Re: [ethereal-dev] AIX: gtk problem solved, now an ethereal problem

[Laurent Deniel <deniel@xxxxxxxxxxx>]

Guy Harris wrote:

> It turns out that the only code that uses it is "pcap_lookupdev()",
> a/k/a "pick some interface and give me its name"; "libpcap" itself
> doesn't use it, it just exports it, "tcpdump" uses it only if you don't
> explicitly tell it what device to use, and Ethereal doesn't use it at
> all, so if it's broken, the only effect would be that you'd have to give
> "tcpdump" a "-i" flag - it'd probably complain if you didn't.

ok
 
> Any idea what the right test should be to see whether to use SIOCGIFCONF
> or OSIOCGIFCONF?  Is there a built-in ifdef we should use (e.g., AIX),
> or do we need to put in, say, a configure-script test?

The only OS i see that uses OSIOCGIFCONF was AIX and it was mandatory in 
order to make it work like the other unix with the `normal' struct ifconf.
So an #ifdef OSIOCGIFCONF should be ok (since it should be declared in the
same include file as SIOCGIFCONF). An #ifdef __aix is harder to handle 
since it depends on the compiler. Otherwise a configure option like 
HAVE_AIX or HAVE_OSIOCGIFCONF is ok also.

Laurent.

--
Laurent DENIEL        | E-mail: deniel@xxxxxxxxxxx
Paris, FRANCE         |         laurent.deniel@xxxxxxxxxxxxxxxxxxxxxxxxxxxx
                      | WWW   : http://www.worldnet.fr/~deniel
    All above opinions are personal, unless stated otherwise.