Create recorder streams from packet sinks ops

Previously, the packet sink push() implementation just set the codec and
notified a wait condition. Then the recorder thread read the codec and
created the AVStream.

But this was racy: an AVFrame could be pushed before the creation of the
AVStream, causing its video_stream_index or audio_stream_index to be
initialized to -1.

Also, in the future, the AVStream initialization might need data
provided by the packet sink open(), so initialize it there (with a
mutex).

app/src/recorder.h

@@ -43,11 +43,8 @@ struct sc_recorder {
 
     // wake up the recorder thread once the video or audio codec is known
     sc_cond stream_cond;
-    const AVCodec *video_codec;
-    const AVCodec *audio_codec;
-    // Instead of providing an audio_codec, the demuxer may notify that the
-    // stream is disabled if the device could not capture audio
-    bool audio_disabled;
+    bool video_init;
+    bool audio_init;
 
     int video_stream_index;
     int audio_stream_index;