)]}'
{
  "commit": "6d0ae84a581ca100edb2ef38cffb425ee75083bf",
  "tree": "863d844fc2e5026b410b0dcc28cfd5810223138f",
  "parents": [
    "9981b66f87c0705b2123de637454603c46823da9"
  ],
  "author": {
    "name": "Leonardo de Moura",
    "email": "leomoura@amazon.com",
    "time": "Tue Oct 20 19:18:28 2026 +0000"
  },
  "committer": {
    "name": "Leonardo de Moura",
    "email": "leomoura@amazon.com",
    "time": "Tue Oct 20 19:18:28 2026 +0000"
  },
  "message": "fix: canonicalize `OfNat` instances of nested numerals in `grind`\n\nThis PR fixes `grind` producing a proof the kernel rejects (`decide (4 \u003d 4) \u003d false`) when a numeral is written as `OfNat.ofNat 4`.\n\nThe canonicalizer replaced the nested numeral argument of `OfNat.ofNat` with a raw literal, but kept the instance `instOfNatNat (OfNat.ofNat 4)`: the instance was resynthesized for its own type, which still mentions the nested numeral, so resynthesis returned it unchanged. The two spellings of `4` then became distinct interpreted nodes in the E-graph, and `grind` closed the goal using the incorrect fact that they denote distinct values. The instance is now resynthesized for `OfNat α n` with the canonical `α` and `n`. In addition, nested numerals of any depth, such as `@OfNat.ofNat Nat (@OfNat.ofNat Nat 4 _) _`, are now reduced to a raw literal.\n\nCloses #15604\n\nCo-Authored-By: Claude Opus 5.5 \u003cnoreply@anthropic.com\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "89ad1682969186417b7abcd13a2bc0aa6e04e1be",
      "old_mode": 33188,
      "old_path": "src/Lean/Meta/Sym/Canon.lean",
      "new_id": "296c06589876eca196f25a5ac36dd45138f27198",
      "new_mode": 33188,
      "new_path": "src/Lean/Meta/Sym/Canon.lean"
    },
    {
      "type": "add",
      "old_id": "0000000000000000000000000000000000000000",
      "old_mode": 0,
      "old_path": "/dev/null",
      "new_id": "ae70415acf119c53d5448376e52d1f6b475b1332",
      "new_mode": 33188,
      "new_path": "tests/elab/grind_15541.lean"
    }
  ]
}
