bug in loop protection macro expansion

[TakagiHitoshi]

p5.js version

1.11.11

What is your operating system?

Windows

Web browser and version

chrome 149.0.7827.104

Actual Behavior

"ReferenceError: _LP0 is not defined"

Expected Behavior

expected to run fine without errors.

Steps to reproduce

When run this code:
--- code begins --
//#つぶやきProcessing #p5js
t=0,h=.03
draw=_=>{
t||createCanvas(W=(w=200)2,W)+noStroke()+colorMode(HSB)
background(1,.1)
for(u=-4;u<4;u+=h)
for(i=0,x=u,y=u;i<t;i++)
[x,y]=[(x-y-x(S=xx+yy))h+x,(x-y+yS)h+y],
rect(X=xw+w,Y=yw+w,(fill(wcos(X-Y+t/17),w,w,.3),8))
t=++t%W}
--- code ends ---
Got this:
ReferenceError: LP0 is not defined
at /sketch.js:12:30
Instead, when run this code (with braces '{}')
--- code begins ---
//#つぶやきProcessing #p5js
t=0,h=.03
draw==>{
t||createCanvas(W=(w=200)2,W)+noStroke()+colorMode(HSB)
background(1,.1)
for(u=-4;u<4;u+=h)
{ // added
for(i=0,x=u,y=u;i<t;i++)
[x,y]=[(x-y-x(S=xx+yy))h+x,(x-y+yS)h+y],
rect(X=xw+w,Y=yw+w,(fill(wcos(X-Y+t/17),w,w,.3),8))
} // added
t=++t%W}
--- code ends ---
Runs fine.

[dsaicharan072-cmyk]

Hey @TakagiHitoshi I understood the problem here can I work on it to resolve it.

[TakagiHitoshi]

Thanks a lot for the quick response. That would be really helpful.

…

On Tue, Jun 16, 2026 at 4:10 PM dsaicharan072-cmyk ***@***.***> wrote: *dsaicharan072-cmyk* left a comment (processing/p5.js-web-editor#4183) <#4183 (comment)> Hey @TakagiHitoshi <https://github.com/TakagiHitoshi> I understood the problem here can I work on it to resolve it. — Reply to this email directly, view it on GitHub <#4183?email_source=notifications&email_token=ASK44J4QWNOVBWO4ZUC2LYD5ADXH3A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZRGU4TAMZYGAZ2M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4715903803>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ASK44JYZVPG3X3VH4FXGISL5ADXH3AVCNFSNUABEKJSXA33TNF2G64TZHM2TOOJUGA2DQNZ3JFZXG5LFHM2DMNZRGQYDOMRSGOQXMAQ> . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS <https://github.com/notifications/mobile/ios/ASK44JZMKWAMEYC3AI3NLYT5ADXH3A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZRGU4TAMZYGAZ2M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJKTGN5XXIZLSL5UW64Y> and Android <https://github.com/notifications/mobile/android/ASK44J3PTOHDVY33EWXZXDL5ADXH3A5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZRGU4TAMZYGAZ2M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLTGN5XXIZLSL5QW4ZDSN5UWI>. Download it today! You are receiving this because you were mentioned.Message ID: ***@***.***>

[inaridarkfox4231]

This is probably the correct code.

Works posted on Twitter (X)
https://x.com/TakagiHitoshi/status/2066649668520681725

upper code:

t=0,h=.03
draw=_=>{
t||createCanvas(W=(w=200)*2,W)+noStroke()+colorMode(HSB)
background(1,.1)
for(u=-4;u<4;u+=h)
for(i=0,x=u,y=u;i<t;i++)
[x,y]=[(x-y-x*(S=x*x+y*y))*h+x,(x-y+y*S)*h+y],
rect(X=x*w+w,Y=y*w+w,(fill(w*cos(X-Y+t/17),w,w,.3),8))
t=++t%W}

lower code (Parentheses added)

t=0,h=.03
draw=_=>{
t||createCanvas(W=(w=200)*2,W)+noStroke()+colorMode(HSB)
background(1,.1)
for(u=-4;u<4;u+=h)
{ // added
for(i=0,x=u,y=u;i<t;i++)
[x,y]=[(x-y-x*(S=x*x+y*y))*h+x,(x-y+y*S)*h+y],
rect(X=x*w+w,Y=y*w+w,(fill(w*cos(X-Y+t/17),w,w,.3),8))
} // added
t=++t%W}

In my environment, when I ran the code in a non-OpenProcessing environment, both codes worked correctly in Chrome, Firefox, Waterfox, and Edge.
The p5.js version is 1.11.11.

However, it seems that the poster encountered problems in their environment, so I thought any additional information would make it easier to address the issue. Since the contributor has already stated that they will work on it, that's all I have to say.

