media: Make media mojo services a component

media: Make media mojo services a component

Previously media mojo services are in multiple small source_sets. This
was okay when they are only depended by one target (i.e. content). But
now we have multiple targets depending on them. For example, content and
media/gpu both depend on cdm_services. This causes duplicate definition
of some global veriables, e.g. g_cdm_manager in mojo_cdm_service.cc.

TBR=jam@chromium.org
BUG=610904
TEST=EME playback works.

Committed: https://crrev.com/d3b5d3f46840aba12d0788a14f14be9739712284
Cr-Commit-Position: refs/heads/master@{#394323}

[xhwang, 2016-05-11 00:53:16]

Patchset #1 (id:1) has been deleted

[xhwang, 2016-05-11 07:11:37]

The CQ bit was checked by xhwang@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-05-11 07:11:50]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1968833002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1968833002/60001

[commit-bot: I haz the power, 2016-05-11 08:18:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-05-11 08:18:08]

Dry run: This issue passed the CQ dry run.

[xhwang, 2016-05-11 16:20:43]

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

[xhwang, 2016-05-11 16:20:44]

ddorwin/jam: PTAL

[watk, 2016-05-11 17:28:17]

watk@chromium.org changed reviewers:
+ watk@chromium.org

[watk, 2016-05-11 17:28:18]

https://chromiumcodereview.appspot.com/1968833002/diff/60001/media/mojo/servi...
File media/mojo/services/BUILD.gn (right):

https://chromiumcodereview.appspot.com/1968833002/diff/60001/media/mojo/servi...
media/mojo/services/BUILD.gn:51: deps = [
Drive by :) gn style guide prefers to put this after sources
https://chromium.googlesource.com/chromium/src/+/master/tools/gn/docs/style_g...

[jam, 2016-05-11 20:36:26]

which files do you want me to review?

[xhwang, 2016-05-11 20:46:01]

On 2016/05/11 20:36:26, jam wrote:
> which files do you want me to review?

I should've made it more clear.

Please OWNERS review everything in content/ and chrome/. The main change is in
media/mojo/services/BUILD.gn in case you wonder what's changed.

[ddorwin, 2016-05-11 23:02:28]

I can't speak to whether this is a good idea, but I reviewed the changes.

https://codereview.chromium.org/1968833002/diff/80001/chrome/browser/chrome_c...
File chrome/browser/chrome_content_browser_client.cc (right):

https://codereview.chromium.org/1968833002/diff/80001/chrome/browser/chrome_c...
chrome/browser/chrome_content_browser_client.cc:313: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
Why is this now required? Is this a bug in DEPS handling?

https://codereview.chromium.org/1968833002/diff/80001/content/gpu/BUILD.gn
File content/gpu/BUILD.gn (right):

https://codereview.chromium.org/1968833002/diff/80001/content/gpu/BUILD.gn#ne...
content/gpu/BUILD.gn:68: deps += [ "//media/mojo/services" ]
This was previously always included. Was that incorrect?

https://codereview.chromium.org/1968833002/diff/80001/content/gpu/gpu_process...
File content/gpu/gpu_process_control_impl.cc (right):

https://codereview.chromium.org/1968833002/diff/80001/content/gpu/gpu_process...
content/gpu/gpu_process_control_impl.cc:11: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
ditto

https://codereview.chromium.org/1968833002/diff/80001/content/shell/BUILD.gn
File content/shell/BUILD.gn (right):

https://codereview.chromium.org/1968833002/diff/80001/content/shell/BUILD.gn#...
content/shell/BUILD.gn:268: deps += [ "//media/mojo/services" ]
ditto on default

https://codereview.chromium.org/1968833002/diff/80001/content/shell/browser/s...
File content/shell/browser/shell_content_browser_client.cc (right):

https://codereview.chromium.org/1968833002/diff/80001/content/shell/browser/s...
content/shell/browser/shell_content_browser_client.cc:63: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
ditto

https://codereview.chromium.org/1968833002/diff/80001/content/utility/BUILD.gn
File content/utility/BUILD.gn (right):

https://codereview.chromium.org/1968833002/diff/80001/content/utility/BUILD.g...
content/utility/BUILD.gn:39: deps += [ "//media/mojo/services" ]
ditto on default

https://codereview.chromium.org/1968833002/diff/80001/content/utility/utility...
File content/utility/utility_process_control_impl.cc (right):

https://codereview.chromium.org/1968833002/diff/80001/content/utility/utility...
content/utility/utility_process_control_impl.cc:14: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
ditto

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/BUI...
File media/mojo/services/BUILD.gn (right):

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/BUI...
media/mojo/services/BUILD.gn:91: "demuxer_stream_provider_shim.cc",
We now build everything regardless of whether a given platform/build
configuration needs/uses it?

