Open UA_PARSER_YAML as binary

markdepalma · web-flow · commit ced19d4611e4 · 2022-04-26T21:14:20.000+02:00
The yaml files should be parsed as UTF8. In text mode, `open` uses `locale.getpreferredencoding`, which unless Python is run in UTF-8 mode (see PEP 540) might be complete nonsense (it's the encoding in `LC_CTYPE` on POSIX systems, and almost certainly a garbage ANSI codepage on windows, likely cp1252 for western windows systems). Open the file in binary mode and pass the binary stream to pyyaml for it to decode to avoid this error. Closes #98
diff --git a/ua_parser/user_agent_parser.py b/ua_parser/user_agent_parser.py
@@ -477,7 +477,7 @@ def GetFilters(
     except ImportError:
         from yaml import SafeLoader
 
-    with open(UA_PARSER_YAML) as fp:
+    with open(UA_PARSER_YAML, 'rb') as fp:
         regexes = yaml.load(fp, Loader=SafeLoader)
 
     USER_AGENT_PARSERS = []
