)]}'
{
  "commit": "2d458da9a8e1927275589c308ae19eae577d2b3c",
  "tree": "1477c03dffaffa478501a299939f0c5ad2db8e85",
  "parents": [
    "c6660615d17e89f45ac0c8a4d62e6c285f1c5636"
  ],
  "author": {
    "name": "Tom de Vries",
    "email": "tdevries@suse.de",
    "time": "Thu Apr 03 17:13:12 2025 +0200"
  },
  "committer": {
    "name": "Tom de Vries",
    "email": "tdevries@suse.de",
    "time": "Thu Apr 03 17:13:12 2025 +0200"
  },
  "message": "[gdb/testsuite] Fix xfail in gdb.ada/array_of_variant.exp\n\nIn commit af2b87e649b (\"[gdb/testsuite] Add xfail for PR gcc/101633\"), I added\nan xfail that was controlled by variable old_gcc, triggering the xfail for\ngcc 7 and before, but not for gcc 8 onwards:\n...\nset old_gcc [expr [test_compiler_info {gcc-[0-7]-*}]]\n...\n\nIn commit 1411185a57e (\"Introduce and use gnat_version_compare\"), this changed\nto:\n...\nset old_gcc [gnat_version_compare \u003c\u003d 7]\n...\nwhich still triggered the xfail for gcc 7, because of a bug in\ngnat_version_compare.\n\nAfter that bug got fixed, the xfail was no longer triggered because the gnatmake\nversion is 7.5.0, and [version_compare {7 5 0} \u003c\u003d {7}] \u003d\u003d 0.\n\nWe could have the semantics for version_compare where we clip the input\narguments to the length of the shortest, and so we\u0027d have\n[version_compare {7 5 0} \u003c\u003d {7}] \u003d\u003d [version_compare {7} \u003c\u003d {7}] \u003d\u003d 1.\n\nBut let\u0027s stick with the current version-sort semantics, and fix this by\nusing [gnat_version_compare \u003c 8] instead.\n\nTested on x86_64-linux.\n\nApproved-By: Tom Tromey \u003ctom@tromey.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "8b83f4ed22adbc0bb06162348b7ba2e9fc6f2c51",
      "old_mode": 33188,
      "old_path": "gdb/testsuite/gdb.ada/array_of_variant.exp",
      "new_id": "83b626cb6d1819b738c0382507060fae050ef9a0",
      "new_mode": 33188,
      "new_path": "gdb/testsuite/gdb.ada/array_of_variant.exp"
    }
  ]
}