[erlang-questions] subnet masking with binary matching

[Kostis Sagonas]

Bob Ippolito wrote:
> 
> You could just do this:
> 
> mask_address(Addr, Maskbits) ->
>   B = list_to_binary(tuple_to_list(Addr)),
>   io:format("address as binary: ~p ~p~n", [B,Maskbits]),
>   Rest = (size(B) * 8) - Maskbits,
>   <<Subnet:Maskbits, _Host:Rest>> = B,
>   Subnet.
> 
> If/when the feature enhancements proposed in that paper end up in
> Erlang, then this code would be shorter and cleaner. It sure would
> make a huge difference in a lot of my code, but even the current
> syntax is much better than bit twiddling other languages.

I second the last sentence.

By the way, the feature enhancements proposed in our paper are already 
in the development version of Erlang/OTP and will most probably be 
released as part of R11B-3.

Jeff's code would need to be slightly modified and use the "bitstr" type 
specifier:

mask_address(Addr, Maskbits) ->
   B = list_to_binary(tuple_to_list(Addr)),
   io:format("address as binary: ~p ~p~n", [B,Maskbits]),
   <<Subnet:Maskbits, _Host/bitstr>> = B,
   Subnet.


Kostis