The Nineteenth Byte

12:07 AM

Wait nvm it works now

Oh cool

@RadvylfPrograms The code loaded from github

@RadvylfPrograms wait 1.'3?tf kinda float is that? Some sort of imperial bit system?

So, as for arrays, there is no array literal syntax. However, the built-in , can be used to "build" arrays, in a somewhat haphazard process that occasionally needs parentheses and occasionally doesn't

@lyxal This is a cool feature of rSN: a ' in the fractional part makes everything after repeat infinitely

So 1.1'2 is 1.1222222222...

And 1.'3 is 4 / 3

Romanp

It returns Null on the online version. Is that intentional?

12:10 AM

What's 1.'2'3?

@Romanp No, but if it does that, it's canon

Frick, seems my regex for detecting numbers is missing '

Well, no changing the language now

Okay, now functions. There are two flavors of functions: functions and built-ins

Both are called identically, either as infix functions or with {}

E.g., f{x} calls f with x as an argument, and f{x}{y} or x f y are two ways of calling it with two arguments

To make a function literal, simply use ~ in the same way you'd use ~. So, x ~ x + 1 is a function to increment a number.

hows't do I name my function

Now, our first ._.MC: Given two ints, divmod them ((x / y, x % y))

@lyxal With variables, since all functions are lambdas. Variable assignment is :

Prolly should've mentioned that earlier lol

a: 1 assigns 1 to a, classic stuff

func : ~ func

Recursion go brrr

Any string can be used as a variable name sometimes

@lyxal Yeah, func is being called with ~ as its first argument and Null as its second

Handy tip: dbg.pars{...} will show you how something was parsed

12:16 AM

oh that's helpful

@RadvylfPrograms x ~ y ~ x/y,(x%y) feels like it should work but knowing you it probably doesn't

x ~ y

curry

Plus I can't test it because you forgot to make the repl work on mobile

and i don't think rsnbatwpl has pattern matching for something that doesn't have actual literal syntax :P

@lyxal You got the precedence wrong

(x/y),x%y should work

12:20 AM

Called it

Knew it wasn't that simple

Still doesn't work on mobile

So, some quick control flow stuff: if {condition} {run} can be used for if statements, cond {condition} {if do} {if don't} can be used for if/else or ternary statements, and for {initial; condition; after} {run} can be used for looping (same w/ while {condition} {run} and do-while {condition} {run})

._.MC: Nth fibonacci number

Handy tip: . can be used to recurse in a (possibly unnamed) lambda

oh yeah when you first released the online interepreter i tried using infix if and that did... not work lmao

Just keep in mind, due to currying, it only works with single-arg functions

@UnrelatedString it actually does, you just gotta switch the intuitive ordering of the condition and thing to run

it was definitely something to do with evaluation working differently from {} calls

I've been busy and it's almost working

12:24 AM

hype

Make sure to use cond not if

@RadvylfPrograms x ~ cond {x < 2}{1}{.{x- 2}+.{x- 1}} but knowing you that's also probably wrong

and not cons either :P

Oh, and x-2 and x-1 will be interpreted as variables, since -s are allows. Try x- 1

You can thank SwiftKey for that :p

@RadvylfPrograms now just hold on a second. Your readme says =

12:27 AM

Does rSNBATWPL have any form of tail recursion?

Quick tip: f $ x is a built-in which calls f with x, so the final {...} in most places can be replaced with a $

@lyxal Oh frick it's = now, yeah

@emanresuA lol no

It works!

Another quick tip: {...} in a context which is clearly not calling the preceding thing, e.g., 1 + {2}, will auto-insert .

@emanresuA That's a local host link

So {.{x- 2}+.{x- 1}} can be golfed to {{x- 2}+{x- 1}}

12:29 AM

Oops

@RadvylfPrograms is that why {} is an infinite loop?

Yep!

._.MC: Pass an integer by reference

oh no

do not do this to me

I'm sorry what.

You implemented a by-reference system?

