gracefully handles binary files (#370)

detect_secrets/core/scan.py

@@ -46,9 +46,13 @@ def scan_file(filename: str) -> Generator[PotentialSecret, None, None]:
         with open(filename) as f:
             log.info(f'Checking file: {filename}')
 
-            lines = _get_transformed_file(f)
-            if not lines:
-                lines = f.readlines()
+            try:
+                lines = _get_transformed_file(f)
+                if not lines:
+                    lines = f.readlines()
+            except UnicodeDecodeError:
+                # We flat out ignore binary files.
+                return
 
             has_secret = False
             for secret in _process_line_based_plugins(

detect_secrets/transformers/config.py

@@ -42,6 +42,7 @@ def parse_file(self, file: IO) -> List[str]:
 def _parse_file(file: IO, add_header: bool = False) -> List[str]:
     """
     :raises: configparser.Error
+    :raises: UnicodeDecodeError
     """
     lines: List[str] = []
     for key, value, line_number in IniFileParser(file, add_header=add_header):

tests/core/scan_test.py

@@ -13,13 +13,23 @@ def test_handles_broken_yaml_gracefully():
             textwrap.dedent("""
             metadata:
                 name: {{ .values.name }}
-        """)[1:].encode(),
+            """)[1:].encode(),
         )
         f.seek(0)
 
         assert not list(scan.scan_file(f.name))
 
 
+def test_handles_binary_files_gracefully():
+    # NOTE: This suffix needs to be something that isn't in the known file types, as determined
+    # by `detect_secrets.util.filetype.determine_file_type`.
+    with tempfile.NamedTemporaryFile(suffix='.woff2') as f:
+        f.write(b'\x86')
+        f.seek(0)
+
+        assert not list(scan.scan_file(f.name))
+
+
 @pytest.fixture(autouse=True)
 def configure_plugins():
     with transient_settings({