fix: read cassette test sources as utf-8

scripts/check_cassettes.py

@@ -56,7 +56,7 @@ def _has_module_vcr_marker(tree: ast.Module) -> bool:
 def _collect_vcr_tests_from_file(path: Path) -> set[str]:
     """Parse a Python test file and return cassette names for VCR-marked tests."""
     try:
-        tree = ast.parse(path.read_text())
+        tree = ast.parse(path.read_text(encoding='utf-8'))
     except SyntaxError:
         return set()
 

tests/test_check_cassettes.py

@@ -0,0 +1,21 @@
+from pathlib import Path
+
+from scripts.check_cassettes import _collect_vcr_tests_from_file
+
+
+def test_collect_vcr_tests_reads_utf8_source(tmp_path: Path) -> None:
+    test_file = tmp_path / 'test_unicode_source.py'
+    test_file.write_text(
+        """\
+import pytest
+
+
+@pytest.mark.vcr
+def test_unicode_source_comment():
+    message = 'Olá, 世界'
+    assert message
+""",
+        encoding='utf-8',
+    )
+
+    assert _collect_vcr_tests_from_file(test_file) == {'test_unicode_source_comment'}