We're lucky there's no pass-by-name bs in this language

12:33 AM

are you sure there isn't

Good point

nscopd-id

That's basically pass-by-name :p

Ffs next thing you'll be telling me there's pass by result

rSN includes pass-by-value, pass-by-reference, and pass-by-name, so that you can choose the most readable one for your scenario (TM)

ATaco

Flashbacks to early autohotkey

12:34 AM

It works!

a = 1; aid = id{a}; (aid) = 2; print{a}

ATaco

“You want an array? Just name a bunch of variables A2 A2 A3…”

Don't give radvylf ideas

rSN does not have comments by the way, to encourage readable code. If you need comments, your code isn't doing its job.

@ATaco I...I didn't know about arrays when I first started using JS, and I'm not proud to admit this but...eval("n" + n) may have occured a few times

oh on

12:36 AM

@RadvylfPrograms well that's just simulating references/pointers, no?

*no

@lyxal Basically, yeah. So you can pass an id to a function to pass a reference to it, and an nscopd-id to pass its name

Where id gets wacky is when you use it as the parameter of a function literal

@RadvylfPrograms what do you mean it's name?

@RadvylfPrograms honest question, did you have any particular strategy for passing those groups of variables around? because i can't think of any way that would be possible, but i also don't know js :)

@lyxal Doesn't seem to work

12:38 AM

a = 1; f = id{a} ~ Null; f{2}; print{a}

@UnrelatedString No, but I also didn't know functions were a thing. One of my early SO questions was asking about how to goto in JS because my only programming experience was on the TI-83

ah

lmao

Also, btw, you can give variables arbitrary names: "x y" = 2; var{"x y"}

Returns null, which means it probably errors

You can use @ to dereference IDs, if you need to do so

What does nscopd-id return?

12:40 AM

An id, but without a scope attached, so it uses the one it's dereferenced in

Because hopefully, your idea of pass by name is completely different to the really cursed method I'm thinking of

@RadvylfPrograms okay good you don't have true pass-by-name

Ah frick my github branches are cursed

Hey whaddya know, I forgot to implement .

Good job

Wait no, worse

I did, but +. is read as a single operator

So you need a space

ATaco

12:43 AM

C has a built in qsort?!

C has nothing built in why qsort?!

._.MC: A function which returns the first n prime numbers

So there's pass by value, pass by reference and pass by what I assume is value-result

@RadvylfPrograms What's up with this?

It's hard to explain how nscopd-ids work, but basically, if I pass nscopd-id{"a"} to a function f = x ~ (a = 1; @{x}), it'll return 1

Does that textually substitute a where x is?

12:47 AM

@emanresuA : should be =, need a ; after the function definition

@lyxal No, so not fully pass-by-name-cursed, but close

You should add ASI

@RadvylfPrograms good.

At least you made one sensible decision

@emanresuA Yeah, it's gotten me a few times. Remembering ;s after for loops is especially tricky

Fun fact about the raisin language: id{'("prints"')} = ncon{"Hello, World!"} is a hello world program

Oh god

What's '(?

Are those char literals?

(ncon's not anything cursed, it just wraps a thing into a single element array)

@emanresuA Yeah

You can't have parens in strings

12:51 AM

... why

Parens are parsed before anything else

...

... why

Well because ; used to be first, but that broke the parens, so I moved parens to the top and downgraded ; to be right before operators

'Do youu have any idea what a tokenizer is

12:54 AM

Yes, and I have all of the pieces of one, but scattered around for various reasons

i'm beginning to think you started by just trying to make a praclang and eventually it was so cursed that you decided to embrace it

Fun fact: a trailing ' does some really cool stuff to the bracket autocompletion

I'd show y'all but you physically cannot dbg.pars it

Basically it eats the } it adds, meaning it's still unbalanced brackets after the autocompletion, which is undefined behavior

@RadvylfPrograms really cool

Knowing this language, I think cool is the wrong term

@emanresuA even with syntax fixes this hangs

Oh, even better, in a blank program ' will be the same as '' since the way I find the thing being quoted is .slice(-1)

