It's so weird how it even preserves spaces. So (x=>x+2).toString() is "x=>x+2"

It even preserves comments

+50 to whoever writes a JS quine using that :P

@Steffan It preserves everything, indenting as well, it just gets it straight from the source code

I know

It's just really cursed

Yeah

I could probably overload one of my modifiers with sort-by. I don't think it needs to preserve arity.

Other similar possibilities are min-by and max-by.

oh, good idea

also maybe map over prefixes

@Steffan How does dyadic filter work?

LDQ: Working on an idea for a tape-based language that's (slightly) more usable than BF. Should cell values all be numbers, with strings encoded as numbers using a bijective base system? Or should there be a separate string type?

@Steffan Not in this case

I have modifiers - functions that operate on functions, so those functions' source is referencing stuff that's above their own scope scope.

I guess you could argue then that your language is compiled to whatever internal representation JS uses for its functions, which is a lower-level "language" than JS itself. Kind of like compiling your esolang into Python bytecode.

¯\_(ツ)_/¯

Also, last night when I finally hooked up the function creator to a stack, it just worked.

\(ツ)/

I suspect it may break at some point in future

put your hands in the air like you don't care :P

lol

@DLosc then all langs are compiled

I don't think so. If you go from the original source code straight to executing each command as you read it, that's interpreted.

@Seggan That's actually a really common cheating quine approach with JS and the easiest way to do it

@DLosc What if you have to construct an AST?

f=x=>"f="+f

@Seggan f=_=>'f='+f

ninja'd

@emanresuA =_=

@emanresuA Then you could perhaps argue that that's a form of compilation, I suppose.

Pip constructs an AST IIRC

Yes

@DLosc I don't think so

@NoHaxJustRadvylf I don't either, but I haven't come up with a knockdown proof as to why.

To me a compiled language is one that turns the code into a different, independently storable and runnable language

@emanresuA so does fig and choco

vyxal

@NoHaxJustRadvylf this

You can't just stick the parsed AST into a file, you'd need to represent it in some way, and what is code if not a string representation of an AST

Interpreters and compilers both construct ASTs

not BF

The line gets kinda blurry between them anyway

"What is grief code if not love persevering a string representation of an AST?" - Vision Radvylf

@user I don't think it does honestly

@user yeah. is python interpreted or compiled? what abt java?

@NoHaxJustRadvylf JIT makes it so you could call it both

@Seggan Python is interpreted, or if you use a JIT implementation, it's interpreted but the interpreter contains another compiler.

@Seggan I mean, there are different implementations of Python

Woohoo!!! 2 + 3 = 5!!!

Aren't all languages interpreted at the lowest level, though?

no

not IMO at least

@user CPython

@user JIT falls pretty cleanly into interpreter territory IMO. Sure, there's a compiler inside the interpreter, but there's different things inside different things all the time.

@user Sure, you could argue machine code is an interpreted language. But the compiler doesn't run the machine code (unless it does, in which case the whole thing together acts as an interpreter)

@user machine code is directly electrical pulses flipping transistors inside a CPU

Which is interpretation not compilation

Or some fourth thing if you feel like it deserves its own category

i dont think JIT is an interpreter

the computer still directly runs the output

just a bit later than AOT

@emanresuA Might wanna get that fixed

@Seggan Did you swap "compiler" and "interpeter"?

"still directly runs the output" sounds like it's talking about an interpreter

no

im talking about the computer

by ur definition everything is interpreted then

> compilation is not restricted to ahead-of-time compilation to native machine code. A compiler is, more generally, a program that converts a program in one programming language into a program in another programming language

Source

Sure, there's a compiler in a JIT interpreter, but the thing as a whole is an interpreter

It doesn't make sense to call an apple pie an apple

@NoHaxJustRadvylf What about a caramel apple? :P

I'd still say no, since "caramel apple" is just short for "apple and the caramel around it" :p

CMQ: "car-muhl" or "car-uh-mell" or "cair-uh-mell" or "cair-muhl"?

