Date: Thu, 21 Dec 2000 15:32:34 -0500 From: Wesley Eddy <email@example.com> Subject: Aha! I was right Message-ID: <20001221153233.A13059@masaka.cs.ohiou.edu>
On Thu, Dec 21, 2000 at 02:07:13PM -0500, Wesley Eddy wrote:
> We've had a couple reports of this compilation problem on Redhat systems now.
> Both of the lines which gcc is rejecting are assignments of what should be
> timevals to timevals. I'm betting that this is yet another problem caused by
> there being seven hundred libpcaps out there, and the one that Redhat includes
> is probably broken in that the ts element of a struct pcap_pkthdr isn't a
> timeval. Not having a Redhat system to look at the pcap.h though, I don't know.
> One solution may be to try another libpcap. I know the Linux-Mandrake one
> works fine, so going to one of their mirrors and getting a new libpcap RPM may
> be all it takes to fix.
I checked up some more on this by looking at an actual Redhat box of a friend
of mine, and I was right, the pcap.h they ship has ts as a bpf_timeval instead
of a plain old timeval. I'm not sure if you can just do a cast in the tcptrace
code to fix it, but in any case, getting a different libpcap should fix it.
This archive was generated by hypermail 2b30 : 01/09/01 EST