12:56 AM

@lyxal Got it to work

@emanresuA nah just takes a while for n = 34

turns out it wasn't hanging

No surprise - it's doing five million recursive calls in a lang that's not particularly known for efficiency

._.MC: Memoized fibonacci

1:14 AM

@RadvylfPrograms rSNBATWPL, 84 bytes: f=n~for{i=1;i<=n;++$i}{p=0;for{b=2;b<i;++$b}{if{not$i%b}{p=1}};if{not{p}}{print$i}};

it prints them, not returns them

Nice!

wait that outputs primes up to n, not n prime numbers

@Steffan Does 0~~ do the not$ shorter?

that doesn't work

You can probably do $p=1 too

1:20 AM

could do not$p tho

I'm especially meaning the not$i%b

IIRC or is short-circuit

So p or print$i is probably shorter

And that trailing ; probably isn't needed

that doesn't work either

Ah, oops

It doesn't short circuit apparently

@Steffan p=b=2 should work

Since p will be set to b=2, which is Null

I'm absolutely sure you can golf at least one of those ifs with the $ trick

if{0~~i%b}$p=1, if{p!1}$print$i

Since rSN does recursion so golfily though, wouldn't be surprised if that'd save a lot over for loops

._.MC: Given a rectangular array of arrays, return an array of the sums of the columns

rSNBATWPL, 81 bytes: f=n~for{i=1;i<=n;++$i}{p=0;for{b=2;b<i;++$b}{if{not$i%b}$p=1};if{not$p}$print$i};

E.g., ((1, 2), (4, 8), (10, 20)) -> (15, 30)

You don't need to count the ; btw, it's needed if you do stuff afterward, but that's also true in a lot of languages

1:31 AM

Why doesn't this return a range?

@emanresuA Need a $

$i

Also <=n't

Oh wack

Add a space

For some reason?

I threw an error!

@RadvylfPrograms rSNBATWPL, 72 bytes: f=a~for{i=0;i<size$a$0;++$i}{s=0;for{j=0;j<size$a;++$j}{s=s+a{j}{i}};s};

Try It Online!

1:35 AM

@emanresuA Frick

That's another incident of aftermath from the awaitification

wait a{j}{i} can be a{j}{i}

I'm going to fix ' floats and that bug, just a sec

i mean a{j}$i lol

Yeah, that's what I was trying to get you to do with the ifs lol

@RadvylfPrograms Okay, DSO will pull that code from github once it's pushed

@Steffan a$j$i?

1:36 AM

nope

you would have to do (a$j)$i

ah frick rtl

OPERATOR PRECEDENCE EXISTS FOR A REASON

@RadvylfPrograms Can/should we unpin the Off-topic TNB message now?

+1 to ^

Go ahead and do it, you're an RO :p

Better to ask for forgiveness than permission (goes to rename TNB)

Lyxal can do it, they haven't used RO magic yet

doneski

1:44 AM

Feel the 12

@RadvylfPrograms not all the time

sometimes it's good to check if someone wants something a certain way to avoid making more work for yourself

Well, the two bugs with rSN have been fixed

on DSO too, pretty impressive

@emanresuA You led me to 6 more errors with that

@RadvylfPrograms It lazy loads the code from github :P

Oh, I figured it was some sort of cool auto-update-detection, but I just realized it's probably like, a fetch lol

It's not a server, it's completely static :P

1:50 AM

._.MC: Given a number, return how many steps it takes for the Collatz conjecture.

Should be pretty simple

pretty simple is the terminology you use when asking in a regular golfing language or practical language

this is radvylf we're talking about

pretty simple doesn't exist :p

@Steffan x~cond{x!1}{{cond{x%2}{1+x*3}$x/2}+1}$0

doesn't work

you need to actually recurse

Wdym "actually recurse"?

rSNBATWPL, 43 bytes: f=x~cond{x!1}{1+f{cond{x%2}{1+x*3}$x/2}}$0;

1:53 AM

(I know it doesn't work, that's a different issue)

@Steffan You can remove the f

and the f=

It'll auto-recurse if something's in {}s with no function name

oh nvm it works

idk what happened

._.MC: Given a number, convert it to binary.