Usually "car-muhl," occasionally "cair-uh-muhl"

@NoHaxJustRadvylf "cair-uh-mell"

I usually say "cair-uh-muhl", seems like at least where I live it's a pretty even mix between them all

@NoHaxJustRadvylf caird coinheringaahing

Also, Trianguish update: Mouse controls are now almost fully working!

im honestly so excited more than i ought to be probably

I just need to add some sensible logic for what to do when you try to do multiple mouse actions at once (e.g., right click and middle-click-drag)

redirect to a [rickroll](www.example.org)

i need to stop being so fail

Ultra secret preview:

umm why are so many mouse actions allowed in the first place lol

@NoHaxJustRadvylf This looks like a puzzle game =)

is this one of those things where theres an instruction pointer and the temporal order of actions matters or is it all mostly instant

@thejonymyster Just convenience stuff. Left click drag for moving the screen around, left click for moving the hexagon cursor thingy, right click for op chooser, right click drag to copy an op from somewhere else on screen, middle click for rotate, and middle click drag for auto wire routing

@DLosc thats what ive been saying lol its like a zachlike

@thejonymyster you're missing the https:// :P

i knowww but its too late to fix it once people see it

OTL

@thejonymyster There's no IP, things can happen simultaneuously (but it's not instant)

scary, hype though

@thejonymyster this is an example of a rickroll :P ez

but that link just leads to never going to give you up by richard astley

Is his name actually richard

my life is a lie

Fun fact: the rickroll turns 35 today

@NoHaxJustRadvylf Richard Paul Astley

...ort for richard? nah

this looks more interesting

(spoiler: rick astley is not shirtless)

@DLosc it might actually make for a fun puzzle game, now that I think about it

Like, you have a limited inventory of pieces, and you have to pass test cases

@DLosc Just like Jelly's does

So [F][filter][x][y] (where F is a dyad) would loop through x where the var is Y, and call F(Y, y) on each

oh god...I just got jumpscared so bad by myself

I forgot I implemented a feature, so I was playing around with another aspect of it, then it worked and I was not ready for it

And with that, I believe Trianguish is now feature complete!

@Steffan Oh, okay. So it's filtering x but with y as a constant second argument to the function.

yes

@NoHaxJustRadvylf is it turing complete?

Optionally :p

The default behavior is restricting numbers to be between 0 and 215, but you can set it to allow arbitrary integers, which allows turing completeness

@NoHaxJustRadvylf why 215? i get its 6^3-1, but why that amount?

Two reasons: I think it's cool to use base-6 for a hexagon-centric language, and it's perfect for representing operators' IDs alongside their rotations, which is important since there's a lot of self-modifying stuff

and did you just stop at 216 cause it was close enough to 256 without going over? :P

or is that where the operator id thing comes in

Mostly just the operator ID thing I think, but I don't remember specifically since I came up with that part of the language multiple years ago

years? Wow

Yeah, I had the idea for this language at least two years ago

@NoHaxJustRadvylf The third

@NoHaxJustRadvylf Thesecond

Caramel sucks, anyway. Compilers > Interpreters > Chocolate > Caramel

@emanresuA I was gonna say something similar, but then I remembered that I started working on Pip in 2014, so... ¯\_(ツ)_/¯

One thing that sucks about compiled languages (or maybe just bad ones, idk) is not having a REPL

Just bad ones

I think it's just bad ones. Haskell has a REPL.

But yes, much harder to make REPLs for compiled languages

Even Java's had a REPL and eval-like stuff for a while now

And with JS, I forget just how convenient it is to be able to hit Ctrl + Shift + C and do some little task in the language I'm most familiar with

@user Typically, caramel is sucked :P

@NoHaxJustRadvylf Yeah, although often Vyxal's better fir me

How do you type it though?

@DLosc Nightmare fuel: caramel that sucks people

@NoHaxJustRadvylf Not Ctrl + Shift + I?

I thought Radvylf meant copying in the terminal?

@DLosc I usually use I yeah, I thought that C went straight to the console but apparently not

(My Chrome-based muscle memory is for Ctrl-Shift-J, but that doesn't work as well on Firefox.)

Ah it's J I was thinking of

TIL that's a shortcut

I've been using Ctrl+Shift+I all this time. Vivaldi makes a new note with Ctrl+Shift+C so I got confused

My personal favorite way to open the console is Ctrl + Shift + W. Only works sometimes though.

