)]}'
{
  "commit": "89230f3595abdfca1e3f7e5621a1d6cc974a79c0",
  "tree": "baf507adb197a1cfad4ef187d7497b3b1a470bf6",
  "parents": [
    "95b7c2f3e060746a4bb7d5a3845a2dc86f19f56b"
  ],
  "author": {
    "name": "Devin Jeanpierre",
    "email": "jeanpierreda@google.com",
    "time": "Wed Mar 15 12:27:02 2023 -0700"
  },
  "committer": {
    "name": "Copybara-Service",
    "email": "copybara-worker@google.com",
    "time": "Wed Mar 15 12:27:38 2023 -0700"
  },
  "message": "Handle colon-less and relative bazel labels in `ir.rs`.\n\nThis has been bugging me for a while now -- not because I\u0027d predict it to be a real issue, but just because it feels weird to rely on a specific formatting of bazel labels. This has bitten me in the past with e.g. unknown commit.\n\nI restrained myself until now, but it\u0027s still kind of bothering me that we crash on `//foo` syntax. I took the most expedient option of fixing it the simplest way I know how.\n\nThe alternative would have been to make the field private, and validate the input at parsing time (e.g. defining `TryFrom` instead of `From`, and validating that the label meets basic validity checks. I chose not to because it\u0027s a lot more work than to accept invalid labels without crashing.\n\nThis probably won\u0027t actually affect anything, but if we ever find ourselves forwarding bazel labels along using non-canonicalized form of `//foo:bar`, we will DTRT. Hopefully that doesn\u0027t hide any problems.\n\nPiperOrigin-RevId: 516897644\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "c16a44403216a9430f1402cef8cd9e6e3275791c",
      "old_mode": 33188,
      "old_path": "rs_bindings_from_cc/ir.rs",
      "new_id": "e7f0e94003701d8ab7a8bc7e657223f9708350d0",
      "new_mode": 33188,
      "new_path": "rs_bindings_from_cc/ir.rs"
    }
  ]
}