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/BUI...
media/mojo/services/BUILD.gn:120: defines = [ "MEDIA_MOJO_IMPLEMENTATION" ]
This is new?

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/BUI...
media/mojo/services/BUILD.gn:121: configs += [
"//build/config/compiler:no_size_t_to_int_warning" ]
Should we comment the reason for this?

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/BUI...
media/mojo/services/BUILD.gn:122: public_configs = [ ":mojo_media_config" ]
This was associated with a TODO about cdm_service. Is this no longer
incorrect/needing correction?

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/BUI...
media/mojo/services/BUILD.gn:145: if (enable_test_mojo_media_client) {
Does this mean we build all the files above multiple times - once for the
product and once for the test?

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/med...
File media/mojo/services/media_mojo_export.h (right):

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/med...
media/mojo/services/media_mojo_export.h:1: // Copyright 2015 The Chromium
Authors. All rights reserved.
Was this copied or should this be updated?

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/med...
media/mojo/services/media_mojo_export.h:15: #define MEDIA_MOJO_EXPORT
__declspec(dllexport)
Does the discussion, including naming, in
https://bugs.chromium.org/p/chromium/issues/detail?id=609564 apply here?

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/moj...
File media/mojo/services/mojo_cdm_allocator.h (right):

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/moj...
media/mojo/services/mojo_cdm_allocator.h:25: class MEDIA_MOJO_EXPORT
MojoCdmAllocator : public CdmAllocator {
Why wasn't this necessary before? Do we really want to export it if it wasn't
necessary before?

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/moj...
File media/mojo/services/mojo_cdm_service.h (right):

https://codereview.chromium.org/1968833002/diff/80001/media/mojo/services/moj...
media/mojo/services/mojo_cdm_service.h:40: static scoped_refptr<MediaKeys>
MEDIA_MOJO_EXPORT LegacyGetCdm(int cdm_id);
Why does only this static method need to be exposed?

https://codereview.chromium.org/1968833002/diff/80001/media/test/BUILD.gn
File media/test/BUILD.gn (right):

https://codereview.chromium.org/1968833002/diff/80001/media/test/BUILD.gn#new...
media/test/BUILD.gn:104: "//media/mojo/services:proxy",
It seems strange that we have a subitem even though we generally use the main
name. I guess that already exists, such as lines 98-99.

[xhwang, 2016-05-12 06:32:23]

xhwang@chromium.org changed reviewers:
+ brettw@chromium.org
- jam@chromium.org

[xhwang, 2016-05-12 06:32:25]

s/jam/brettw since this change is mainly about gn refactoring.

brettw: Could you review this CL from GN's perspective. I am merging smaller
source_sets into a "component". Also, I have two questions about nogncheck.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/chrome/browser/c...
File chrome/browser/chrome_content_browser_client.cc (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/chrome/browser/c...
chrome/browser/chrome_content_browser_client.cc:313: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
On 2016/05/11 23:02:27, ddorwin wrote:
> Why is this now required? Is this a bug in DEPS handling?

// nogncheck is recommended for this kind of conditional includes:

https://chromium.googlesource.com/chromium/src/+/master/tools/gn/docs/referen...

But I actually don't see "nogncheck" used anywhere in chrome/.

+brettw: Do you know what's the status and recommendation of "gn check" in
chrome/?

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/gpu/BUIL...
File content/gpu/BUILD.gn (left):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/gpu/BUIL...
content/gpu/BUILD.gn:60: "//media/mojo/services:application_factory",
+brettw: This was added in https://codereview.chromium.org/1833193002, even
though we have the conditional deps below at line 69. And in the source code we
have a conditional include.

Are we preferring unconditional deps to nogncheck now?

