# Help! Very baffling code behavior, simple loop breaks pattern

**URL:** <https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581>\
**Category:** Troubleshooting\
**Created:** [November 9, 2025, 8:40pm UTC](https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581 "2025-11-09T20:40:07Z")\
**Posts on this page:** 7\
**Page:** 1

<div class="post-metadata">

**Author:** ![cjn566](https://forum.electromage.com/user_avatar/forum.electromage.com/cjn566/32/3029_2.png) [@cjn566](https://forum.electromage.com/u/cjn566)\
**Post date:** [November 9, 2025, 8:40pm UTC](https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581/1 "2025-11-09T20:40:07Z")

</div>

tl;dr: any simple loop breaks my pattern. The loop doesn’t do anything, is super short, and uses unique variables. Both a for loop and a while loop do this. By all reason they should have no effect whatsoever, but it breaks a specific seemingly unrelated part of the pattern.

Long version: I’m working on this pattern, it’s a modified copy of another pattern I’ve made that works fine. 1D pixel space. It works like this: “spots” of light spawn with random traits - location, hue, speed, size, they float around, despawn, repeat. They new version of the code introduces “gravity”, so the spots are attracted to eachother. This means each spot needs to process it’s relation to every other spot, hence the nested loop. But every time I add even the most benign inner loop, then new spots spawn with a size of 0.

Here’s a video showing the problem: [https://drive.google.com/file/d/1BpZOx\_3hQQbqXdIMshqzykYVvnHAX\_vF/view?pli=1](https://drive.google.com/file/d/1BpZOx_3hQQbqXdIMshqzykYVvnHAX_vF/view?pli=1)

And here’s the full code:

```javascript

var decay = 0.98
var G = 0.00001

var numSpots = 2
var fadeTime = 1200

export var position = array(numSpots)
export var size = array(numSpots)
export var speed = array(numSpots)
export var hue = array(numSpots)
export var deathTimeRemaining = array(numSpots)
export var fadeIn = array(numSpots)

baseTime = 2
offsetTime = 0.15

var tVals = array(12)
tVals[0] = baseTime + random(offsetTime)
tVals[1] = baseTime + random(offsetTime)
tVals[2] = baseTime + random(offsetTime)
tVals[3] = baseTime + random(offsetTime)
tVals[4] = baseTime + random(offsetTime)
tVals[5] = baseTime + random(offsetTime)
tVals[6] = baseTime + random(offsetTime)
tVals[7] = baseTime + random(offsetTime)
tVals[8] = baseTime + random(offsetTime)
tVals[9] = baseTime + random(offsetTime)

var RMAX = [1, 0.15, 0.5, 1, 200];
var RMIN = [0, 0.03, -0.5, 0, 100];

function initializeSpot(spot) {
  for (var k = 0; k < 5; k++) {
    var Tw = time(tVals[k]);
    var Tc = time(tVals[2*k]);

    var range = RMAX[k] - RMIN[k];
    var w = wave(Tw) * range;
    var c = wave(Tc) * (range - w) + RMIN[k];
    var v = c + random(w);

    if (k === 0) position[spot] = v
    else if (k === 1) size[spot] = v // <-- THIS IS ZERO WHEN THERE IS A LOOP
    else if (k === 2) speed[spot] = v
    else if (k === 3) hue[spot] = v
    else if (k === 4) deathTimeRemaining[spot] = v;
  }

  if (size[spot] <= 1/256) size[spot] = 0.02;

  fadeIn[spot] = 0;
}

// Initialize all spots
for (var i = 0; i < numSpots; i++) {
	initializeSpot(i);
}

export function beforeRender(delta) {

  for (var i = 0; i < numSpots; i++) {
    if (size[i] <= 1/256) {
      // recycle dead spots
      deathTimeRemaining[i] -= delta;
      if (deathTimeRemaining[i] <= 0) initializeSpot(i); 
      continue;
    }
    var xi = position[i];
    var mi = size[i];

    // for (y = 0; y < 3; y++) {} <-- HERE IS THE PROBLEM
    
    // z = 0 <-- OR HERE
    // while(z < 5){
    // z++
    // }
    
    
    // var mj = size[j]
    // if (mj > 1/256){
  
    // // signed shortest separation x_j - x_i ∈ [-0.5, 0.5]
    // var d = position[j] - xi
    // if (d > 0.5) d -= 1
    // else if (d < -0.5) d += 1
  
    // if (d != 0 && d < 0.3 && d > -0.3){
  
      // // 1/r^2 with sign via s = sign(d); no sqrt needed
      // var r2 = d * d;
      // var invr2 = 1 / r2;
      // var sgn = d > 0 ? 1 : -1;
  
      // // a_i = G * sj * sgn / r^2; a_j = -G * si * sgn / r^2
      // var fac = delta * G * sgn * invr2; // common factor for this pair
      // speed[i] += fac * mj * delta; // += a_i * delta
      // speed[j] -= fac * mi * delta; // += a_j * delta
        
        // rv = speed[j] - speed[i];
        // if(rv != 0 && abs(rv) * delta < 0.5){
        // tHit = -d/rv
        // if(tHit > 0 && tHit <= delta){
        // // COLLISION
        // speed[i] = (mi * speed[i] + size[j] * speed[j]) / (mi + size[j])
            
        // mi += size[j]
        // size[j] = 0
        // }
        // }
    // }
    // }
    // }

    // size decay (your logic)
    mi *= decay
    size[i] = mi

    // integrate position; wrap without modulo
    var x = xi + (speed[i] * delta)/1000;
    if (x >= 1) x -= 1;
    else if (x < 0) x += 1;
    position[i] = x;

    if (fadeIn[i] < fadeTime) fadeIn[i] += delta;
  }
}

export function render3D(index, x, y, attr) {
	if(index >= 215) return
	
		var v = 0
		var h
		for (var i = 0; i < numSpots; i++) {
			if (size[i] > 1/256) {
				dist = abs(index/215 - position[i])
				if(dist > 1 - size[i]){
					dist = 1-dist
				}
				if(dist < size[i]){
					var scale = fadeIn[i] / fadeTime
					prox = ((size[i] - dist) / size[i]) * scale
					if(prox > v){
						v = prox
						h = hue[i]
					}
				}
			}
		}
		
		
		
	ring = false
	invert = false
	if(attr > 0){
		invert = true
		if(attr >= 0.9){
			ring = true
		}
	}
	
	hsv(index > 141? -h : h, 1, v)
}

