[Xastir] Xastir / Solaris 10 Sparc

[Chris Bell]

> > 
> > Hello Dick and thank's for your answer.
> > Yes...solved with this...but then we don't TX our position.
> > 
> > Anybody more with this problem?
> 
> It is pretty similar to what happens on HPUX, actually. I haven't
> looked at that in quite a while - I started trying to debug but
> got lost in the threading stuff, and haven't gotten back to it. 
> 

I started looking into this... [after working around another X issue
that has back-burnered my solaris work for a while] On my system
giving snprintf() a string pointer that is NULL makes it segfault.
yummy.

I am not sure who is wrong... that we are letting an arg be NULL, or
solaris segfaulting (and Linux's glibc not...) I am leaning toward
both, with Linux glibc being the only sane tolerant behaviour.  

I tried using the snprintf in src/snprintf.c, but the system fprintf()
dies the same way (in debug statements).  

So the fix is to use "" instead of NULL everywere a string might be
used in a *printf() function.  The big culprit right now is the
path/unproto_path variables.  I have not looked in too deep to see if
there are others that could be NULL as well...
interface.c:output_my_aprs_data() is where the normal runtime crashes
on net tx, and interface.c:output_my_data() crashes with debug set to
1.

I am going to see how many places it needs to be changed, and test
that while waiting to see if anyone else has any bright ideas... 

I actually have some HPUX boxes now, I will have to see if any have a
dev kit installed to test porting there too!  

Chris.

KD6ZWR
cbell at junknet.com