Re: DNS resolver for NSE

[doug () hcsw org]

On Tue, Jun 26, 2007 at 11:55:54PM +0200 or thereabouts, Diman Todorov wrote:
> As far as I remember your DNS resolver does use nsock. Providing  
> hooks between your resolver and NSE does pose a challenge but I don't  
> see how it would block the whole Nmap process. After all the whole  
> point of using nsock is to have parallelized network I/O.

Good point. nmap_dns.cc does use nsock and could have some of its
functionality available to NSE scripts if it was refactored a little.
The biggest problems are that it wants its own nsock loop and that it
only does reverse DNS.

> oh contrair. The garbage collection in Lua is quite flexible. In fact  
> stoiko extended it to collect and close open but unneeded nsockets. I  
> am not familiar with weak hash tables but are they something like  
> this http://lua-users.org/wiki/WeakTablesTutorial ?

I'm sorry, I was wrong. After reading those wiki links you pasted I
realise that Lua offers more gc control than I thought. Oh well,
as the french say, ploo suh change. :)

> And what exactly  
> do you mean by a 'real format function' anyway?

I mean a format function that can be macro compiled to CODE instead
of just a function call to an interpreter (ala printf and lua's
format). Also, one that offered even 1/10th of the features of CL's
format could be considered "real".

> Speaking of Lua macros, I thought about adding at least a simple  
> macro system to NSE but I decided against it.

As much as I love macros, I must say I am AGAINST adding macros to
a language with signifigant syntax like Lua. Macros only work once
you've ELIMINATED syntax, and once you've eliminated syntax you have
lisp. :)

Best,

Doug

Attachment: signature.asc
Description: Digital signature

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