loop() not implemented in coffeescript mode?

Processing with Other Languages

benj.n.su..

loop() not implemented in coffeescript mode?

in Processing with Other Languages • 5 months ago

hey all.

i was creating a simple on/off switch as i often do when debugging, println-ing lots of stuff etc.

in processing, that looks like this, and it works as i'd expect it to:

boolean onOff = true;
int i = 0;
void setup(){
}
void draw(){
println(i);
i++;
}
void keyPressed(){
if(onOff){
println("turning off");
noLoop();
} else if(!onOff) {
loop();
println("turning on");
}
onOff = !onOff;
}

The equivalent coffeeScript version in CS mode will not even run. When I comment out loop(), it works just fine. Can anyone confirm?

onOff = true
i = 0;
setup: ->
size 200, 200
draw: ->
println i
i++
keyPressed: ->
if onOff
println "turning off"
noLoop()
if not onOff
println "turning on"
loop()
onOff = !onOff

Replies(18)

GoToLoop

Re: loop() not implemented in coffeescript mode?

5 months ago

I wasn't aware that CoffeeScript had become a new mode in Processing!

My knowledge of it is almost nil. Nevertheless, here's my silly attempt:

draw: -> print "#{frameCount} - "

mousePressed: -> keyPressed()

keyPressed: ->
  if looping = not looping then println "turning on..."
  else println "\nturning off..."

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

Thanks GoToLoop. I think I might be missing some code from your example. As far as I can tell, your example does not make use of the native loop() from the Processing API. I can successfully create a program that prints turning on and off, it's just that using loop() causes the program to not run at all. I can even stop the program using noLoop(), I just can't restart it using loop().

As an aside, since you said you are new to the CS Mode, there is one caveat to it and that is that all of the core API functions that you'd have to write void in front of such as setup(), draw(), mousePressed() must use object literal notation when calling the function. The stuff inside does not need to be that way. See my example above, or: https://github.com/fjenett/coffeescript-mode-processing

GoToLoop

Re: loop() not implemented in coffeescript mode?

4 months ago

... must use object literal notation when calling the function.

You mean put a colon right after function's name and remove the = sign?

Variable looping is a "secret" field in Processing/Java's API

which determines whether draw() is to be auto-called or not.

That is the same field which loop() & noLoop() actually ends up modifying!

However, if that variable exists in Processing/JS's API, it has another name!

And I dunno if that can be reached to be modified as Processing/Java's API generously allows!

And I dunno neither JS nor CS yet!

Back to Processing/CS mode, it seems like that the fact that loop()

was not implemented there is not a good sign!

It means it's way behind official Processing/JS mode!

In my opinion, Processing/CS mode would be more viable if it was "compiled" into Processing/JS 1st,

instead of trying to get a direct JS compilation.

Doing so, anything offered by official Processing/JS mode would also be available to Processing/CS!

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

re: object literal notation, exactly. all the functions akin to the ones i mentioned go like this:

setup: ->
      size 400,400
      background 255

draw: ->
      println "drawing! yay!"

mousePressed: ->
      noLoop

functions you write on your that are not inside an object literal you created use -> or => depending on circumstance.

i don't know Java so i'm not familiar with what a secret field is, but i guess it's not so secret anymore! there is a variable called looping as part of the p.loop in PJS. You can compare it's use here. It's too complicated for me to make out how it relates to the rest of the running of the sketch.

turns out the loop() is a bug in CS Mode that the creator is aware of. you can always write CS directly in a file attached to the canvas but i think you have to call the processing object when using the PJS API (p.background). Personally, I find that slightly annoying which is why I'm enjoying CS mode despite it's short comings of which so far there are only two others: file tab usage differs slightly (true for JS mode too), and syntax highlighting is not yet implemented.

I'm not sure of the ins and outs of how the conversion process works but I imagine it would be significantly more complicated to convert it to P/PJS first. Code written in CS that is not in the core API would have to use the CS compiler anyway. While coffeescript compiles directly to javascript (the motto is "it's just javascript") it has a very opinionated way of doing so. Not incidentally, the person who created CS mode is the exact same person who created JS mode so I imagine he has his reasons for how he implemented it.

GoToLoop

Re: loop() not implemented in coffeescript mode?

