Re: Heatbug.java question

[David Smith]

On 18 Jan 2000, Marcus G. Daniels wrote:

> >>>>> "DS" == David Smith <address@hidden> writes:
> 
> DS> OK, here are the diffs:
> 
> Ok, the problem is that you've dropped the modulus on bestCell, so
> the coordinates go outside of the Discrete2d boundaries and memory
> dereferences end up being invalid, and a segfault occurs. 
>   

It seems to me that this uncovers a bug in HeatBug.java source. Here is a 
snip from the source in the Swarm CD:

    // After choice of ideal spot is made, there's a chance of
    // random move.  (Note the normalization of coordinates to [0,
    // worldSize). The current space library does not enforce
    // boundary conditions.)
    if ((Globals.env.uniformDblRand.getDoubleWithMin$withMax (0.0, 1.0)) 
        < randomMoveProbability)
      {
        // pick a random spot
        newX = 
          x + Globals.env.uniformIntRand.getIntegerWithMin$withMax (-1, 1); 
        
        newY = 
          y + Globals.env.uniformIntRand.getIntegerWithMin$withMax (-1, 1);
        
        // normalize coords
        newX = (newX + worldXSize) % worldXSize;      
        newY = (newY + worldYSize) % worldYSize;
      }

See where the closing brace is? With it there, the coords never get 
normalized unless they were picked up from the random move. I moved the 
"}" up above "// normalize coords" and no more crashes. 

David Smith

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.