Irritating bytearray behavior

[MRAB]

On 2019-09-17 02:37, Ian Pilcher wrote:
> I am using a bytearray to construct a very simple message, that will be
> sent across the network.  The message should always be 20 bytes:
> 
>     2 bytes - address family (AF_INET or AF_INET6) - network byte order
>     2 bytes - (padding)
>     4 or 16 bytes - IP address
> 
> The size of the IP address is dependent on whether it is an IPv4 address
> (4 bytes) or an IPv6 address (16 bytes).  In the IPv4 case, it should be
> followed by 12 bytes of padding, to keep the message size consistent.
> 
> Naïvely, I thought that I could do this:
> 
>      ip = ipaddress.ip_address(unicode(addr))
>      msg = bytearray(20)
>      msg[1] = socket.AF_INET if ip.version == 4 else socket.AF_INET6
>      msg[4:] = ip.packed
>      sock.sendto(msg, dest)
> 
> This doesn't work in the IPv4 case, because the bytearray gets truncated
> to only 8 bytes (4 bytes plus the size of ip.packed).
> 
> Is there a way to avoid this behavior copy the contents of ip.packed
> into the bytearray without changing its size?
> 
The truncation you're seeing is perfectly normal. It's the same as for 
lists.

Try this:

     msg[4 : 4 + len(ip.packed)] = ip.packed

Alternatively, you could build the message a bytestrings and then pad 
with the .ljust method.