Removing AudioSystem::GetAudioManager() usage from speech recognition

Removing AudioSystem::GetAudioManager() usage from speech recognition.

After https://codereview.chromium.org/2763383002/ lands, AudioSystem::GetAudioManager() will be removed: it's only purpose was to assist in gradually switching clients from AudioManager to AudioSystem interface.
However, it leaked a bit in speech recognition, so - fixing it.

The plan is:

1) After the last client (WebRTC audio private API) is switched to AudioSystem interface, all the related AudioManager methods will be made protected to be used by AudioSystem only; so essentially AudioManager interface will become a factory for audio streams.

2) Further, all audio stream clients will be switched to asynchronous stream creation interface, and then to Mojo.

BUG=672468
Review-Url: https://codereview.chromium.org/2785463005
Cr-Commit-Position: refs/heads/master@{#460763}
Committed: https://chromium.googlesource.com/chromium/src/+/b6082fccb8e9fc2f21871508bdcf79ca567326ad

[o1ka, 2017-03-29 17:22:44]

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

[commit-bot: I haz the power, 2017-03-29 17:23:10]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[o1ka, 2017-03-29 17:33:58]

Description was changed from

==========
Removing AudioSystem::GetAudioManager() usage from speech recognition

BUG=672468
==========

to

==========
Removing AudioSystem::GetAudioManager() usage from speech recognition.

After https://codereview.chromium.org/2763383002/ lands,
AudioSystem::GetAudioManager() will be removed: it's only purpose was to assist
in gradually switching clients from AudioManager to AudioSystem interface.
However, it leaked a bit in speech recognition, so - fixing it.

The plan is:

1) After the last client (WebRTC audio private API) is switched to AudioSystem
interface, all the related AudioManager methods will be made protected to be
used by AudioSystem only; so essentially AudioManager interface will become a
factory for audio streams.

2) Further, all audio stream clients will be switched to asynchronous stream
creation interface, and then to Mojo.

BUG=672468
==========

[o1ka, 2017-03-29 17:33:58]

olka@chromium.org changed reviewers:
+ dalecurtis@chromium.org

[o1ka, 2017-03-29 17:42:28]

Dale - PTAL.
This is re: your comment on AudioSystem::GetAudioManager() in
https://codereview.chromium.org/2763383002/, which I agree with. I forgot to
remove the method there, and now I've realized the leak of GetAudioManager()
usage in Speech.

WDYT about the plan? Who may be unhappy if we close direct access to
AudioManager device info interfaces (served by AudioSystem now)? Content
embedders?

Thanks!

[DaleCurtis, 2017-03-29 17:52:10]

I don't think anyone will be. They all have access via AudioManager::Get(); so
lgtm

[o1ka, 2017-03-29 18:01:59]

On 2017/03/29 17:52:10, DaleCurtis wrote:
> I don't think anyone will be. They all have access via AudioManager::Get(); so
> lgtm

My question is rather about unpublising AM methods: the only access to them will
be via AS. (And we have AS::Get() as well)

[commit-bot: I haz the power, 2017-03-29 18:13:54]

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

[commit-bot: I haz the power, 2017-03-29 18:13:55]

Dry run: This issue passed the CQ dry run.

[o1ka, 2017-03-29 19:49:12]

olka@chromium.org changed reviewers:
+ avi@chromium.org

[o1ka, 2017-03-29 19:49:12]

avi@chromium.org: could you RS as an owner?

[DaleCurtis, 2017-03-29 19:50:45]

Hmm, I dunno about that plan. It's too soon to tell if that will work out. AM
does a lot of things. I'm definitely supportive of paring it down to AS only
being visible, but nominally that's what the AM interface is already supposed to
be.

[o1ka, 2017-03-29 19:55:59]

On 2017/03/29 19:50:45, DaleCurtis wrote:
> Hmm, I dunno about that plan. It's too soon to tell if that will work out. AM
> does a lot of things. I'm definitely supportive of paring it down to AS only
> being visible, but nominally that's what the AM interface is already supposed
to
> be.

Ok, thanks! Let's continue the discussion later via mail.
There are concerns regarding keeping AM public: it's supposed to live in another
process soon.

[o1ka, 2017-03-30 07:31:25]

olka@chromium.org changed reviewers:
+ tommi@chromium.org

[o1ka, 2017-03-30 07:32:14]

tommi@ - please RS speech

[tommi (sloooow) - chröme, 2017-03-30 07:56:53]

lgtm

[o1ka, 2017-03-30 12:30:12]

Thanks dalecurtis@ and tommi@!

avi@ please RS content/browser/browser_main_loop.cc

[Avi (use Gerrit), 2017-03-30 14:57:16]

lgtm

[o1ka, 2017-03-30 15:13:02]

Thanks dalecrutis@, tommi@ and avi@.

[o1ka, 2017-03-30 15:13:08]

The CQ bit was checked by olka@chromium.org

[commit-bot: I haz the power, 2017-03-30 15:13:50]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-03-30 15:24:14]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1490886788704200, "parent_rev":
"d5543110dfdaa0b8fbe7f4b69797cef6b0f5dca4", "commit_rev":
"b6082fccb8e9fc2f21871508bdcf79ca567326ad"}

[commit-bot: I haz the power, 2017-03-30 15:25:07]

Description was changed from

==========
Removing AudioSystem::GetAudioManager() usage from speech recognition.

After https://codereview.chromium.org/2763383002/ lands,
AudioSystem::GetAudioManager() will be removed: it's only purpose was to assist
in gradually switching clients from AudioManager to AudioSystem interface.
However, it leaked a bit in speech recognition, so - fixing it.

The plan is:

1) After the last client (WebRTC audio private API) is switched to AudioSystem
interface, all the related AudioManager methods will be made protected to be
used by AudioSystem only; so essentially AudioManager interface will become a
factory for audio streams.

2) Further, all audio stream clients will be switched to asynchronous stream
creation interface, and then to Mojo.

BUG=672468
==========

to

==========
Removing AudioSystem::GetAudioManager() usage from speech recognition.

After https://codereview.chromium.org/2763383002/ lands,
AudioSystem::GetAudioManager() will be removed: it's only purpose was to assist
in gradually switching clients from AudioManager to AudioSystem interface.
However, it leaked a bit in speech recognition, so - fixing it.

The plan is:

1) After the last client (WebRTC audio private API) is switched to AudioSystem
interface, all the related AudioManager methods will be made protected to be
used by AudioSystem only; so essentially AudioManager interface will become a
factory for audio streams.

2) Further, all audio stream clients will be switched to asynchronous stream
creation interface, and then to Mojo.

BUG=672468

Review-Url: https://codereview.chromium.org/2785463005
Cr-Commit-Position: refs/heads/master@{#460763}
Committed:
https://chromium.googlesource.com/chromium/src/+/b6082fccb8e9fc2f21871508bdcf...
==========

[commit-bot: I haz the power, 2017-03-30 15:25:08]

Committed patchset #1 (id:1) as
https://chromium.googlesource.com/chromium/src/+/b6082fccb8e9fc2f21871508bdcf...