Allowing a custom audio buffer size in WebRtcAudioCapturer

content/renderer/media/webrtc_audio_capturer.h

Index: content/renderer/media/webrtc_audio_capturer.h
diff --git a/content/renderer/media/webrtc_audio_capturer.h b/content/renderer/media/webrtc_audio_capturer.h
index ca10ff2ab399b782d9f1749e796c21540134056e..f1ada132fd5c97de08addb87a2c0bedb30f8ff5a 100644
--- a/content/renderer/media/webrtc_audio_capturer.h
+++ b/content/renderer/media/webrtc_audio_capturer.h
@@ -139,10 +139,12 @@ class CONTENT_EXPORT WebRtcAudioCapturer
   // desires to provide their own captured audio data. Client is responsible
   // for calling Start() on its own source to get the ball rolling.
   // Called on the main render thread.
+  // buffer_size is optional. Set to 0 to let it be chosen automatically.

[henrika (OOO until Aug 14), 2015/05/14 15:19:51]

Perhaps mention kDontUseBufferSizeParameter

[Charlie, 2015/05/14 17:30:53]

I think Tommi prefers the explicit zero.

   void SetCapturerSourceInternal(
       const scoped_refptr<media::AudioCapturerSource>& source,
       media::ChannelLayout channel_layout,
-      float sample_rate);
+      int sample_rate,
+      int buffer_size);
 
   // Starts recording audio.
   // Triggered by AddSink() on the main render thread or a Libjingle working