrote:
> Now html_viewer and tests are fixed.

which parts do you want me to look at? content & components lgtm

[Ken Rockot(use gerrit already), 2015-07-21 17:29:06]

this approach lgtm

https://codereview.chromium.org/1230313010/diff/20001/content/renderer/render...
File content/renderer/render_frame_impl.cc (right):

https://codereview.chromium.org/1230313010/diff/20001/content/renderer/render...
content/renderer/render_frame_impl.cc:4985:
shell_ptr->ConnectToApplication(request.Pass(), GetProxy(&service_provider),
Hmm this is going to be a recurring pattern. I think I'll put out a follow-up CL
that gives RenderFrame a direct ConnectToApplication interface through a
persistent ShellPtr.

[xhwang, 2015-07-21 22:25:13]

rebase only

[xhwang, 2015-07-21 22:30:54]

https://codereview.chromium.org/1230313010/diff/20001/content/renderer/render...
File content/renderer/render_frame_impl.cc (right):

https://codereview.chromium.org/1230313010/diff/20001/content/renderer/render...
content/renderer/render_frame_impl.cc:4985:
shell_ptr->ConnectToApplication(request.Pass(), GetProxy(&service_provider),
On 2015/07/21 17:29:06, Ken Rockot wrote:
> Hmm this is going to be a recurring pattern. I think I'll put out a follow-up
CL
> that gives RenderFrame a direct ConnectToApplication interface through a
> persistent ShellPtr.

Acknowledged.

[xhwang, 2015-07-21 22:50:33]

The CQ bit was checked by xhwang@chromium.org

[xhwang, 2015-07-21 22:50:33]

The patchset sent to the CQ was uploaded after l-g-t-m from rockot@chromium.org,
jam@chromium.org
Link to the patchset: https://codereview.chromium.org/1230313010/#ps40001
(title: "rebase only")

[commit-bot: I haz the power, 2015-07-21 22:51:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1230313010/40001

[commit-bot: I haz the power, 2015-07-21 23:39:58]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-07-21 23:41:03]

Patchset 3 (id:??) landed as
https://crrev.com/c0299ae3c8a93f9d19d2367e4cc7568231f708bb
Cr-Commit-Position: refs/heads/master@{#339780}