> I also think the change to remove mfprintf is not a
> good idea.  I checked the
> function on my computer, it works fine.  No crashes.
>  I don't see anything
> broken about it.  I suspect that maybe you are wrong
> about the problem with
> mfprintf.  At any rate, that band-aid will never fix
> the bug.  It's like the
> random changes you made for the closest planet bug,
> they had nothing to do
> with the real problem and didn't fix it.

Okay, I remember the condition which causes mfprintf

Manually start a robot in the normal port 2592,
request it to join a team that is not allowed. 

It will try to join that team 3 times, then fail.
Instead of an graceful exit or joining a different
team, it executes the mfprintf, and SIGSEVs. 

The robot defaults to joining Roms. 

So what I did was:

gdb robot

run -h localhost

during a Fed vs. Ori game. 

I can try it again, and see if it coredumps or not.