UDP proxy bug in python version less than 3.7

[yuchting]

Here's a simple bug I met and I will give a reminder to resolve.

socket.type will be changed after settimeout/setblocking calling. And pysocks will not send proxy data after these 2 functions called, because socks.sockssocket.sendto just use socket.type to adjust how it will send message.

just look in python3.6.8:

Python 3.6.8 (default, Apr  2 2020, 13:34:55) 
[GCC 4.8.5 20150623 (Red Hat 4.8.5-39)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import socket
>>> udp_socket = socket.socket(socket.AF_INET, socket.SOCK_DGRAM) 
>>> udp_socket.type
<SocketKind.SOCK_DGRAM: 2>
>>> udp_socket.setblocking(1)
>>> udp_socket.type
<SocketKind.SOCK_DGRAM: 2>
>>> udp_socket.setblocking(0)
>>> udp_socket.type
2050       <----------------- look, changed
>>> 

in python3.8:

Python 3.8.3 (tags/v3.8.3:6f8c832, May 13 2020, 22:37:02) [MSC v.1924 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import socket
>>> udp_socket = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
>>> udp_socket.type
<SocketKind.SOCK_DGRAM: 2>
>>> udp_socket.setblocking(0)
>>> udp_socket.type                     
<SocketKind.SOCK_DGRAM: 2>   <----------------- look, won't be changed
>>>

in python official document, they saied:

• Changed in version 3.7: The method no longer applies SOCK_NONBLOCK flag on socket.type.

so the following code of pysocks will has a small bug:

    def sendto(self, bytes, *args, **kwargs):
        if self.type != socket.SOCK_DGRAM:  # <---------------- here
            return super(socksocket, self).sendto(bytes, *args, **kwargs)
        if not self._proxyconn:
            self.bind(("", 0))

of course all code like this equal judgement for check whether is UDP socket might have same issue in python version less then 3.7, I hope you will fix it in next version.

please don't mind, just reminder, thanks for your project to help me.

[yuchting]

sorry, this RP has resolved this issue:
#128

I will close this issue.