[xhwang, 2016-05-13 18:02:21]

brettw: kindly ping :)

[brettw, 2016-05-13 21:12:42]

"gn check" isn't yet enabled in chrome/browser which is why it was working
before. From looking at this, I'm pretty sure these annotations will be
required, so feel free to add them.

You'll have to decide for the design of your component whether unconditional
deps are required or better. It doesn't make sense to have it appear twice as in
the example you pointed out.

[xhwang, 2016-05-16 05:39:52]

comments addressed

[xhwang, 2016-05-16 05:41:21]

Addressed two trivial comments. Mostly replies for discussion. PTAL again.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/chrome/browser/c...
File chrome/browser/chrome_content_browser_client.cc (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/chrome/browser/c...
chrome/browser/chrome_content_browser_client.cc:313: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
On 2016/05/12 06:32:24, xhwang wrote:
> On 2016/05/11 23:02:27, ddorwin wrote:
> > Why is this now required? Is this a bug in DEPS handling?
> 
> // nogncheck is recommended for this kind of conditional includes:
> 
>
https://chromium.googlesource.com/chromium/src/+/master/tools/gn/docs/referen...
> 
> But I actually don't see "nogncheck" used anywhere in chrome/.
> 
> +brettw: Do you know what's the status and recommendation of "gn check" in
> chrome/?

Per brettw's comment, I think this is the right thing to do.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/gpu/BUIL...
File content/gpu/BUILD.gn (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/gpu/BUIL...
content/gpu/BUILD.gn:68: deps += [ "//media/mojo/services" ]
On 2016/05/11 23:02:27, ddorwin wrote:
> This was previously always included. Was that incorrect?

We should either only always include, or only conditional include. I think
conditional include makes more sense.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/gpu/gpu_...
File content/gpu/gpu_process_control_impl.cc (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/gpu/gpu_...
content/gpu/gpu_process_control_impl.cc:11: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
On 2016/05/11 23:02:27, ddorwin wrote:
> ditto

ditto

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/shell/BU...
File content/shell/BUILD.gn (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/shell/BU...
content/shell/BUILD.gn:268: deps += [ "//media/mojo/services" ]
On 2016/05/11 23:02:27, ddorwin wrote:
> ditto on default

ditto

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/shell/br...
File content/shell/browser/shell_content_browser_client.cc (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/shell/br...
content/shell/browser/shell_content_browser_client.cc:63: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
On 2016/05/11 23:02:27, ddorwin wrote:
> ditto

ditto

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/utility/...
File content/utility/BUILD.gn (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/utility/...
content/utility/BUILD.gn:39: deps += [ "//media/mojo/services" ]
On 2016/05/11 23:02:27, ddorwin wrote:
> ditto on default

ditto

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/utility/...
File content/utility/utility_process_control_impl.cc (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/content/utility/...
content/utility/utility_process_control_impl.cc:14: #include
"media/mojo/services/mojo_media_application_factory.h"  // nogncheck
On 2016/05/11 23:02:27, ddorwin wrote:
> ditto

ditto

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
File media/mojo/services/BUILD.gn (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/BUILD.gn:91: "demuxer_stream_provider_shim.cc",
On 2016/05/11 23:02:27, ddorwin wrote:
> We now build everything regardless of whether a given platform/build
> configuration needs/uses it?

Yes. Previously I was trying to keep things separate so that we only build
things that we need. But it becomes a pain to maintain. Now I need to use
component (shared/static library). Having multiple components would be even
harder: we need to have different MEDIA_MOJO_XXX_EXPORT for each of them.

By building everything, we rely on the linker to remove symbols that are not
reachable. So it won't affect production build. It does increase build time a
little bit when mojo media is enabled, but makes the build file much easier to
manage.

Another option is to include things based on the config above, e.g. only include
audio decoder service when ENABLE_MOJO_AUDIO_DECODER is true. That's doable. But
again I am not sure whether it's worth it.

WDYT?

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/BUILD.gn:120: defines = [ "MEDIA_MOJO_IMPLEMENTATION" ]
On 2016/05/11 23:02:27, ddorwin wrote:
> This is new?

Yes, this is needed since we are "component" now.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/BUILD.gn:121: configs += [
"//build/config/compiler:no_size_t_to_int_warning" ]
On 2016/05/11 23:02:28, ddorwin wrote:
> Should we comment the reason for this?

