Added an empty AudioTransportProxy to AudioState.

webrtc/audio/audio_state.h

 /*
  *  Copyright (c) 2015 The WebRTC project authors. All Rights Reserved.
  *
  *  Use of this source code is governed by a BSD-style license
  *  that can be found in the LICENSE file in the root of the source
  *  tree. An additional intellectual property rights grant can be found
  *  in the file PATENTS.  All contributing project authors may
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
 #ifndef WEBRTC_AUDIO_AUDIO_STATE_H_
 #define WEBRTC_AUDIO_AUDIO_STATE_H_
 
+#include <memory>
+
+#include "webrtc/api/audio/audio_mixer.h"
 #include "webrtc/api/call/audio_state.h"
+#include "webrtc/audio/audio_transport_proxy.h"
 #include "webrtc/audio/scoped_voe_interface.h"
 #include "webrtc/base/constructormagic.h"
 #include "webrtc/base/criticalsection.h"
 #include "webrtc/base/thread_checker.h"
+#include "webrtc/modules/audio_device/include/audio_device_defines.h"
 #include "webrtc/voice_engine/include/voe_base.h"
 
 namespace webrtc {
 namespace internal {
 
 class AudioState final : public webrtc::AudioState,
                          public webrtc::VoiceEngineObserver {
  public:
   explicit AudioState(const AudioState::Config& config);
   ~AudioState() override;
 
   VoiceEngine* voice_engine();
+
+  rtc::scoped_refptr<AudioMixer> mixer() const;

[ossu, 2016/11/03 14:02:04]

I believe this should just return an AudioMixer& (or possibly an AudioMixer*,
but can it ever be null?) C.f. the discussion in the previous CL about
unique_ptr.

[aleloi, 2016/11/08 11:46:41]

We decided to leave the mixer refcounted for the time being.

   bool typing_noise_detected() const;
 
  private:
   // rtc::RefCountInterface implementation.
   int AddRef() const override;
   int Release() const override;
 
   // webrtc::VoiceEngineObserver implementation.
   void CallbackOnError(int channel_id, int err_code) override;
 
+  // The Audio Device currently connected or nullptr.
+  AudioDeviceModule* audio_device();
+
   rtc::ThreadChecker thread_checker_;
   rtc::ThreadChecker process_thread_checker_;
   const webrtc::AudioState::Config config_;
 
   // We hold one interface pointer to the VoE to make sure it is kept alive.
   ScopedVoEInterface<VoEBase> voe_base_;
 
   // The critical section isn't strictly needed in this case, but xSAN bots may
   // trigger on unprotected cross-thread access.
   rtc::CriticalSection crit_sect_;
   bool typing_noise_detected_ GUARDED_BY(crit_sect_) = false;
 
   // Reference count; implementation copied from rtc::RefCountedObject.
   mutable volatile int ref_count_ = 0;
 
+  // Transports mixed audio from the mixer to the audio device and
+  // recorded audio to the VoE AudioTransport.
+  std::unique_ptr<AudioTransportProxy> audio_transport_proxy_;

[ossu, 2016/11/03 14:02:04]

Why does the AudioTransportProxy need to be allocated on the heap?

[aleloi, 2016/11/08 11:46:41]

You are right, it doesn't! But now initialization of AudioState is a little
different and happens in another order. In an earlier iteration of this CL the
transport was not always initialized. Now that I changed the tests to always
provide a transport it is and we don't need a pointer.

+
   RTC_DISALLOW_IMPLICIT_CONSTRUCTORS(AudioState);
 };
 }  // namespace internal
 }  // namespace webrtc
 
 #endif  // WEBRTC_AUDIO_AUDIO_STATE_H_