)]}'
{
  "commit": "9c34454bf50b7e08d750a0e0fa9123e9d4ff2b90",
  "tree": "0d903752ff4fbaca78a2d0f184c7df00e974649f",
  "parents": [
    "05e43daf7e8d716a45a02e7fe9b3da8da968b6de"
  ],
  "author": {
    "name": "Jonathan Drolet",
    "email": "jonathan.drolet@gmail.com",
    "time": "Mon Aug 03 21:44:24 2026 -0400"
  },
  "committer": {
    "name": "Jonathan Drolet",
    "email": "jonathan.drolet@gmail.com",
    "time": "Thu Aug 06 22:40:55 2026 -0400"
  },
  "message": "Fix MSVC crash in FindSubtreeStartingAtOffsetTest: avoid constexpr substr provenance bug\n\nkFindSubtreeTestSubstring was a namespace-scope const std::string_view\ninitialized from kFindSubtreeTestText.substr(1, 3), a constant expression\n(kFindSubtreeTestText is constexpr). Under MSVC this constant-folds into an\nindependently materialized string literal elsewhere in .rdata, instead of\npreserving that the result is a view into kFindSubtreeTestText\u0027s own\nstorage -- confirmed via cdb: the two buffers were over 45KB apart in the\ncompiled binary.\n\nFindSubtreeStartingAtOffset() then correctly (given the corrupted input)\nreturned nullptr, which the test dereferenced, causing an SEH access\nviolation.\n\nFix: compute the substring at runtime inside the test fixture constructor,\nusing pointer arithmetic against the already-live kFindSubtreeTestText\nobject instead of a separately compile-time-folded global.\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "ace228482e38d558a0f601bbc23caff16af73768",
      "old_mode": 33188,
      "old_path": "verible/common/text/tree-utils_test.cc",
      "new_id": "0a775309dea6865f7f058a2c451832ab7ec87d0a",
      "new_mode": 33188,
      "new_path": "verible/common/text/tree-utils_test.cc"
    }
  ]
}