I assume there's no builtin for this

@emanresuA What does DSO use to join the header, code, and footer?

newlines (IIRC)

Doesn't seem like it:

... never mind, looking at the code

2:02 AM

rSNBATWPL, 44 bytes: f=x~(while{x}{a=conc{x%2}$a;x=floor$x/2};a);

Try It Online!

@Steffan Seems like infix would shorten that conc

recursive:

rSNBATWPL, 39 bytes: f=x~cond{x>1}{conc{f$floor$x/2}$x%2}$x;

oh right, x~cond{x>1}{conc{{floor$x/2}}$x%2}$x

36

rSNBATWPL, 34 bytes: x~cond{x>1}{{floor$x/2}conc x%2}$x

rSNBATWPL, 28 bytes: x~cond{x>1}{{x/2}conc x%2}$x

2:53 AM

._.MC: Given a rectangular array of arrays, find the product of the sums of all columns with positive products

(This one's intended to get y'all to use filt, rdc, flip, etc.)

is there a complete list of functions?

(almost typo'd that as fynctions)

No, but most of the useful ones are listed in the README. I'd just ctrl+f the source code for builtin: if you want to catch them all

I'll type up a gist listing them all, actually

But it'll be an hour or so

'cause I'm busy

@RadvylfPrograms there's only one instance of builtin: though

3:23 AM

Oops

partial: then

Oh frick there's over 100 builtins this is gonna be time consuming

rSNBATWPL, 111 bytes: f=a~(y=1;for{i=0;i<size$a$0;++$i}{s=0;p=1;for{j=0;j<size$a;++$j}{s=s+a{j}$i;p=p*a{j}$i};y=y*cond{p<0}{1}$s};y);

3:35 AM

._.MC: Shortest empty array literal (two categories: requires parentheses (except at the far right), and doesn't require parentheses)

I have four bytes for the first category

Oh boy, copy does some really neat stuff with IDs and functions

conc{}{}

actually conc{}$

Too bad con wouldn't work here

@Steffan Since this already requires parens, ()conc

3:54 AM

raisin-batwaffle built-in reference: gist.github.com/Radvylf/b8dd14834a0ef519238e0b9076807a76

._.MC: Make a dictionary. Should be a function which allows, at minimum, getting an item (with an arbitrary key) and setting an item (with arbitrary data)

An example of the interface could be dict = Dict{}; to make one, dict{do}{"a"} to retrieve an item, and dict{don't}{"b"}{100} to retrieve one, but some creativity is fine

You can assume no global variables already exist other than the ones in the current version

with the ubiltin table, for the earlier cmc: rSNBATWPL, 59 bytes: f=x~rdc{strfor{filt{flip$x}$a~rdc{a}{*}>=0}$k~rdc{k}{+}}$*;

(the one of the products off the sums of the transposed with positive products)

4:10 AM

I think infixing some of that will save a lot of bytes

E.g., couldn't rdc{a}{*} be a rdc*?

Wait no

The k rdc+ one could though

@RadvylfPrograms Actually even in parens this is shorter

oh no we are learning rSNBATWPL

*runs away from the computer*

5:02 AM

oh damn its LYAL again

and we're doing raisin batwaffle this is going to be amazing

lol

> Fun fact: a trailing ' does some really cool stuff to the bracket autocompletion

this is one of the most terrifying things about this language ive read so far

beautiful

why does this hang, im trying to do a collatz conjecture program

@Ginger much better

wait does *= not actually change the value of a variable

what the fuck

@RadvylfPrograms what did you do......

5:15 AM

but if theres spaces it works the other way

i love this language

lmao parser bugs go brerrr

it not a bug, its a feature!

parser feature go weeee

Wow

yeah randomly removing spaces can change whether a program works or hangs infinitely

5:19 AM

@des54321 need a $

i will make a programming language called "pSBATWPL" :P

I just found an old unpublished/incomplete esolang of mine

what is it?

It's one where the syntax looked like this:

/¯\
\_ \
 ¯ /¯
 _/

ooh

i like it

5:31 AM

noodles?

