Binding media stream audio track to speech recognition [renderer]

content/renderer/speech_recognition_dispatcher.h

Index: content/renderer/speech_recognition_dispatcher.h
diff --git a/content/renderer/speech_recognition_dispatcher.h b/content/renderer/speech_recognition_dispatcher.h
index bae0e5d92a484cbd706533010a8ffd28e3f9e5fa..9ac43f55ed6edbfeee9b5e6c635140fecab99d1a 100644
--- a/content/renderer/speech_recognition_dispatcher.h
+++ b/content/renderer/speech_recognition_dispatcher.h
@@ -8,14 +8,23 @@
 #include <map>
 
 #include "base/basictypes.h"
+#include "base/memory/shared_memory.h"
+#include "base/message_loop/message_loop_proxy.h"
 #include "content/public/common/speech_recognition_result.h"
 #include "content/public/renderer/render_view_observer.h"
+#include "third_party/WebKit/public/platform/WebMediaStreamTrack.h"
 #include "third_party/WebKit/public/platform/WebVector.h"
 #include "third_party/WebKit/public/web/WebSpeechRecognitionHandle.h"
 #include "third_party/WebKit/public/web/WebSpeechRecognizer.h"
 
+
+namespace media {
+class AudioParameters;
+}
+
 namespace content {
 class RenderViewImpl;
+class SpeechRecognitionAudioSourceProvider;
 struct SpeechRecognitionError;
 struct SpeechRecognitionResult;
 
@@ -36,6 +45,11 @@ class SpeechRecognitionDispatcher : public RenderViewObserver,
   virtual bool OnMessageReceived(const IPC::Message& message) OVERRIDE;
 
   // blink::WebSpeechRecognizer implementation.
+  virtual void attach(const blink::WebSpeechRecognitionHandle&,
+                             const blink::WebMediaStreamTrack&,
+                             blink::WebSpeechRecognizerClient*) OVERRIDE;
+  virtual void detach(const blink::WebSpeechRecognitionHandle&,
+                             blink::WebSpeechRecognizerClient*) OVERRIDE;
   virtual void start(const blink::WebSpeechRecognitionHandle&,
                      const blink::WebSpeechRecognitionParams&,
                      blink::WebSpeechRecognizerClient*) OVERRIDE;
@@ -53,6 +67,13 @@ class SpeechRecognitionDispatcher : public RenderViewObserver,
   void OnRecognitionEnded(int request_id);
   void OnResultsRetrieved(int request_id,
                           const SpeechRecognitionResults& result);
+  void OnAudioTrackReady(int session_id, const media::AudioParameters& params,
+                         base::SharedMemoryHandle handle, uint32 length);
+
+  // for receiving audio data via audio track

[henrika (OOO until Aug 14), 2014/08/25 14:46:00]

nit, capital 'F' end with period '.' as all other comments.

+  void OnAudioData(int request_id);
+  void OnAudioError(int request_id);
+  void OnAudioChunkProcessed(int request_id);
 
   int GetOrCreateIDForHandle(const blink::WebSpeechRecognitionHandle& handle);
   bool HandleExists(const blink::WebSpeechRecognitionHandle& handle);
@@ -61,6 +82,21 @@ class SpeechRecognitionDispatcher : public RenderViewObserver,
   // The WebKit client class that we use to send events back to the JS world.
   blink::WebSpeechRecognizerClient* recognizer_client_;
 
+  // Media stream audio track that the speech recognition connects to.
+  // Accessed on the render thread.
+  blink::WebMediaStreamTrack audio_track_;
+
+  bool audio_track_set_;

[henrika (OOO until Aug 14), 2014/08/25 14:46:00]

Comment.

...and, capitals for all comments below ;-)

+
+  // Used by policy on which audio sources are allowed

[henrika (OOO until Aug 14), 2014/08/25 14:46:00]

Rather unclear comment to me. What is meant by "used by policy"?

+  bool is_allowed_audio_track_;
+
+  // audio sink used to provide audio from the track
+  scoped_ptr<SpeechRecognitionAudioSourceProvider> audio_source_provider_;
+
+  // for sending signal back to browser
+  scoped_refptr<base::MessageLoopProxy> render_loop_;
+
   typedef std::map<int, blink::WebSpeechRecognitionHandle> HandleMap;
   HandleMap handle_map_;
   int next_id_;