)]}'
{
  "log": [
    {
      "commit": "656ecc6676367d7483667550fa91046d7a8c0719",
      "tree": "4f934b1feda7c7032da3a62d09f4e101929ff5fa",
      "parents": [
        "313c4be4cfc0b0e1bb2551efed6489333b195bfb"
      ],
      "author": {
        "name": "Zsolt Haraszti",
        "email": "zharaszt@ciena.com",
        "time": "Wed Dec 28 15:08:23 2016 -0800"
      },
      "committer": {
        "name": "Zsolt Haraszti",
        "email": "zharaszt@ciena.com",
        "time": "Thu Dec 29 00:15:43 2016 -0800"
      },
      "message": "PONSIM: PON simulator with real dataplane handling\n\nThis was needed because neither CPQD nor OVS can handle\nboth zero-tagged packets and 802.1ad (QinQ).\n\n- extensive unittest proves ponsim functional correctness\n  (for the common use-cases needed in the PON scenario)\n- integrated with frameio and coupled with a rather\n  simple gRPC NBI, ponsim can be operated from Voltha\n  just like a real PON system\n- posim_olt/_onu adapters added to Voltha to work on\n  ponsim\n- CLI can be used to preprovision and activate a PONSIM\n  instance (e.g., preprovision_olt -t ponsim_olt -H localhost:50060)\n- Some of olt-oftest:olt-complex testcases can be run on\n  the ponsim device (in vagrant/Ubuntu environment),\n  but there are some remaining issues to work out:\n  - barrier calls in OF do not guaranty that the flow\n    is already installed on the device. This is a generic\n    issue, not just for ponsim.\n  - the whole test framework is inconsistent about zero-\n    tagged vs. untagged frames at the ONUs, while ponsim\n    is rather pedantica and does exactly what was defined\n    in the flows.\n\nChange-Id: I0dd564c932416ae1566935492134cb5b08113bdc\n"
    }
  ]
}