)]}'
{
  "commit": "bf42afa0df142896bea074f0d4d1d4fd6d3309ab",
  "tree": "89b7e08bc7b3ae56c764107a3af303af3662b951",
  "parents": [
    "f7e697617c77a7d8a1c47db178b095e11368171a"
  ],
  "author": {
    "name": "Tom Sepez",
    "email": "tsepez@chromium.org",
    "time": "Wed Feb 13 23:26:06 2019 +0000"
  },
  "committer": {
    "name": "Chromium commit bot",
    "email": "commit-bot@chromium.org",
    "time": "Wed Feb 13 23:26:06 2019 +0000"
  },
  "message": "Handle creation of CGFAS_Decimals from floats with bigger values.\n\nAlas, the current implementation is 96 bits; we\u0027d need 128 bits to\ncover the full range of an IEEE float.\n\nDividing a float by 1e32 does not produce a 32-bit shift. The current\ncode \"works\" because 64 bits are accidentally carried into |plo| with\n|phi, pmid| getting 0s (mercifully) from the botched divisions (except\nin some large cases) and then redistributed.\n\nBeef up some tests while we\u0027re at it.\n\nChange-Id: I01922d368c52008f9e242909820afeaae796da5f\nReviewed-on: https://pdfium-review.googlesource.com/c/50630\nReviewed-by: Lei Zhang \u003cthestig@chromium.org\u003e\nCommit-Queue: Tom Sepez \u003ctsepez@chromium.org\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "4783ca1d1091032798b50ff24f8035ae5273120a",
      "old_mode": 33188,
      "old_path": "xfa/fgas/crt/cfgas_decimal.cpp",
      "new_id": "9765c93061d1e5a64c4ba6a3ed961785eafa2bfd",
      "new_mode": 33188,
      "new_path": "xfa/fgas/crt/cfgas_decimal.cpp"
    },
    {
      "type": "modify",
      "old_id": "b1581b51ca7e444123e20a8b53b14ced37fb6e6a",
      "old_mode": 33188,
      "old_path": "xfa/fgas/crt/cfgas_decimal_unittest.cpp",
      "new_id": "5b714e32debf5efe26c40daf767116b5943fe8b0",
      "new_mode": 33188,
      "new_path": "xfa/fgas/crt/cfgas_decimal_unittest.cpp"
    }
  ]
}