The parameters for Logging.enableTracing() were creating the impression that they control level and severity of one tracing system and they are meant to be used together. In fact the "levels" parameter controlled one tracing system (WEBRTC_TRACE), and the "severity" parameter was responsible for a completely different one: setting the severity level above which log messages from LOG() will be directed to the platform-specific debug output (logcat on Android). The method signature suggested that the "path" parameter applied to both systems - while it was only meaningful for the WEBRTC_TRACE; LOG messages were directed to ADB logcat no matter what the Path value was. It is possible to redirect LOG messages to a file, but that is done using a completely different set of APIs - PeerConnectionFactory.startInternalTracingCapture(). I've separated these two methods to make it more clear which of the parameters controls which system. NOTRY=true Review-Url: https://codereview.webrtc.org/2110853003 Cr-Commit-Position: refs/heads/master@{#13334}
This directory holds a Java implementation of the webrtc::PeerConnection API, as
well as the JNI glue C++ code that lets the Java implementation reuse the C++
implementation of the same API.
To build the Java API and related tests, build with OS=android in $GYP_DEFINES.
To use the Java API, start by looking at the public interface of
org.webrtc.PeerConnection{,Factory} and the org.webrtc.PeerConnectionTest.
To understand the implementation of the API, see the native code in jni/.