BJ> I know I've seen it reported that the MEX virtual 
 BJ> machine was broken.  I think I've now updated to 
 BJ> the version of Maximus directly out of the source 
 BJ> forge site (plus the files in max-3.03-etc.tar.gz), 
 BJ> but I'm not sure....
 ...
 WG> I'm looking at the MEX VM on-and-off -- it's been frustrating me.

:(

 WG> The only changes which ave made it into the CVS 
 WG> versions of max is a modification of the grammar so 
 WG> that it will compile correctly under real yacc (as 
 WG> opposed to GNU bison). 

Ok.  I assume it will still work with GNU bison also....  
Must, since I recompiled from the CVS tree....

 WG> The problem is that for some reason, MEX is 
 WG> dereferencing pointers to unallocated memory when its 
 WG> trying to pull stuff from its heap. I originall through 
 WG> the problem was that Scott was coercing pointers into 
 WG> 32 bit scalars and back, but I have since tested under 
 WG> a 32-bit platform (sparc) and it still blows up.

I can attest that MEX is blowing up on my intel based linux system.  :( 
Guess I'm going to have to figure out how to use the debugger tools under
my Linux setup.  I liked the GUI interface NeXTStep / OpenStep had with the
GNU compiler and debugger.....  That was very nice.....  Back to using
regular (command line driven) GDB if I don't spot something for the KDE or
GNOME desktop environments.....

 WG> Bascially, it's understand-the-code-then-hit-the-debugger time.

Ok......

Let me know where you can use a second person to bounce ideas off of.....

Take care.....

Bob Jones, 1:343/41


--- Maximus/2 3.01