[qvwm-e:00310] Re: Forwarded Message From CP Spam Shield

James Newlin wrote:

>>If you've compiled qvwm with debugging symbols (which is the default,
>>i.e. with -g passed to the compiler) you should be able to get gdb to
>>tell you what function it's segfaulting in. That should make figuring
>>out what the problem is a whole lot easier. My guess is that the
>>segfault is in some X or Xt function that SCO has diddled with. The
>>commercial Unix vendors are notorious for diddling with X...
>>    
>>
>
>Thanks for the input.   When I run backtrace in gdb on qvwm and the core
>dump, it says its in _istart().     I get the same error outside of X.  Does
>this mean that the program isn't compiling correctly?    There were no
>errors reported when compiling.    Could I replace SCO's version of X?
>
>Thanks,
>
>James
>
>
>  
>

It looks like you have a big problem. The _istart function is most 
likely part your system's crt library (crt0.o or crt1.o, possibly in 
/lib or /usr/lib), which means your compiled version of qvwm is falling 
over early in the process of starting with X.

As for replacing SCO's X, I think that would be a monumental task. It 
could break every program that's linked to your current X libraries. To 
fix that you'd have to relink, or possibly recompile, all of those 
programs. It would be easier to find a solution inside the qvwm source 
code. I don't think qwm is at fault, my guess is that SCO's doing 
something that's out of compliance with the ICCCM (Inter Client 
Communication Conventions Manual).

See the "X-Windows Disaster" chapter of the "Unix Hater's Handbook" 
before proceeding:

http://www.art.net/~hopkins/Don/unix-haters/x-windows/disaster.html