```

Somebody please help me find some clue as to WTF is going on here. This installation is getting installed in 48 hours, so I’m freaking out a tiny bit.

---

<div class="post-metadata">

**Author:** ![zranger1](https://forum.electromage.com/user_avatar/forum.electromage.com/zranger1/32/301_2.png) [@zranger1](https://forum.electromage.com/u/zranger1)\
**Post date:** [November 10, 2025, 1:47am UTC](https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581/2 "2025-11-10T01:47:35Z")

</div>

There is something strange going on here that I haven’t managed to track down.

The (sort of) good news is, if you get rid of the `continue`, the pattern goes back to working as expected. The loops work, and you can add as much code as you want. Far as I can tell, the only effect on the pattern of leaving the `continue` out is that spots restart on the current frame, rather than on the next one.

Here’s what I saw while debugging:

If you put _any_ loop in the code after the

```auto
    if (size[i] <= 1/256) {
      // recycle dead spots
      deathTimeRemaining[i] -= delta;
      if (deathTimeRemaining[i] <= 0) initializeSpot(i); 
      continue;
    }

```

block, (where the commented loop blocks are) the pattern fails as described. Using exported debug variables to track the calculations in `initializeSpot()` shows that `v` is being calculated correctly, and it is apparently called properly if `k == 1` to set `size[spot]`. But if a second loop is present, `size[spot]` remains 0 outside of `initializeSpot()`.

It’s not execution time related. I added a huge block of random trig functions in place of the secondary loops, and everything worked fine.

Maybe there’s math I’m not seeing that’s stomping the `size` array, but I’m slightly worried that in this case, `continue` is somehow misbehaving. Would love for someone to find the totally obvious thing that I’m not seeing. Reserved word as variable name? Out-of-bounds write to another array, maybe?

---

<div class="post-metadata">

**Author:** ![cjn566](https://forum.electromage.com/user_avatar/forum.electromage.com/cjn566/32/3029_2.png) [@cjn566](https://forum.electromage.com/u/cjn566)\
**Post date:** [November 10, 2025, 3:16am UTC](https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581/3 "2025-11-10T03:16:20Z")

</div>

Thanks for catching that! I had a suspicion that `continue` was problematic and pruned it from the area of code I was working in, but I didn’t see that one there.

Still very odd, but now I know not to use continue and I can keep working on this.

---

<div class="post-metadata">

**Author:** ![wizard](https://forum.electromage.com/user_avatar/forum.electromage.com/wizard/32/4_2.png) [@wizard](https://forum.electromage.com/u/wizard)\
**Post date:** [November 10, 2025, 3:31am UTC](https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581/4 "2025-11-10T03:31:54Z")

</div>

Could be a bug with `continue` I’ll take a look and boil this code down to make a regression test. Thanks for posting it!

---

<div class="post-metadata">

**Author:** ![maumau](https://forum.electromage.com/user_avatar/forum.electromage.com/maumau/32/3384_2.png) [@maumau](https://forum.electromage.com/u/maumau)\
**Post date:** [November 13, 2025, 5:22pm UTC](https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581/5 "2025-11-13T17:22:47Z")

</div>

I don’t see an obvious issue with stepping on an array but have you tried using an ‘else’ instead of ‘continue’ to see if that fixes the issue?

```auto
    if (size[i] <= 1/256) {
      // recycle dead spots
      deathTimeRemaining[i] -= delta;
      if (deathTimeRemaining[i] <= 0) initializeSpot(i); 
    } else {
      var xi = position[i];
      var mi = size[i];
      ...

```

---

<div class="post-metadata">

**Author:** ![cjn566](https://forum.electromage.com/user_avatar/forum.electromage.com/cjn566/32/3029_2.png) [@cjn566](https://forum.electromage.com/u/cjn566)\
**Post date:** [November 13, 2025, 11:12pm UTC](https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581/6 "2025-11-13T23:12:20Z")

</div>

That’s what I ended up doing, yes. Once I was aware that `continue` was a problem, that was a pretty clear workaround.

---

<div class="post-metadata">

**Author:** ![system](https://forum.electromage.com/uploads/default/original/1X/98123d16088a5451b5a585e24c0f2a40dd76061e.png) [@system](https://forum.electromage.com/u/system)\
**Post date:** [March 13, 2026, 11:12pm UTC](https://forum.electromage.com/t/help-very-baffling-code-behavior-simple-loop-breaks-pattern/4581/7 "2026-03-13T23:12:31Z")

</div>

This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.