On Wed, Jan 03, 2001 at 11:12:53PM -1000, Loukas wrote:
> Hey guys,
> great work with libpcap and tcpdump.
> I've been trying to get libpcap to install on Darwin(MacOSX). 
> I managed to get it to install (libpcap-0.5.2 install log), witha  
> couple of small changes. tcpdump comes pre-installed on 
> Darwin and i also installed snort (snort-1.6.3 install log).
> Now the problem is, i can sniff on en0 but not on ppp0 
> (atest log). I thought you might find this interesting and 
> could possibly add macosx support.

These are mostly problem reports, not patches, so they should go to
"tcpdump-workers", not "patches".

>From the configure log:

        checking for gcc... no
        checking for cc... cc
        checking whether the C compiler (cc  ) works... yes
        checking whether the C compiler (cc  ) is a cross-compiler... no
        checking whether we are using GNU C... yes
        checking whether cc accepts -g... yes
        checking gcc version... 
        ./configure: parse error: condition expected: -gt [955]

That one is very odd; there's a clause in that code that does

        if test $ac_cv_lbl_gcc_vers -gt 1 ; then
                V_CCOPT="-O2"
        fi

but unless, somehow, "$ac_cv_lbl_gcc_vers" was empty, that shouldn't
have had a problem.

Now, if somehow Apple insists on calling GCC "cc" but not "gcc", it
could be that the code got confused somehow.

What does "cc -v" print?

In the log of snort and tcpdump:

        [lister:Library/Desktop/snort-1.6.3] root# tcpdump -v -i en0
        tcpdump: listening on en0
        <blah>
        [lister:Library/Desktop/snort-1.6.3] root# tcpdump -v -i ppp0
        tcpdump: ppp0: Device not configured

Well, *is* device ppp0 configured?  What does "ifconfig -a" report (I'm
assuming - hoping - here that MacOS X has a modern "ifconfig" that
supports "-a")?

If it's configured, you might want to ask Apple why you can't attach a
BPF device to it, as that's probably the fault of MacOS X, not of
libpcap.

The changes mentioned were

        cc -O -I.  -DHAVE_SYS_IOCCOM_H=1 -DHAVE_SYS_SOCKIO_H=1 -DHAVE_ETHER_HOSTTON=1 
-DHAVE_STRERROR=1 -DHAVE_SOCKADDR_SA_LEN=1  -c ./pcap-bpf.c
        ./gencode.h:182: bad attribute specification, expecting identifier, found 
`volatile'
        ./gencode.h:182: illegal function definition, found `)'
        make: *** [pcap-bpf.o] Error 1

        i got rid of this error by commenting out the following in gencode.h:
        /*#if __STDC__
        __dead void bpf_error(const char *, ...)
            __attribute__((volatile, format (printf, 1, 2)));
        #endif*/

        after that make worked fine:

and

        to get 'make install' to work, i had to create a user 'bin'
        and mkdir /usr/local/lib

For the first one, the 0.6 code doesn't do the "__dead void" stuff, or
the "__attribute__" stuff, although it's a bit surprising that a
compiler that the configure script thought was GCC didn't accept them.

For the second one:

        Makefile.in is wired, in 0.6, to use "bin" as the owner; we
        may want to go back to making that platform-dependent.

        Makefile.in doesn't create "$(DESTDIR)/$(LIBDEST) before
        installing the libraries in it; I'll check in code to do so into
        the main and x.6 branches.

> ps. I also tried recompiling tcpdump with the new libpcap 
> installed, but i had no success. i have a log of that as well if 
> youare interested

Yes, we'd want a log of that.




-
This is the TCPDUMP workers list. It is archived at
http://www.tcpdump.org/lists/workers/index.html
To unsubscribe use mailto:[EMAIL PROTECTED]?body=unsubscribe

Reply via email to