4 months ago

From the Processing.JS source link you gave, it's indeed true that variable looping exists there.

However, rather than storing boolean values as the 1 from Processing/Java, it's a regular numerical one!

But anyways, just above it, there's 1 called doLoop. And it's being initialized as true!

So, just made some tweaks from my poorly CS attempt above, and also, a Java one for comparison.

Java:

// https://github.com/processing/processing/blob/processing-0197/core/src/processing/core/PApplet.java#L562

void draw() {
  print(frameCount + " - ");
}

void mousePressed() {
  keyPressed();
}

void keyPressed() {
  if (looping = !looping)  println("turned on...");
  else println("\nturned off...");
}

CS:

# https://github.com/processing-js/processing-js/blob/master/processing.js#L2241

draw: -> print "#{frameCount} - "

mousePressed: -> keyPressed()

keyPressed: ->
  if doLoop = not doLoop then println 'turned on...'
  else println '\nturned off...'

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

i tried it out and while it did work as a program onto itself, it didn't stop and start the actual sketch. i don't think CS Mode has access to the PJS object's variables. ah well. i might be able to get by on slowing down the framerate. I really only use it when I want to see how the numbers are playing out and they're zooming by at 30 frames/second.

again.

thanks for hashing it out with me!

GoToLoop

Re: loop() not implemented in coffeescript mode?

4 months ago

Alas, I was kinda expecting you to be able to tell me how to access those internal variables in case I'd fail.

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

Alas, I am flattered, but you are greatly over estimating the extent of my capabilities.

Next time I post in issue to the creator's repository, I'll throw in that question.

if you can't wait, here's the link: https://github.com/fjenett/coffeescript-mode-processing/issues?state=closed

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

update!

Jeremy Ashkenas stopped by the GitHub issue list to explain that loop is a keyword in CoffeeScript.

per his example:

loop
print '.'
break if finished

GoToLoop

Re: loop() not implemented in coffeescript mode?

4 months ago

Perhaps I've got a patch solution while waiting for a definitive fix.

Just be warned that I don't have any Processing 2 installed, much less CS mode!

So I can't check out whether it works or not!

I've found out that CS allows insertion of expressions or even entire blocks of raw JS by the use of backticks: `raw JS`

Even if it works, I still wanna know how to access internal Processing.JS class variables through CS!

# https://github.com/processing-js/processing-js/blob/master/processing.js#L2241
# https://github.com/fjenett/coffeescript-mode-processing/issues/16
# http://forum.processing.org/topic/loop-not-implemented-in-coffeescript-mode

isLooping = on;

setup: ->

draw: -> print "#{frameCount} - "

mousePressed: -> keyPressed()

keyPressed: ->
  if isLooping = not isLooping
    println 'doLoop is on...'
    `loop()`
  
  else
    println '\ndoLoop is off...'
    noLoop()

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

DUDE.

It worked!

The only thing standing in the way of your code working by being directly cut and pasted into the IDE is that you have to add a setup: -> function.

That was amazing. Well done, sir. Well done.

I hope you have a GitHub account so you can post your solution to the issue tracker. If you do not and don't want to create one, let me know and I'll post and give credit with a link back here.

Tremendous.

GoToLoop

Re: loop() not implemented in coffeescript mode?

4 months ago

YAY! How nice I've got it!

So, a setup: -> is needed? Funny I don't need that in P/Java.

I don't have an account @ GitHub.com. So plz, feel free to post it there.

And also, ask them about accessing internal P/JS class variables!

Hope next Processing beta version be more friendly to Linux and has a working <CTRL+T>.

So I can program in CS for real rather than a pseudo make-believe one. HEHEHE

And here's the link about embedding JS into CS:

http://CoffeeScript.org/#embedded

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

Done!

https://github.com/fjenett/coffeescript-mode-processing/issues/16

PS. I just realized that it took a person aptly named GoToLoop to figure out how to use loop(). Of course we now know you will have to change it to GoTo`Loop` if you ever want it to compile to Coffeescript.

GoToLoop

Re: loop() not implemented in coffeescript mode?

4 months ago

Thx for posting there!

My nick is about using keyword GoTo to get back to label Loop:

A pity neither Java nor JS/CS haven't implemented goto keyword!

I'd surely use it at times!

