Re: heatbugs cold

[Ginger Booth]

Terry, Nelson,

    Thanks so much for the suggestions!  That's what I was hoping for, some
more ideas and a little narrowing down of what's wrong.

    Terry, that 32-char suggestion really deserves to work.  I suspect it 
won't (the tcl I've seen has short stuff), but it deserves to. :)  Will sift
carefully anon.  
 
> Damn, I'm sorry to hear this. It doesn't sound that bad to me, though,
> mostly BLT problems.

    Fwiw, all other BLT tests appear to work.  Also fwiw, I didn't mean to
suggest the situation was dire, just that "I'd gotten nowhere" since last 
time I yelped, on the problems I'd had last time I yelped.

> >    I tried Nelson's suggestion and built testGraph.  It compiles and dies:
> >        log10: SING error (yes, SING, not sign)
> >        Floating exception (core dumped)

> Oh boy. Could you get a stack trace so we can see who is generating
> the error? Run under gdb and type "where" when it crashes.

I decided "gdb" wasn't even worth building, because, unfortunately, gcc 
refuses to do the "-g" thing under HPUX.  I vaguely recall some technical
reason why in the documentation.  I let it crash under xdb, the HP debugger.
The assembly code meant nothing to me.  You grok that junk?  At the time,
I decided the blt demos/graph and heatbugs test were closer to mostly-working,
thus easier to debug than going into testGraph and doing print-trace.

The "it isn't -that-" suggestions are the most helpful of all--thanks!
Especially readline, since there didn't seem anything I could do about it.
I was expecting heatbugs to eventually look like the screen shot in the
Swarm pages, and it doesn't look that way as long as I've run it.  (2e+29 
generations ought to be long enough :).  (And five minutes.  Actually, the
green dots appear to be running away from the red dots, which chase after.)

Well, off to play.

Thanks again,
    Ginger