Re: ncat/sys_wrap.c

[Fyodor <fyodor () insecure org>]

On Tue, Jan 06, 2009 at 01:20:59PM -0600, Kris Katterjohn wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 01/06/2009 08:19 AM, Gisle Vanem wrote:
> > Why do we need to override vsscanf() like in
> > sys_wrap.c? The _asm statements in this file is
> > only good for MSVC. What about MIngW?
>
> I'm not sure why it's needed, but I tracked down its origins to Mixter's r8439
> with the log message "Fix win32 portability issues".  I'm not sure if he was
> testing with VC++2008 during this commit, but judging by the other SVN logs he
> may have been trying to.  I'll try to look into whether this is needed at all.

Thanks Kris.  With your Ncat VC++ successes last summer, this may no
longer be needed at all.  It looks like the "vsscanf" with the
assembly code is only used by "Sscanf" in the same file.  But we
should probably remove that and the rest of sys_wrap.* and use the
same portability methods used in the rest of Nmap.  Visual C++ seems
to have the normal sscanf() anyway since we seem to already be using
it in other parts of Nmap and I don't see any nbase redefinition.

If you have time to try to remove the sys_wrap* stuff, that would be
great.

Cheers,
-F

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