Merge pull request #89 from benner/fix/message-file-utf8-encoding

fix(cli): read commit message file as UTF-8

Author: benner

src/git_commit_guard/__init__.py

@@ -706,7 +706,7 @@ def _parse_args():  # noqa: PLR0915 Too many statements (59 > 50)
         message = ""
     elif args.message_file:
         rev = None
-        message = _strip_comments(args.message_file.read_text().strip())
+        message = _strip_comments(args.message_file.read_text(encoding="utf-8").strip())
     elif args.rev:
         rev = args.rev
         message = _strip_comments(_get_message(rev))

tests/test_git_commit_guard.py

@@ -1260,6 +1260,18 @@ def test_from_message_file(self, tmp_path, capsys):
             assert main() == 0
         assert "all checks passed" in capsys.readouterr().out
 
+    def test_from_message_file_utf8_non_ascii(self, tmp_path):
+        f = tmp_path / "msg"
+        f.write_bytes(
+            "fix: handle non-ascii\n\nbody\n\n"
+            "Signed-off-by: Nerijus Bendžiūnas <a@b.com>".encode()
+        )
+        with patch(
+            "sys.argv",
+            ["cg", "--message-file", str(f), "--disable", "signature"],
+        ):
+            assert main() == 0
+
     def test_from_stdin(self):
         stdin = MagicMock()
         stdin.isatty.return_value = False