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

update.

bug is officially fixed. funny thing; now we've come near full loop to the second solution you suggested, except it wasn't a solution because it hadn't been implemented yet.

from fjenett here:

I added "doLoop" as replacement for "loop" and added a check/warning to the compiler. Plus, there is an example showing the use of both the loop control structure and the noLoop() / doLoop() combo.

developing solutions that haven't even been invented yet. ahead of your time.

it is finished.

GoToLoop

Re: loop() not implemented in coffeescript mode?

4 months ago

Nice they've made up a new name for previous non-working loop().

Although it's the same name as the internal PJS variable -> doLoop.

Nevertheless, the backticks trick is useful when you want raw JS "real power"!

Seems like they didn't pay much attention to that extra request about accessing internal PJS objects though.

However, after following the newest commit below:

https://github.com/fjenett/coffeescript-mode-processing/commit/aca53157e978e8a69bcdf9b3d4d26b4221851b6c
loadStrings = processing.loadStrings;
loadType = processing.loadType;
log = processing.log;
- loop = processing.loop;
+ doLoop = processing.loop;
mag = processing.mag;
map = processing.map;
match = processing.match;

I see various Processing function names as alias of their analogous processing object function!

So, I've just thought that maybe the elusive & secret internal variable doLoop

can be accessed by prefixing processing. to it:

processing.doLoop

Since I can't compile/run it, can you plz check out whether my previous example works w/ this "hack" now?

# https://github.com/processing-js/processing-js/blob/master/processing.js#L2241 # https://github.com/fjenett/coffeescript-mode-processing/issues/16 # http://forum.processing.org/topic/loop-not-implemented-in-coffeescript-mode

setup: ->

draw: -> print "#{frameCount} - "

mousePressed: -> keyPressed()

keyPressed: ->
  if processing.doLoop = not processing.doLoop then println 'doLoop is on!'
  else println '\ndoLoop is off!'

benj.n.su..

Re: loop() not implemented in coffeescript mode?

4 months ago

nope. tested. does not work.

i don't think that the IDE has any access to the processing objects internals. if you were doing this in a .js/.coffee file, you'd have to get the processing instance first. i'm now venturing beyond the area of my knowledge so i'll stop talking here.

Here's an example of what will work using the new doLoop(), FYI: print does not appear to be working either, must use println.

onOff = true;
setup: ->
draw: ->
#prints one frameCount then stops
println "#{frameCount} - "
noLoop()
mousePressed: -> keyPressed()
keyPressed: ->
#prints next frameCount
doLoop()

GoToLoop

Re: loop() not implemented in coffeescript mode?

4 months ago

Hey BenjNSugar! Thx for trying! Well, it's fixed anyways!

But my request was more about how to have the same freedom to access the internals of Processing

in JS/CS as much as I have in Java!

Alas! Developers of PJS were not as "friendly" as the ones in PJava.

More than half of internal variables were sealed away against any public access! And that includes doLoop !

I didn't know JS had any means to control access to "classes" as Java!

Even though I still dunno much about JS nor CS, I've searched for some way to hack my way in. But failed!

Seems like what determines if a member variable is protected or public is the internal usage or not of prefix this.

So, a variable declared w/o anything can only be accessed by a function and any sub-functions of it.

To open it up, they have to be declare it w/ this. as prefix.

Or create another variable w/ this, and then assign it w/ the reference of a variable you want public access.

In short, I'd have to hack PJS's source code @ function "class" Processing @ line #1389,

from doLoop to this.doLoop. Or even p.doLoop, since p = this there too!

But then, it'd only work for me alone, since every1 else would be using the original processing-1.4.1.js file!

Ah! And something strange I've found out in PJS source @ lines 941 to 1082:

They've created a "hash" variable colors{} filled w/ lotsa literal color names w/ corresponding hexadecimal web color values, like: aliceblue, blanchedalmond, darkgoldenrod, ghostwhite, mintcream, tomato, etc.

However, like doLoop, it's privileged to PJS internal functions, since it wasn't declared w/ this.

And even puzzling, it doesn't seem to be used anywhere at all!!!

I'd really like to be able to use color names instead of numbers sometimes.

Hope they free 'em up at next version!

Well, that's it. Till next time!

Top Reply