)]}'
{
  "commit": "cb208105eb5674fcc0fa2a4999ea0cc4f06bc8c6",
  "tree": "7d8af3dc06a6ab979edad13d902738dbcce34a00",
  "parents": [
    "52f0642b075c9f81c63f554203cee84cbc1f5bf8"
  ],
  "author": {
    "name": "Simon Marchi",
    "email": "simon.marchi@efficios.com",
    "time": "Wed Jul 09 11:35:12 2025 -0400"
  },
  "committer": {
    "name": "Simon Marchi",
    "email": "simon.marchi@polymtl.ca",
    "time": "Fri Aug 01 00:25:54 2025 -0400"
  },
  "message": "gdb/dwarf: sort units when writing index\n\nThe order of all_units can\u0027t be relied on when writing the CU and TU\nlists to .gdb_index or .debug_names.\n\nBoth the .gdb_index and .debug_names writers expect that all_units\ncontains comp units followed by type units.  As of this commit, when\nreading a DWARF 5 .debug_info, the all_units vector is ordered based on\nthe order the units appear in .debug_info, where type units can be\ninterleaved with comp units.\n\nIt probably worked fine with DWARF 4, where type units were in a section\nof their own (.debug_types).  They were read after comp units, and\ntherefore after them in the all_units vector.\n\nChange the writers to use a common function that splits the units in two\nlists (comp units and type units).  Sort both lists by section offset.\nThis is more than required, but it should help produce a stable and\npredictable output.\n\nChange-Id: I5a22e2e354145e3d6b5b2822dc2a3af2f9d6bb76\nApproved-By: Tom Tromey \u003ctom@tromey.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "0efd6a20e5e42b88c91a412c7ed734c17fdc4527",
      "old_mode": 33188,
      "old_path": "gdb/dwarf2/index-write.c",
      "new_id": "d5b198ba34734480573e3cb29a613764576c4baa",
      "new_mode": 33188,
      "new_path": "gdb/dwarf2/index-write.c"
    }
  ]
}