Re: [PATCH] Old massping doesn't compile

[David Fifield <david () bamsoftware com>]

On Sat, Sep 29, 2007 at 03:46:07PM -0500, Kris Katterjohn wrote:
> Hey everyone!
>
> As most of you know, David's great massping changes are in /nmap.  A
> #define in targets.cc allows the choice between David's changes and
> the old massping system.
>
> However, undefining "NEW_MASSPING" causes the compilation to halt
> because "icmpscan" isn't in "struct pingtech" ( targets.h)

I put that #define in way near the beginning of the migration. Then the
data structures changed around it, as you've seen.

> I've attached a patch that adds it back and compilation proceeds, but I didn't
> want to commit it because I thought that there might have been a good reason
> for removing it but it was just never finished so my patch isn't necessary, or
> there might be other problems (that I haven't encountered) and my patch only
> fixes some of it.

We decided that there was no difference between rawicmpscan and
icmpscan, so I merged them together. rawicmpscan used an Ethernet handle
and icmpscan used a raw socket. Now ultra_scan just uses whatever's
appropriate.

> Or it was just a mistake and I should've committed it, and sorry if that's the
> case :)

We haven't had a lot of problems reported with the new host discovery
code, so I'm going to nuke the old massping code. But it's probably not
bad to have a working massping, at least so it exists in the SVN
repository. I would rather see the change like this (which will require
some other small changes to make NEW_MASSPING visible to targets.h):

        struct pingtech {
          unsigned int rawicmpscan: 1,
#ifndef NEW_MASSPING
            icmpscan: 1,
#endif
            connecttcpscan: 1,
            rawtcpscan: 1,
            rawudpscan: 1;
        };

David Fifield

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