Update audio code to use size_t more correctly, webrtc/modules/audio_processing/

webrtc/modules/audio_processing/agc/agc.h

Index: webrtc/modules/audio_processing/agc/agc.h
diff --git a/webrtc/modules/audio_processing/agc/agc.h b/webrtc/modules/audio_processing/agc/agc.h
index dd4605e812e9cbbf14470da9ad96ea4f12d471bd..08c287f82057c613e3c3b27b9871b9255f35043b 100644
--- a/webrtc/modules/audio_processing/agc/agc.h
+++ b/webrtc/modules/audio_processing/agc/agc.h
@@ -27,10 +27,10 @@ class Agc {
 
   // Returns the proportion of samples in the buffer which are at full-scale
   // (and presumably clipped).
-  virtual float AnalyzePreproc(const int16_t* audio, int length);
+  virtual float AnalyzePreproc(const int16_t* audio, size_t length);
   // |audio| must be mono; in a multi-channel stream, provide the first (usually
   // left) channel.
-  virtual int Process(const int16_t* audio, int length, int sample_rate_hz);
+  virtual int Process(const int16_t* audio, size_t length, int sample_rate_hz);
 
   // Retrieves the difference between the target RMS level and the current
   // signal RMS level in dB. Returns true if an update is available and false