)]}'
{
  "commit": "fda3549c01ea0a1d2ee00fe0033c35dc36796fc8",
  "tree": "2ef96e4c81fca716906b41bbaef5978394533a07",
  "parents": [
    "75d1b01b7fc49fae805514cd568c30691d135c14"
  ],
  "author": {
    "name": "Daniel Sanders",
    "email": "daniel.sanders@imgtec.com",
    "time": "Wed Feb 17 15:02:33 2016 +0000"
  },
  "committer": {
    "name": "Daniel Sanders",
    "email": "daniel.sanders@imgtec.com",
    "time": "Wed Feb 17 15:02:33 2016 +0000"
  },
  "message": "Merging r261088:\n------------------------------------------------------------------------\nr261088 | dsanders | 2016-02-17 13:16:31 +0000 (Wed, 17 Feb 2016) | 21 lines\n\n[libcxx] Fix definition of regex_traits::__regex_word on big-endian glibc systems\n\nSummary:\nOn glibc, the bits used for the various character classes is endian dependant\n(see _ISbit() in ctypes.h) but __regex_word does not account for this and uses\na spare bit that isn\u0027t spare on big-endian. On big-endian, it overlaps with the\nbit for graphic characters which causes \u0027-\u0027, \u0027@\u0027, etc. to be considered a word\ncharacter.\n\nFixed this by defining the value using _ISbit(15) on MIPS glibc systems. We\u0027ve\nrestricted this to MIPS for now to avoid the risk of introducing failures in\nother targets.\n\nFixes PR26476.\n\nReviewers: hans, mclow.lists\n\nSubscribers: dsanders, cfe-commits\n\nDifferential Revision: http://reviews.llvm.org/D17132\n\n------------------------------------------------------------------------\n\n\ngit-svn-id: https://llvm.org/svn/llvm-project/libcxx/branches/release_38@261097 91177308-0d34-0410-b5e6-96231b3b80d8\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "f1f3264d32020529ccf12f00adc088d56091b725",
      "old_mode": 33188,
      "old_path": "include/regex",
      "new_id": "1139d8fb2a90efb64223734d07e398800980be2f",
      "new_mode": 33188,
      "new_path": "include/regex"
    }
  ]
}