[Xastir] Xastir is stack smashing?

[Mike Markowski]

On Fri 18-Jun-04 at  929 EDT, Mike Markowski wrote:
> 
> On Fri 18-Jun-04 at  240 EDT, Curt Mills wrote:
> > Stack smashing sounds like a bad thing, and it is.  It's possible
> > that on some systems Xastir just crashes, while on others it could
> > return meaningful error messages.  We'll have to keep an eye on
> > things.  If Xastir starts crashing more while doing certain things,
> > we might have introduced some new bugs into the code somewhere.
> 
> In this case, Curt, xastir doesn't even start.  The output in my 1st
> post on this is all that ever happens.  Over the weekend I'll see
> if I can use strace to pinpoint the problem and hack at it a bit.
> 
> I haven't changed kernels, though, which is the weird thing, because I
> agree with you, Chris, that it sounds like a kernel issue.  I'll
> keep you posted on any progress I make.

Ok, it's taken me a week longer than expected to get to this, but
I've made some progress that maybe will be more meaningful to the
developers than me.  I'll continue hacking at it, but am just
passing on this little discovery.

In the CVS main.c from this past week (not sure offhand which day...)
on line 7524 is the code:

	children[ac++] = da;

If I put a 'return' just before this line, no stack smashing error.
If I put the return just after the line, I get the error.

The funny thing is that this is on a P-III laptop.  On my Athlon 3200
at work, compiling with AX.25 and (I think) the same graphics library
support as here at home, xastir runs just fine using the same day's
cvs update.

I'll go dig into 'da' and see if I can figure out what's funny about
it on my P-III versus the Athlon at work.  For what it's worth, soundmodem,
gMFSK, xlog/hamlib, are all still working fine.  This is the only ham
or general purpose software on the machine that seems to be affected.

Mike "can't believe I'm hacking on a beautiful Sunday" Markowski
AB3AP