A segfault occurs in QtTermTCP on both 32bit and 64bit x86 systems while
attempting a K2K session between two instances of QtTermTCP without any
node software in between, with QTT configured for the default split
screen mode "Single Window (Term + Mon)", and using the AGW interface
with QtSoundModem. The segfault does not occur if QTT is configured for
Single Window (Term only), MDI, or Tabbed. It also does not occur if
using KISS rather than AGW.
I attempted to describe this back in January at /g/bpq32/message/40577 and followed up with a patch
against QtTermTCP in February at /g/bpq32/message/40647
that solved the issue here. In this specific instance it was while using
QtTermTCP 0.0.0.73. The segfault continues to occur with the latest QTT
0.0.0.77 Beta 1.
I read where others were using QTT for direct K2K sessions so assume I
am not using QTT for something it was not intended for... if I am then
please ignore this post!
The segfault is easy to reproduce using two systems which for this
example I am calling system A and system B. Start QTS on both A and B
with AGW enabled on a port of your choosing. I used the default port
8000 for testing.
Start a fresh copy of QTT on both system A and B letting QTT create a
new .ini file. On both A and B systems enable QTT Listen on the default
port 8015. Again on both A and B within the QTT AGW Setup enable both
the interface and the monitor and set Ports to 1. Then set a unique
Terminal Call for each and make sure the port in the AGW tnc setup
sections match whatever port was specified for AGW in QTS. Save the QTT
config then stop and restart QTT.
An AGW connect from either A to B or from B to A works. To reproduce the
segfault simply attempt sending something TO the system that issued the
connect FROM the system that was connected to. A gdb backtrace should
look very similar to the one I provided at /g/bpq32/message/40577 .
I reworked the original QTT .73 patch and attached it for use with QTT
.77 Beta 1 source. Rework or ignore as seen fit.