llSetPos and Math Errors....

hi...has anyone else run across this problem: I'm attempting to create a non-physical swarm script for some animals I'm working on. it works reasonably well, but periodically seems to generate math errors. I can't figure out where i have values that are turning complex.

the error is somewhere in the following lines of code (within a sensor event):

            if( llVecMag(direc)>3)
            {
            llSetPos(<mypos.x+mult*INC*direc.x,mypos.y+mult*INC*direc.y,gz+zshift>);
            }
            if(llVecMag(direc)<=1)
            {
                llSetPos(<mypos.x+-mult*INC*direc.x,mypos.y+-mult*INC*direc.y,gz+zshift>);
            }

where direc is the target direction for the object in the swarm.

how, exactly, can llVecMag be producing complex values, unless the positional numbers going in are themselves complex.

I`d like to suggest that you ask people in the Artificial Life forum.

Toodle-oo!

Math errors are usually divide-by-zeros - can't quite see how this would be happening in the fragment you show. I would suggest loading the new position into a vector variable first, though, and passing this to llSetPos(), rather than doing the whole calculation within the function call argument list. This would simplify the logic for the LSL interpreter, and also allow you to print out the new vector position value, just to check it is what you expect.

Yeah, llVecMag shouldn't be giving math errors. Get some readouts of the values going in, see if you can track it down.