Re: Two NSE questions

[David Fifield <david () bamsoftware com>]

On Tue, Sep 29, 2009 at 03:33:34PM -0500, Ron wrote:
> On 09/29/2009 03:24 PM, David Fifield wrote:
> > Ron, if you propose an interface for setting the source address I'll
> > look into writing the binding. Just show me a snippet of your code as it
> > would work with the new function, and any other cases you think need to
> > be handled.
>
> I'm not particularly picky with how it should work, but the most logical  
> (to me) would be to change the connect() function from this:
>
> connect(hostid, port, protocol)
>
> To this:
>
> connect(hostid, port, protocol, lhostid, lport)
>
> I'm not sure if setting the local address makes sense, but setting the  
> local port in a connect call would be ideal. Another option would be:
>
> local socket = nmap.new_socket()
> set_local_port(socket, 67)
> connect(host, 68, 'udp')
> ...
>
> That's a little uglier, in my mind, but it is sufficiently usable.

Hmm, I like the second option because it's like the way to set a
timeout. I also think of a source address as a property of a socket, not
of a connection. But I think I would make it a method of a socket
object, like this:

local socket = nmap.new_socket()
socket:set_local_port(67)
connect(host, 68, 'udp')

David Fifield

_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://SecLists.Org