Done

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/BUILD.gn:122: public_configs = [ ":mojo_media_config" ]
On 2016/05/11 23:02:27, ddorwin wrote:
> This was associated with a TODO about cdm_service. Is this no longer
> incorrect/needing correction?

cdm_service doesn't need this but has to have it for the reason documented in
the TODO. Now we have the app in this target and we should always have this
public_configs.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/BUILD.gn:145: if (enable_test_mojo_media_client) {
On 2016/05/11 23:02:28, ddorwin wrote:
> Does this mean we build all the files above multiple times - once for the
> product and once for the test?

This is used for test in chromium by developers, not in unittests. The files
above will only be built once.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
File media/mojo/services/media_mojo_export.h (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/media_mojo_export.h:1: // Copyright 2015 The Chromium
Authors. All rights reserved.
On 2016/05/11 23:02:28, ddorwin wrote:
> Was this copied or should this be updated?

Done.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/media_mojo_export.h:15: #define MEDIA_MOJO_EXPORT
__declspec(dllexport)
On 2016/05/11 23:02:28, ddorwin wrote:
> Does the discussion, including naming, in
> https://bugs.chromium.org/p/chromium/issues/detail?id=609564 apply here?

This is equivalent to MEDIA_EXPORT. I'd like to follow the current convention
for now. If we decide to change, then we should change all XXX_EXPORT in
chromium.

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
File media/mojo/services/mojo_cdm_allocator.h (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/mojo_cdm_allocator.h:25: class MEDIA_MOJO_EXPORT
MojoCdmAllocator : public CdmAllocator {
On 2016/05/11 23:02:28, ddorwin wrote:
> Why wasn't this necessary before? Do we really want to export it if it wasn't
> necessary before?

Previously we use a source_set, where this file will always be statically linked
into the depending target, so there's no need to export anything. Now we are
using a component, which will be a shared library in component builds. We need
XXX_EXPORT for unittest (we have a test for this class).

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
File media/mojo/services/mojo_cdm_service.h (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/mojo/servi...
media/mojo/services/mojo_cdm_service.h:40: static scoped_refptr<MediaKeys>
MEDIA_MOJO_EXPORT LegacyGetCdm(int cdm_id);
On 2016/05/11 23:02:28, ddorwin wrote:
> Why does only this static method need to be exposed?

See the comment above for some context. This is called by AVDA, which lives in a
different target (media_gpu).

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/test/BUILD.gn
File media/test/BUILD.gn (right):

https://chromiumcodereview.appspot.com/1968833002/diff/80001/media/test/BUILD...
media/test/BUILD.gn:104: "//media/mojo/services:proxy",
On 2016/05/11 23:02:28, ddorwin wrote:
> It seems strange that we have a subitem even though we generally use the main
> name. I guess that already exists, such as lines 98-99.

Yeah. We do have a plan to move proxies to another folder. Then this will become
something like //media/mojo/proxies. But I am sure the name convention "proxy"
has been deprecated. I need to find out what the new name should be.

[ddorwin, 2016-05-16 21:23:38]

lgtm

Thanks.

[xhwang, 2016-05-17 04:58:38]

rebase only

[xhwang, 2016-05-17 05:01:18]

The CQ bit was checked by xhwang@chromium.org

[xhwang, 2016-05-17 05:01:18]

The patchset sent to the CQ was uploaded after l-g-t-m from ddorwin@chromium.org
Link to the patchset: https://codereview.chromium.org/1968833002/#ps120001
(title: "rebase only")

[commit-bot: I haz the power, 2016-05-17 05:02:11]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1968833002/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1968833002/120001

[commit-bot: I haz the power, 2016-05-17 05:08:39]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-05-17 05:08:40]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[xhwang, 2016-05-17 05:29:42]

Description was changed from

==========
media: Make media mojo services a component

Previously media mojo services are in multiple small source_sets. This
was okay when they are only depended by one target (i.e. content). But
now we have multiple targets depending on them. For example, content and
media/gpu both depend on cdm_services. This causes duplicate definition
of some global veriables, e.g. g_cdm_manager in mojo_cdm_service.cc.

