)]}'
{
  "log": [
    {
      "commit": "5beb4810060cda2afea1c89f98caf5c337763c0a",
      "tree": "377deb29ea3236a151adf09f32806b79f642304c",
      "parents": [
        "bcfd6df4b678246993f3c6ae12c0027aba041aa2"
      ],
      "author": {
        "name": "Mikhail Naganov",
        "email": "mnaganov@google.com",
        "time": "Fri Dec 14 10:48:17 2018 -0800"
      },
      "committer": {
        "name": "Mikhail Naganov",
        "email": "mnaganov@google.com",
        "time": "Fri Dec 14 10:48:17 2018 -0800"
      },
      "message": "audioserver: Use \u0027_exit\u0027 instead of \u0027exit\u0027 in HalDeathHandler\n\nUsing \u0027exit\u0027 from an RPC threadpool thread is not safe, as \u0027exit\u0027\nruns atexit handlers that destroy global objects. This can interfere\nwith code still running on other threads.\n\n\u0027_exit\u0027 does not run atexit handlers, just terminates the process.\n\nBug: 116665972\nTest: kill android.hardware.audio@2.0-service, check logcat\nChange-Id: I5391a659e359e0ca5bba580f1c51dea5df3ea562\n"
    },
    {
      "commit": "7f16ea4d1f6554f653babd0a63b7e97ce1061826",
      "tree": "df8f17772ce517fda38d1120e849effc463e0d9e",
      "parents": [
        "a4fa4760fa417473324e009b3f4117a2d539c486"
      ],
      "author": {
        "name": "Mikhail Naganov",
        "email": "mnaganov@google.com",
        "time": "Thu Nov 16 08:29:49 2017 -0800"
      },
      "committer": {
        "name": "Mikhail Naganov",
        "email": "mnaganov@google.com",
        "time": "Thu Nov 16 08:31:52 2017 -0800"
      },
      "message": "libaudiohal: Use \u0027exit\u0027 in HalDeathHandler instead of fatal log\n\nFatal log is registered as a process crash. Since these crash\nreports do not bring in any useful info, it\u0027s better to just\nexit the process instead.\n\nBug: 69379459\nTest: kill audio HAL, check logcat\nChange-Id: I0d83d496998248faf3669a1fe8c31944b839bb3e\n"
    },
    {
      "commit": "d621ac82e648c8ef395068edb1af2747f49b700a",
      "tree": "3c37b8d97c0c7feec6c567d407d05c8343ec9e6e",
      "parents": [
        "a72ef28f8cc2b894119b470fe6c93d1abe566398"
      ],
      "author": {
        "name": "Mikhail Naganov",
        "email": "mnaganov@google.com",
        "time": "Thu Jan 12 17:17:45 2017 -0800"
      },
      "committer": {
        "name": "Mikhail Naganov",
        "email": "mnaganov@google.com",
        "time": "Tue Jan 17 13:42:54 2017 -0800"
      },
      "message": "Monitor crashes of audio hal service via linkToDeath\n\nHwbinder \u0027linkToDeath\u0027 mechanism is used to track abrupt\ndisappearance of server hal objects due to hal server crash.\n\nUpon hal server crash, notifications can be sent to binder\nclients if needed, then the audioserver crashes itself as the\ncurrent audio HAL design does not provide an easy way to restore\nall the transient objects (devices, streams, effects) in their\npre-crash state.\n\nBug: 34158778\nTest: In binderized mode, kill audio hal server\nChange-Id: If9265711edc445984601fcb5c29d2c173a370762\n"
    }
  ]
}