Re: HP-UX compiles - redux...

[Rick Jones <rick.jones2 () hp com>]

> > cc: "pcap-int.h", line 217: warning 573: Parameter list is 
> > inconsistent for "yylex".
>
>
> What is the argument list that "yylex()" takes?

Honestly, I have no clue.

> > here is the diff for pcap-dlpi.c:
>
>
> I've checked that in (with the code inside "#ifdef bogus"/"#endif" 
> removed rather than #ifdeffed out),

that's fine - I wasn't sure what was the desired disposition of the code there - 
 might have been that someone would want to pass an ebuf into that routine and 
take-out the snprintf in the caller.

> along with some changes that might
> get rid of the compiler warnings - are there still warnings with version 
> 1.106?

CVS?  I know notink about CVS :) I presume it will appear in "current" tomorrow 
right?

(Ok, Sgt Schultz doesn't really translate well in global email :)

> Does that code work when sending packets?  (I think I have a test 
> program for that at home that I can send you.)

since tcpdump didn't compile I didn't have occation to test libpcap.

> > now, there is an ip6_hdr in ip6.h.  However, the inclusion of ip6.h is 
> > conditional on INET6:
> >
> > #ifdef INET6
> > #include "ip6.h"
> > #endif
> >
> > but none of that IPv6 stuff where the error hits is conditional on 
> > INET6. Without a TOUR installed, HP-UX 11.11 knows little to nothing 
> > about IPv6.
> >
> > I'm not quite sure what to do to fix that code - if I should make it 
> > ifdef INET6 and return -1 as if it was an unknown IP version, or what.
>
>
> I need to see what, if anything, still needs to be done in my attempt to 
> get rid of the INET6 definition in tcpdump (so that the code is there 
> regardless of whether the OS itself knows about IPv6), and get that 
> checked in.
>
> That might fix the problem.

FWIW, while I was looking at ip6.h, it _appeared_ that it might ass-u-me an 
existing defintion for in6_addr

rick jones
-
This is the tcpdump-workers list.
Visit https://lists.sandelman.ca/ to unsubscribe.