Problem binding sock for captive portal



  • I've got a WiPy serving html over http just fine on port 80 in AP mode.

    However, I'm also trying to set up a captive portal so that any dns request within the WiPy's broadcasted network is redirected to the configuration page.

    When I go to bind the socket to port 53, the execution just hangs forever. I simplified the code down to:

    import socket
    import network
    
    ap = network.WLAN()
    ap.init(network.WLAN.AP, ssid='dns test')
    
    def start():        
        udps = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
        udps.setblocking(False)
        udps.settimeout(10)#seconds
        print('just before bind')
        udps.bind(('',53))
        print('just after bind')
    

    when you call start() from the repl, you get:

    
    >>> start()
    just before bind
    
    

    ...the repl never returns

    The bind call on port 80 for my http setup works just fine and returns quickly.

    Any idea what I'm doing wrong here?


  • administrators

    Not that I can see from looking at the current documentation.



  • is it possible to control that behavior from micropython?


  • administrators

    My guess would be that the RTOS running underneath micropython is using the port to resolve DNS requests in AP mode... not sure if that was done recently to fix the reverse DNS lookup timeout telnet bug in MacOS or if this was working like that from the start.

    In STA mode your code works without problems...


Log in to reply
 

Pycom on Twitter