)]}'
{
  "log": [
    {
      "commit": "2c3b2da3049627264b7c6b449a1622f002210f03",
      "tree": "f99fadbfef22ff3b7709a505623371c22e39566c",
      "parents": [
        "ee578c0330319f04a48bccbdb26b53fea0388d04"
      ],
      "author": {
        "name": "John Grossman",
        "email": "johngro@google.com",
        "time": "Thu Aug 02 17:08:54 2012 -0700"
      },
      "committer": {
        "name": "John Grossman",
        "email": "johngro@google.com",
        "time": "Wed Aug 08 11:28:15 2012 -0700"
      },
      "message": "AudioFlinger: fix timed audio\n\n(cherry picked from commit e20ac92c564a2f4e8123885807abdf0a78de0dd7)\n\n\u003e AudioFlinger: fix timed audio\n\u003e\n\u003e Addresses Bug 6900517.\n\u003e\n\u003e Finish up support for timed audio in the new FastMixer world.  Pay special\n\u003e attention to remaining lock-less and voluntary yield free on the FastMixer\n\u003e thread.  This fixes audio playback for Q on JB.\n\u003e\n\u003e Change-Id: Iaf815e58a1b1d0a0190051794bec8dc5c9231785\n\u003e Signed-off-by: John Grossman \u003cjohngro@google.com\u003e\n\nChange-Id: I9bd687acc345a05867af48e71116690fdb0ce1b5\nSigned-off-by: John Grossman \u003cjohngro@google.com\u003e\n"
    },
    {
      "commit": "288ed2103d96f3aabd7e6bea3c080ab6db164049",
      "tree": "dec8e7b3a993e86baa8916ba70dcd6b3c053d062",
      "parents": [
        "0c0abd4ad26971b5fba94734137fe0bb1a590ab6"
      ],
      "author": {
        "name": "Glenn Kasten",
        "email": "gkasten@google.com",
        "time": "Wed Apr 25 17:52:27 2012 -0700"
      },
      "committer": {
        "name": "Glenn Kasten",
        "email": "gkasten@google.com",
        "time": "Mon Apr 30 13:00:31 2012 -0700"
      },
      "message": "Fix race condition for non-started fast tracks\n\nThis required re-implementing how fast tracks are considered active.\nNow, they use the same logic as normal tracks, except underrun is ignored.\n\nOther changes:\n - add framesReady() to AudioBufferProvider interface\n - rebased\n - add track underrun counter state to fast mixer dump state\n - move dumpsys header to Track::appendDumpHeader()\n   so it closer to where tracks are dumped\n - display track state in dumpsys as a character code\n - measure and display warmup time and cycles in dumpsys\n - copy in the presentation complete code\n - add ExtendedAudioBufferProvider for framesReady() which returns size_t\n - simplify underrun tracking\n - deferred reset track after stop()\n - add comments\n\nChange-Id: I7db8821bc565230ec76da1f9380fe3fb09735e5b\n"
    },
    {
      "commit": "010662326b9c43c703725f933e95e0897f8a6bdd",
      "tree": "29a39db792c50d29f266d9df86af2d1de2d4c752",
      "parents": [
        "a1472d9883e35edd280201c8be3191695007dfd4"
      ],
      "author": {
        "name": "Glenn Kasten",
        "email": "gkasten@google.com",
        "time": "Mon Feb 27 11:50:44 2012 -0800"
      },
      "committer": {
        "name": "Glenn Kasten",
        "email": "gkasten@google.com",
        "time": "Mon Apr 02 07:44:17 2012 -0700"
      },
      "message": "Non-blocking audio I/O interface, WIP\n\nYet another abstraction similar to AudioTrack::Buffer and AudioBufferProvider,\nbut with support for streaming, non-blocking, and eventually PTS.\n\nThis is intended to be used as follows:\n - primary HAL output stream implements a Sink\n - primary HAL input stream implements a Source\n - Pipe implements a Sink\n - PipeReader implements a Source or TimedSource (not shown yet),\n   which supports \"read at PTS\"\n - fast AudioTrack on server side will implement a Source using cblk\n - normal AudioTrack on server side will not be changed initially\n - fast AudioRecord on server side will implement a Sink using cblk\n - normal AudioRecord on server side will not be changed initially\n - fast mixer thread will read from Sources and write to a Sink,\n   or (unlikely) implement a Source and multiple Sinks\n - Visualization and PCM logger will read from Source or TimedSource\n - A2DP normal mixer will be connected directly to its output stream\n   and there will be a kind of OutputTrack for duplication that will\n   read from a Sink with non-blocking write fed by the fast mixer.\n\nPatch set 3 changes:\n - Add more implementations of NBAIO interfaces:\n   added SourceAudioBufferProvider, MonoPipe, MonoPipeReader.\n - Added Format_sampleRate and Format_channelCount.\n - Extract out the roundUp() method.\n - Respond to most comments from previous code review.\n - The new classes are untested.\n\nPatch set 4 changes:\n - Fix bugs in MonoPipe::write() and MonoPipeReader::read()\n - Fix bug initializing mFrameBitShift too early\n - renamed roundUp() to roundup()\n - Fix Android.mk\n - Add LOG_TAG an LOG_NDEBUG, use ALOG_ASSERT and utils/Log.h instead of assert\n - Fix build warnings\n - Move constructor and destructor bodies from .h to .cpp\n - Line length 100\n - Following naming conventions for #include double-include protector macros\n - Include what you use\n - More NBAIO logging\n - MonoPipe write can be blocking\n\nPatch set 5 changes:\n - Address code review comments\n - Use a static library so unused implementations don\u0027t take memory\n - Comment out libsndfile dependency\n - Remove debugging LOGV and LOG_NDEBUG\n\nPatch set 6 changes (would be 6 at old location, actually 2 at new location):\n - Address code review comments on patchset 5\n - For MonoPipe, allow the full pipe to be used, no need to omit one slot\n - Don\u0027t do atomic releasing stores unless needed\n\nStill to do:\n - I\u0027m not happy with the Pipe class names\n - Update build/ for new static library?\n\nChange-Id: Ie6c61f05ce06b676b033be448a8ef9025a2ffcfd\n"
    }
  ]
}