wallyri@hotmail.com

The following simple patch highlights a weird problem. If you run 'top' at the same time as you run this patch (turn on the metro), you'll see the RSIZE, RPRVT, and %CPU start to increase steadily. There seems to be some sort of memory leak. I can't locate which lines of the javscript cause the problem. Does anyone know what's causing this?

tmp += "target " + tgt + ", " + str + " " + val + " " + time;

The following simple patch highlights a weird problem.  If you run 'top' at the same time as you run this patch (turn on the metro), you'll see the RSIZE, RPRVT, and %CPU start to increase steadily.  There seems to be some sort of memory leak.  I can't locate which lines of the javscript cause the problem.  Does anyone know what's causing this?

function mess(tgt, str, val, time) {
    var time2 = time | 0;
    var tmp = "";
    tmp += "target " + tgt + ", " + str + " " + val + " " + time;
    //tmp += str + " " + val + " " + time;                                                                               
    //outlet(0, "target "+tgt);                                                                                          
    outlet(0,tmp);
  }

var val = 0;
function randrange(range){
  val = Math.random()
  return val* range;
}

function randvoice(tgt,cf){
  var freq = 220+randrange(60);
  var amp = -40+randrange(20);
  var cf2 = cf+randrange(100);
  var bw = 150+randrange(75);
  var pan = 0.5+randrange(0.5);

  mess(tgt,"freq", freq,500);
  mess(tgt,"amp",amp,500);
  mess(tgt,"cf",cf2,500);
  mess(tgt,"bw",bw,500);
  mess(tgt,"pan",pan,500);
}

function bang(){
  for(var i = 0; i < 3; i++){
    var cf = i*300;
    for(var j = 0; j < 6; j++){
      randvoice(i,cf );
    }
  }
}

javascript-memory-usage

On May 11, 2007, at 9:44 PM, isaac wrote:

> The following simple patch highlights a weird problem. If you run 

> 'top' at the same time as you run this patch (turn on the metro), 

> you'll see the RSIZE, RPRVT, and %CPU start to increase steadily. 

> There seems to be some sort of memory leak. I can't locate which 

> lines of the javscript cause the problem. Does anyone know what's 

Looks like you're bloating the symbol table inside mess(). Do you 

need to use strings/symbols? Why not an atom list? Commas and args 

don't take the same effect as message boxes as JS strings anyway. 

Common confusion, but you typically don't want to use strings for 

composite messages anyway, but rather a list of items. See my message 

to Barry earlier today for one of many examples.

> The following simple patch highlights a weird problem.  If you run  
> 'top' at the same time as you run this patch (turn on the metro),  
> you'll see the RSIZE, RPRVT, and %CPU start to increase steadily.   
> There seems to be some sort of memory leak.  I can't locate which  
> lines of the javscript cause the problem.  Does anyone know what's  
> causing this?

Looks like you're bloating the symbol table inside mess(). Do you  
need to use strings/symbols? Why not an atom list? Commas and args  
don't take the same effect as message boxes as JS strings anyway.  
Common confusion, but you typically don't want to use strings for  
composite messages anyway, but rather a list of items. See my message  
to Barry earlier today for one of many examples.

Thanks, it looks as though that fixes it. Makes programming easier too, since I don't need a 'fromsymbol.' Interestingly, RZISE and RSHRD still go up, but it looks like they level off after enough time, so its probably nothing.

Thanks, it looks as though that fixes it.  Makes programming easier too, since I don't need a 'fromsymbol.'  Interestingly, RZISE and RSHRD still go up, but it looks like they level off after enough time, so its probably nothing.  

On May 12, 2007, at 10:51 AM, isaac wrote:

> Interestingly, RZISE and RSHRD still go up, but it looks like they 

> level off after enough time, so its probably nothing.

Yes, the JS memory heap can grow to 16-32MB or something like that 

until the JS VM starts garbage collecting, which would explain what 

> Interestingly, RZISE and RSHRD still go up, but it looks like they  
> level off after enough time, so its probably nothing.

Yes, the JS memory heap can grow to 16-32MB or something like that  
until the JS VM starts garbage collecting, which would explain what  
you're seeing there.

javascript memory usage?