Re: [Swarm-Support] Segmentation Violation

[Alex Olivas]

Steve Railsback wrote:

> Alex Olivas wrote:
>
> > I modified bug slightly to "interact" with nearest neighbor bugs.
> > I would like for an individual bug's step to depend on the state
> > of the bugs around it.  I defined a simple method called getRole,
> >
> > - (int)getRole;
> > {
> >  return role;
> > }
> >
> > My code compiles just fine, but when I run it crashes in the
> > bug's step method at the line...
> >
> >  int neighborRole = [[world getObjectAtX: xPos+deltaX Y: yPos+deltaY] 
> > getRole];
>
>
> One thing to check very carefully is whether (xPos+deltaX) or 
> (yPos+deltaY) could ever exceed the size of world. If they do, I think 
> you'll get a segmentation fault.
>
> Steve R.
That actually turned out to be the problem.  I thought it odd that if 
the position were off the grid
[world getObjectAtX: xPos+deltaX Y: yPos+deltaY] would return a non-nil 
value.
The line just before the one above checks this...

if([world getObjectAtX: xPos+deltaX Y: yPos+deltaY]!= nil){
     //printf("x: %d y: %d\n",xPos+deltaX,yPos+deltaY);
     //xprint([world getObjectAtX: xPos+deltaX Y: yPos+deltaY]);
     int neighborRole = [[world getObjectAtX: xPos+deltaX Y: 
yPos+deltaY] getRole];

Thanks,
Alex.