AecDump interface

webrtc/media/engine/webrtcvoiceengine.cc

Index: webrtc/media/engine/webrtcvoiceengine.cc
diff --git a/webrtc/media/engine/webrtcvoiceengine.cc b/webrtc/media/engine/webrtcvoiceengine.cc
index 8fcda89962ee768153b9d6337a862309b7e67aab..774b0af0f689879220e19f8d3e3479b31d3ca8ec 100644
--- a/webrtc/media/engine/webrtcvoiceengine.cc
+++ b/webrtc/media/engine/webrtcvoiceengine.cc
@@ -999,7 +999,8 @@ bool WebRtcVoiceEngine::StartAecDump(rtc::PlatformFile file,
     return false;
   }
   StopAecDump();
-  if (apm()->StartDebugRecording(aec_dump_file_stream, max_size_bytes) !=
+  if (apm()->StartDebugRecording(aec_dump_file_stream, max_size_bytes,

[the sun, 2017/03/28 22:35:29]

Could we get rid of these calls on the APM and just have one which sets the
AecDumper implementation? Then we could just have one method on
WebRtcVoiceEngine too - creating a dumper for either of rtc::PlatformFile or a
file name would be handled higher up the stack and we'd not need to know about
such details here...

[aleloi, 2017/04/06 15:46:10]

Sure! I've added a new APM::StartDebugRecording method that takes an AecDump.
However, I havn't changed anything here and not deprecated the old ones. This CL
doesn't contain the implementation of AecDump yet, and
StartDebugRecording(AecDump) does nothing. Therefore the old calls are still
here.

+                                 nullptr) !=
       webrtc::AudioProcessing::kNoError) {
     LOG_RTCERR0(StartDebugRecording);
     fclose(aec_dump_file_stream);
@@ -1013,7 +1014,7 @@ void WebRtcVoiceEngine::StartAecDump(const std::string& filename) {
   RTC_DCHECK(worker_thread_checker_.CalledOnValidThread());
   if (!is_dumping_aec_) {
     // Start dumping AEC when we are not dumping.
-    if (apm()->StartDebugRecording(filename.c_str(), -1) !=
+    if (apm()->StartDebugRecording(filename.c_str(), -1, nullptr) !=
         webrtc::AudioProcessing::kNoError) {
       LOG_RTCERR1(StartDebugRecording, filename.c_str());
     } else {