Re: Proposal for adding new option to nmap

[Fyodor <fyodor () insecure org>]

On Mon, Jul 05, 2010 at 02:27:24PM -0600, David Fifield wrote:
> On Mon, Jul 05, 2010 at 03:37:57PM +0530, kirubakaran S wrote:
>
> This option is only really useful if it produces clean output, without
> any startup or junk messages. Ideally, it prints out only the selected
> scripts, one per line, to standard output, and quits. I understand that
> this might not be easy to implement because of the need to initialize
> the script engine and avoid code paths that produce extra output.

Maybe there should be some unique way to distinguish the --script-list
lines.  For example, they could all start with "SL: ".  My feeling is
that it is easy to parse those out, and provides a few significant
advantages:

o We don't have to remember in the future to make sure all output
  message either fall outside the Nmap --script-list code path or are
  properly guarded against printing anything when --script-list is
  requested.  It seems likely that someone will eventually add an
  output line somewhere which inadvertently breaks this.

o It means we don't have to go through all the existing Nmap output
  and add ugly guard messages to prevent it from printing in
  --script-list mode.

o It means that if we ever want to add a feature to --script-list
  (like include the supported script phases or categories), we can do so
  without breaking compatibility.  We would (for a time) keep printing
  the old stuff with the "SL: " prefix, and use a new prefix for the new
  format.  I think that is cleaner than having to add a different
  command-line option.

> If --script-list can't easily work in this way, then we might as well just
> use -d2.

I'm not against the -d2 approach, but you do lose the "format
guarantee" aspect.  We would have to remember not to change that
debugging output, and obtaining the full path name is more of a pain
with -d2 format since its on a separate Fetchfile line.

> Either way, we will have to document in the source code that the
> specific output format is being used by an external program (Zenmap) and
> it can't be changed or removed.

Yes.

> I am thinking that the script list output should have the full absolute
> path to every script, not just the basename.

Makes sense.

> The output might be
>
> /usr/share/nmap/scripts/http-auth.nse
> /usr/share/nmap/scripts/http-date.nse
> ...
> /home/david/custom/test-script.nse

Note that on Windows it may include spaces like:

C:\Program Files (x86)\Nmap\scripts\http-auth.nse

Cheers,
Fyodor
_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://seclists.org/nmap-dev/