🚀 go-pugleaf

Hi, there,

   I am developing GUI seismic modelling package using XVT Portability
Toolkit. A design program comes with toolkit can design interface and
generate makefile for me. The program I wrote is compiled fine with
SUN cc and gcc. I do not have any problem at all when I run it on SPARC.

   However, when I run the program on MS-Window 3.1 (compiled with Borland C
4.0), I had a general protection problem. Then I tried to track where the
problem was and I found two places I really do not undersand (Sorry for not
including the entire program  since it onsists of more than twenty files an 
the size is over 1 Mb):

A: 

           max_bed_point *= 5;
           nblk = (no_bed+10)*(max_bed_point+2);

I added two printf statements as
           printf("NO_BED = %d and MAX_BED_POINT = %d\n",no_bed,max_bed_point);
           printf("NBLK = %d\n",nblk);

I got NO_BED = 32 and MAX_BED_POINT = 230 (they were correct) but I got
      NBLK = 0 which should be 9744. Sun CC and GCC gave me 9744 here.


B:  The code here is part of my input module, *no_unit here is actually
no_bed shown above

      printf("READ_MODEL A\n");
      for(i=0; i<(*no_unit);i++)
         {no_point_unit[i] = 0; no_point_work[i]=0; printf("NO_UNIT=%d and I =
%d\n",*no_unit,i);}
      printf("READ_MODEL B\n");

   Guess what happened, the loop went on and on until a general protection
problem was caused, in this case i reached 447. However, *no_unit printed
out was 32. Note, only 32*4 bytes allocated for no_point_work and (32+11)*4
bytes allocated for no_point_unit.

   Beside I did something very silly, I took out the two assignation statements
and then the loop went on and on with no stopping. I had to reboot the machine.


   After a while, I thought there might be a virus in my PC, I used f-prot
to scan the HD, YES, I found one called antiexe. I was so glad to find it.
I used a clean boot disk to reboot the system, the virus had gone. I delete
all object files and recompiled the program again. I still had the same
problem.

   Then I write a small program with the same statements and assign them
the same values as above, compiled and run it, OK, no problem.

  Now I really do not understand what the problem was and come here
for your help

  Could any one kindly advise me what mistakes I could make to cause this
kind of problem ? Sound like a memory overlap to me, but why the program run OK
on SPARC compiled with either SUN cc or gcc. Is the program too large for
BORLAND C to handle ?, Is it because the PC had too less memory (only 4 MB) ?.
Does BORLAND C 4.0 have any bugs ?, Or possibly some virus which f-prot
could not track ?  PLEASE PLEASE advise and cc jim@gl.rhbnc.ac.uk a copy
if reply to the newgroups.


  Thank in advance


Jim Hu