3.4.0 changed argument positions for constructor

Version: redis-py 3.4.0, redis 4.0.9

Platform: Python 3.6.6, Ubuntu 18.04

Description: Version 3.4.0 added an extra username argument to the redis.Redis constructor, but because it was added near the front of the list, and the arguments are not marked keyword-only, it breaks code that passes arguments to the constructor positionally.

Here’s a (somewhat contrived) example, that works on 3.3.11 but breaks on 3.4.0. This is the cause of jamesls/fakeredis#260.

#!/usr/bin/env python3
import redis

r = redis.Redis('localhost', 6379, 0, None, None, None, None, None, None, None, 'utf-8')
print(r.ping())

On 3.4.0 the ‘utf-8’ gets interpreted as unix_socket_path instead of encoding, leading to

Traceback (most recent call last):
  File "/home/bmerry/work/sdp/env3/lib/python3.6/site-packages/redis/connection.py", line 550, in connect
    sock = self._connect()
  File "/home/bmerry/work/sdp/env3/lib/python3.6/site-packages/redis/connection.py", line 892, in _connect
    sock.connect(self.path)
FileNotFoundError: [Errno 2] No such file or directory

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "./bad2.py", line 5, in <module>
    print(r.ping())
  File "/home/bmerry/work/sdp/env3/lib/python3.6/site-packages/redis/client.py", line 1357, in ping
    return self.execute_command('PING')
  File "/home/bmerry/work/sdp/env3/lib/python3.6/site-packages/redis/client.py", line 881, in execute_command
    conn = self.connection or pool.get_connection(command_name, **options)
  File "/home/bmerry/work/sdp/env3/lib/python3.6/site-packages/redis/connection.py", line 1178, in get_connection
    connection.connect()
  File "/home/bmerry/work/sdp/env3/lib/python3.6/site-packages/redis/connection.py", line 555, in connect
    raise ConnectionError(self._error_message(e))
redis.exceptions.ConnectionError: Error 2 connecting to unix socket: utf-8. No such file or directory.

[andymccurdy]

I moved the username arg to the end of the list in 5a1f3c4. 3.4.1 will be published shortly.

[bmerry]

Yes, it’s a fakeredis issue, and I’m hoping to get a fix out this week.