)]}'
{
  "commit": "d7be62c714fef2c2bf00bf1e173a2d1fc50fddb5",
  "tree": "418d8640bd12c56f0a1014000c43ed09f554b60f",
  "parents": [
    "119ebdf7496e51d583789d4e5a5d6964bb4aca7a"
  ],
  "author": {
    "name": "bnc",
    "email": "bnc@google.com",
    "time": "Tue Oct 06 07:22:37 2020 -0700"
  },
  "committer": {
    "name": "Copybara-Service",
    "email": "copybara-worker@google.com",
    "time": "Tue Oct 06 07:23:18 2020 -0700"
  },
  "message": "Improve http2::HuffmanEncodeFast() by skipping rest of iteration if possible.\n\nWhile I originally thought that a loop without branches results in fastest\nencoding, benchmarks show that skipping to the next iteration if the Huffman\ncode has ended actually speeds up encoding.  This is probably due to the fact\nthat most codes are short (by design of the Huffman code).\n\nOf course this encoding algorithm would be a lot simpler if it was only for\n64-bit big endian platforms, as a single |\u003d instruction would do the job.\n\nPiperOrigin-RevId: 335634321\nChange-Id: Ia95c52b15c3d50600e40cd0beef914773bf2c068\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "4331f1c7e489122878d611caa9eda835345c8388",
      "old_mode": 33188,
      "old_path": "http2/hpack/huffman/hpack_huffman_encoder.cc",
      "new_id": "a5eeaf91e4c9ffff2eb3f0941d69881c07e2ab1a",
      "new_mode": 33188,
      "new_path": "http2/hpack/huffman/hpack_huffman_encoder.cc"
    }
  ]
}