[ns-users] (even) More on Seg Fault

[rik@scs.leeds.ac.uk]

This is now getting quite strange but I hope it's not something I've
got wrong ;-)

The stability of my simulation seems to be (partially?) dependent on
the _ordering_ of the statements in my recv function (.tcl script). I
have had occasion when the ordering and binding of the instantiation of
the FTP/TCP agents has been the difference between the simulation
running and causing a seg fault.

Furthermore, we have two RH5.2 Linux machines. A given script may run
at both the prompt and in gdb on one machine, but only in gdb on the
other. I've even had cases where the script will run once in gdb, but
not a second time.

FYI, the trace indicates the segfault occurring in either the
aforementioned packet.h (line 217) or in TCPSink. This would show the
FTP source starting, but having a few problems :-(

thanks,
rik wade
ATM-MM Group
University of Leeds
UK