According to this post, this issue does not seem to occur in OpenProcessing, so I conducted my tests with that in mind.
https://x.com/TakagiHitoshi/status/2066766740097474724

[TakagiHitoshi]

By peeking actual running code (after adding the loop protection codes):
It seems to me that there is no initialization for _LP0. Therefore javascript blames the lack of it.
For a temporary workaround on this, I added _LP0 (and _LP1 also) initialization before the entire code, like this:
_LP0=_LP1=Date.now()+1e9

(draw=)
"_ => {
print(Date.now(), draw.toString());
t || createCanvas(W = (w = 200) * 2, W) + noStroke() + colorMode(HSB);
background(1, 0.1);
var _LP1 = Date.now();
for (u = -4; u < 4; u += h) {
if (Date.now() - _LP1 > 100) {
window.loopProtect.hit(8);
break;
}
for (i = 0, x = u, y = u; i < t; i++) {
if (Date.now() - _LP0 > 100) {
window.loopProtect.hit(9);
break;
}
[x, y] = [
(x - y - x * (S = x * x + y * y)) * h + x,
(x - y + y * S) * h + y
], rect(X = x * w + w, Y = y * w + w, (fill(w * cos(X - Y + t / 17), w, w, 0.3), 8));
}
}
t = ++t % W;
}"

[inaridarkfox4231]

While not strictly necessary, when posting source code, using code blocks makes it easier to copy and paste, which I think makes it easier for contributors to work with the code.

https://developer.mozilla.org/ja/docs/MDN/Writing_guidelines/Howto/Markdown_in_MDN

[TakagiHitoshi]

Thanks for your tip.

…

On Tue, Jun 16, 2026 at 8:50 PM INARI_DARKFOX ***@***.***> wrote: *inaridarkfox4231* left a comment (processing/p5.js-web-editor#4183) <#4183 (comment)> While not strictly necessary, when posting source code, using code blocks makes it easier to copy and paste, which I think makes it easier for contributors to work with the code. https://developer.mozilla.org/ja/docs/MDN/Writing_guidelines/Howto/Markdown_in_MDN — Reply to this email directly, view it on GitHub <#4183?email_source=notifications&email_token=ASK44J3LICKP76PXVOL3DJD5AEX6TA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZRHA2DAMBUGYZ2M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLDGN5XXIZLSL5RWY2LDNM#issuecomment-4718400463>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ASK44JYEVN24ETWUEBUDOSD5AEX6TAVCNFSNUABEKJSXA33TNF2G64TZHM2TOOJUGA2DQNZ3JFZXG5LFHM2DMNZRGQYDOMRSGOQXMAQ> . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS <https://github.com/notifications/mobile/ios/ASK44J7G76R24J7K4KKGP6T5AEX6TA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZRHA2DAMBUGYZ2M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJKTGN5XXIZLSL5UW64Y> and Android <https://github.com/notifications/mobile/android/ASK44JYFS4HWFC3JRAYKYZ35AEX6TA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTINZRHA2DAMBUGYZ2M4TFMFZW63VHNVSW45DJN5XKKZLWMVXHJLTGN5XXIZLSL5QW4ZDSN5UWI>. Download it today! You are receiving this because you were mentioned.Message ID: ***@***.***>

[inaridarkfox4231]

I didn't realize you were talking about p5.Editor. I apologize.

I have confirmed that it is indeed not working.

[inaridarkfox4231]

Reference: Differences in handling between OpenProcessing and p5.editor.

function setup() {
  createCanvas(400, 400);
    noLoop();
  draw=_=>{
  for(i=0;i<4;i++)
    for(k=0;k<2;k++)
    console.log(i),
    console.log(k)
  }
  console.log(draw.toString())
}

p5.Editor

OpenProcessing

Personally, I don't like the range of parentheses aren't explicitly defined, so I'm hesitant to call this a bug. However, for those who prefer this consistent behavior, it might be worth addressing.

This issue is related to the editor's specifications, so I don't think the version is relevant. The same behavior occurs in version 2.3.0 as well.

[inaridarkfox4231]

Personally, I don't think it's a good idea for p5.Editor to follow the OpenProcessing approach.
This is because OpenProcessing cannot detect infinite loops if you write code like the following:

/*
  The OpenProcessing method cannot detect infinite loops.
  This is just one example.
  Running this code with OpenProcessing will crash the browser.
*/

function setup() {
  createCanvas(400, 400);
  draw=_=>{
    background(220);
    let r=0;
    for(i=0;i<3;i--){
    console.log(1);
    r++;
    if(r>999999){noLoop()}
  }
}
console.log(draw.toString())
}

noLoop() will not be executed. The browser will stop before that for safety reasons.
Running this code with OpenProcessing will crash the browser.

This is a different reason, since p5.Editor is also used by beginners, I believe that the specification that generates an error if the parentheses in a for loop are omitted is necessary from an educational standpoint.

I don't think it's a good idea to introduce a system that can't avoid fatal errors because you want to write messy code.

[ksen0]

(For context why this was added in the firstp lace, #4080 may be helpful)

[inaridarkfox4231]

I wasn't aware of this because no reference for this issue was provided, but it seems that a solution is in progress with this pull request.

pull request: #4184