Just for the record, I use command-line gdb. It has the advantage of
being very fast to use once you've learnt the command names (and I
only use about 5 or 6). You can use it to debug running processes
(e.g. pipeline monoliths etc)
% gdb $KAPPA_DIR/kapppa_mon 19234
where the number at the end is the process id of the running command.
When using valgrind, I tell valgrind to give me the option of running
up gdb each time valgrind detects an error, in order to investigate
what is going on. I have "valgrind" aliased to
valgrind --db-attach=yes --db-command="gdb %f %p" --track-origins=yes
David
|