What does that one do?

Close the window :p

Sometimes I'll have a terminal open behind it

Isn't that Ctrl+W?

That's the tab

Oh right

Weird, Ctrl+Shift+W did nothing to my window

Which is probably good lol

If Ctrl + Shift + T didn't exist I'd have added a warning :p

Eh, this browser's set to open all previous tabs on startup anyway

@NoHaxJustRadvylf These days, I can open a new browser tab/window and enter a short URL to do some little task in one of the languages I'm most familiar with. Which is very nice. Thanks @Dennis, @pxeger, and @emanresuA :)

True lol

And hopefully soon RTO will be another of those short URLs, which is going to have REPLs for every language (where possible)

It's weird how so many of these TIO clones popped up around the same time

There was also TiB but Bubbler's not working on that (at least not right now)

I think we all started discussing TIO after someone brought up how increasingly outdated the language versions on it were or something, and then we all spent a while talking about replacing it, and that inspired a bunch of us to make clones

Speaking of, I should really be working on RTO not Trianguish lol

Ah that makes sense

@NoHaxJustRadvylf Do whatever makes you happy

Or gives you more anguish, I guess, given that you seem to be a masochist

@user Trianguish makes me happy, but not working on RTO makes me sad. So Trianguish only has a slight lead, due to my inability to focus on the long term :p

Well, if you have a hard time deciding, you can do what I do: don't work on either of them and instead waste time doing nothing. You'll feel anxious/guilty about it for a few weeks but after that you'll have lost the motivation/momentum and you can move on to a new project to excitedly start work on and then abandon

Relatable :P

Same :p

@user I came here to have a good time, and I'm feeling really attacked right now :P

See The Tarpit for my attempt at an explanation.

I bet there were a lot of corner cases to cover :P

Lol

Opinion: Regexes should default to matching the whole string, and instead of ^$, you can put a $ at the start, end, or both to indicate there can be additional stuff there

That leaves less flexibility for optionally matching an entire string, or just a substring

e.g. (^...$)|(...)|(^...)|(...$)

idk about you but I usually just use regex to extract a part of a string

The entire string could contain all sorts of junk that no one wants to waste an .* on

@Zionmyceliaadamancy That would still work, just in reverse

E.g., ($1$)|2 for "contains a 1 or is the string 2"

@user Hm. The main thing I use regex for is validating the formatting of stuff.

(that and parsing HTML)

@DLosc Thanks for that. It means a lot to me to have DSO, which started as a hobby project, usable to people other than me.

" No-op
. Wall
= Push 0
0-9 Mul 10 and add 0-9
~ Pop and discard
$ Dup
> Move front to rear (rotate right once)
< Move rear to front (rotate left once)
[ Move to previous deque
] Move to next deque
{ Carry 1 to previous deque
} Carry 1 to next deque
+-*/%! add sub mul div mod not
^ neg
)( inc dec
;`:' numin charin numout charout
_\ space newline
& debug; dump current deque as space-separated numbers
# height
@ halt
? random in [0, n)

The command set for the "tape of deques" thingy I mentioned before

LDQ: Should I stop trying to cram everything into non-alphanumeric ascii?

Looks good :)

You need a 0 command

Oh wait nvm I'm stupid

There is "push 0" near the top

Isn't that =?

IIRC chris pressey had something older than that

@Seggan it's been done. but personally I don't think it's a proper quine because the call works by reading the source code, which isn't allowed under proper quine rules

now, very old versions of Firefox actually had a decompiler which reconstructed the source from the byte code, which was fun

@Neil Let me guess, it produced misleading results?

@Zionmyceliaadamancy Martin Ender mentions Emmental in Labyrinth readme

That was what I was thinking of

@emanresuA actually I was thinking of ${new RegExp} (see if you can predict what that evaluates to before doing it)

bah, of course that was in backquotes, but chat...

Is it (?:)

yup

Nvm, it includes the /

oh yeah I forgot that bit

also it acted as a form of pretty-print

mind you the source code reading version wasn't foolproof, I think there was at least one version where if you edited the source file and then tried to do .toString() on a function it would crash because it couldn't find the code

Lol