BUG=610904
TEST=EME playback works.
==========

to

==========
media: Make media mojo services a component

Previously media mojo services are in multiple small source_sets. This
was okay when they are only depended by one target (i.e. content). But
now we have multiple targets depending on them. For example, content and
media/gpu both depend on cdm_services. This causes duplicate definition
of some global veriables, e.g. g_cdm_manager in mojo_cdm_service.cc.

TBR=jam@chromium.org
BUG=610904
TEST=EME playback works.
==========

[xhwang, 2016-05-17 05:29:43]

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

[xhwang, 2016-05-17 16:09:12]

TBRing jam@ for trivial changes in content/.

jam@: Feel free to review the changes. Otherwise I'll CQ this CL later today.

[xhwang, 2016-05-17 18:40:31]

The CQ bit was checked by xhwang@chromium.org

[xhwang, 2016-05-17 18:40:31]

The patchset sent to the CQ was uploaded after l-g-t-m from ddorwin@chromium.org
Link to the patchset:
https://chromiumcodereview.appspot.com/1968833002/#ps140001 (title: "fix windows
error")

[commit-bot: I haz the power, 2016-05-17 18:41:01]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1968833002/140001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1968833002/140001

[commit-bot: I haz the power, 2016-05-17 18:50:00]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-05-17 18:50:02]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[xhwang, 2016-05-17 19:15:53]

rebase only

[xhwang, 2016-05-18 00:31:46]

The CQ bit was checked by xhwang@chromium.org

[xhwang, 2016-05-18 00:31:46]

The patchset sent to the CQ was uploaded after l-g-t-m from ddorwin@chromium.org
Link to the patchset: https://codereview.chromium.org/1968833002/#ps160001
(title: "rebase only")

[commit-bot: I haz the power, 2016-05-18 00:32:54]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1968833002/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1968833002/160001

[commit-bot: I haz the power, 2016-05-18 03:38:10]

Description was changed from

==========
media: Make media mojo services a component

Previously media mojo services are in multiple small source_sets. This
was okay when they are only depended by one target (i.e. content). But
now we have multiple targets depending on them. For example, content and
media/gpu both depend on cdm_services. This causes duplicate definition
of some global veriables, e.g. g_cdm_manager in mojo_cdm_service.cc.

TBR=jam@chromium.org
BUG=610904
TEST=EME playback works.
==========

to

==========
media: Make media mojo services a component

Previously media mojo services are in multiple small source_sets. This
was okay when they are only depended by one target (i.e. content). But
now we have multiple targets depending on them. For example, content and
media/gpu both depend on cdm_services. This causes duplicate definition
of some global veriables, e.g. g_cdm_manager in mojo_cdm_service.cc.

TBR=jam@chromium.org
BUG=610904
TEST=EME playback works.
==========

[commit-bot: I haz the power, 2016-05-18 03:38:13]

Committed patchset #8 (id:160001)

[commit-bot: I haz the power, 2016-05-18 03:39:19]

Description was changed from

==========
media: Make media mojo services a component

Previously media mojo services are in multiple small source_sets. This
was okay when they are only depended by one target (i.e. content). But
now we have multiple targets depending on them. For example, content and
media/gpu both depend on cdm_services. This causes duplicate definition
of some global veriables, e.g. g_cdm_manager in mojo_cdm_service.cc.

TBR=jam@chromium.org
BUG=610904
TEST=EME playback works.
==========

to

==========
media: Make media mojo services a component

Previously media mojo services are in multiple small source_sets. This
was okay when they are only depended by one target (i.e. content). But
now we have multiple targets depending on them. For example, content and
media/gpu both depend on cdm_services. This causes duplicate definition
of some global veriables, e.g. g_cdm_manager in mojo_cdm_service.cc.

TBR=jam@chromium.org
BUG=610904
TEST=EME playback works.

Committed: https://crrev.com/d3b5d3f46840aba12d0788a14f14be9739712284
Cr-Commit-Position: refs/heads/master@{#394323}
==========

[commit-bot: I haz the power, 2016-05-18 03:39:21]

Patchset 8 (id:??) landed as
https://crrev.com/d3b5d3f46840aba12d0788a14f14be9739712284
Cr-Commit-Position: refs/heads/master@{#394323}