)]}'
{
  "commit": "d3bcf13165899ba1a2acfbe57cf4a1365010de39",
  "tree": "c7a7d94692e1f29c7169d4e6b6c781342ad14ef3",
  "parents": [
    "fcf25833bc4d057377da1ea755f57b2f5126e555"
  ],
  "author": {
    "name": "Nick Harper",
    "email": "nharper@chromium.org",
    "time": "Thu Dec 18 15:47:37 2014 -0800"
  },
  "committer": {
    "name": "Adam Langley",
    "email": "agl@google.com",
    "time": "Fri Dec 19 18:18:08 2014 +0000"
  },
  "message": "Properly handle key_len\u003d0 for HMAC\n\nThe expectation when calling HMAC with key\u003dNULL and keylen\u003d0 is to compute\nHMAC on the provided data with a key of length 0 instead of using the\n\"previous\" key, which in the case of HMAC() is whatever bytes happen to be\nleft on the stack when the HMAC_CTX struct is allocated.\n\nChange-Id: I52a95e262ee4e15f1af3136cb9c07f42f40ce122\nReviewed-on: https://boringssl-review.googlesource.com/2660\nReviewed-by: Adam Langley \u003cagl@google.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "1d37b9e46ecf950a8b23906a03594d744df992f7",
      "old_mode": 33188,
      "old_path": "crypto/hmac/hmac.c",
      "new_id": "bb96606c0703b6ab04a2647dd5e37aca975e5166",
      "new_mode": 33188,
      "new_path": "crypto/hmac/hmac.c"
    },
    {
      "type": "modify",
      "old_id": "7b851967fafb620b4e16285d373b13adca928565",
      "old_mode": 33188,
      "old_path": "crypto/hmac/hmac_test.c",
      "new_id": "def773b1a1da40dc7bc7ad10ece9215ef4afd3bd",
      "new_mode": 33188,
      "new_path": "crypto/hmac/hmac_test.c"
    }
  ]
}