/* flax in c */
typdef struct flax_obj_t {
    void *arr;
    char type;
    int len;
    int is_scalar;
};

typedef struct flax_atom_t {
    int arity;
    flax_obj_t (*call)(int arg_count, flax_obj_t *); /* 1 or 2 arguments */
} flax_atom_t;

Don't do it

/* vyxal in c */
typdef struct vyxal_obj_t {
    void *arr;
    char type;
    int len;
    int is_scalar;
};

/* stack */
static vyxal_obj_t[69];

yesterday, by emanresu A

Speaking of which, @pxeger are you going to tell us what your golflang idea is? (radvylf has)

Aug 15, 2021 at 15:34, by pxeger

@PyGamer0 currently top secret

> Aug 15, 2021

5:46 AM

> Aug 15

huh that date sounds familiar

New Posts

6:06 AM

0

Q: Reverse hex cipher

Characters in strings are sometimes represented as their ASCII hexadecimal codes. Printable characters have two hex digits in their representation. Swapping those digits leads to another character, which will be our output. The table of relevant character codes can be found on Wikipedia. Details ...

6:19 AM

@NewPosts How did my answer to this get so many upvotes lol

it funny

It's a palindrome simply because C and H are their own inverses.

6:44 AM

aaaaand someone downvoted it for some reason lmao

tried porting arnauld's golf

damn you, python, and your sensible behavior on ill-formed inputs

Adám

7:01 AM

@UnrelatedString Me too, but it turns out two chars longer than what I have: ,80⎕DR 4⌽11⎕DR⍪⍞ Try it online!

nice

7:15 AM

@UnrelatedString idk why people downvote good answers....

7:42 AM

@emanresuA yeah i wont make a golflang in c

i will make k in c then a golflang :P

10:02 AM

@PyGamer0 as you can see i am serious about this :P

All I can say is... good luck

10:21 AM

TBH I'd say the hard part of K is train handling

i started that on 10th April

The rest is just simple operators determined by arity (cough jelly cough)

@PyGamer0 ... it's 25 may?

thing is: i have to do it in C

@emanresuA wdym

yes today is 25th of May

@emanresuA initial commit date was in April

^

10:23 AM

@PyGamer0 You hadn't edited it, I thought you meant you were gonna start it on 10 april next year

@emanresuA oh oof

anyways i have done
61 commits so far

Mostly named u, it appears

btw:

in the k tree, 17 hours ago, by Razetime

Announcement: ngn/k is indefinitely frozen from 24 May 2022

@emanresuA the beauty of type overloading

I see

10:25 AM

@emanresuA u for update :P

Golf your language, not your code or commit messages

Don't become Arthur Whitney

Or Radvylf

o_o

I'd say C is a good choice if you want speed, but then you kinda feel obligated to provide said speed :P

ngn/k does some sort of bytecode compilation iirc

But for an array language, dynamically allocating memory like that is gonna be a pain

idk i have never written an array language before

10:27 AM

@emanresuA just use linked lists lol

@PyGamer0 ... flax

@emanresuA is that an array language?

I'd say so

~~its not on the apl wiki~~

Its main datatype is arrays and it provides powerful operations on those arrays and quicks to modify those operations

@PyGamer0 That can be rectified.

aplwiki.com/wiki/Jelly

10:29 AM

i know that exists

Apr 17 at 5:04, by DLosc

> mix of apple and bacon sprinkled with jelly

Apr 17 at 5:03, by PyGamer0

flax looks like a cursed mix of apl and bqn with sprinkles of jelly: +´⍳10

in flax, 17 secs ago, by PyGamer0

Also flax v0.x.x are all beta versions. I might push breaking changes a lot but once I feel that I will not make any major breaking changes then I will release v1.0.0

also owo == 🦉

Once you reach v1.0.0 I'll start trying to compile minor versions to wasm with some duct tape

ok nice

My current approach of manually pulling the files with a script, and writing them into the emscripten filesystem, produces some weird bugs that I'm not sure how to deal with

But if I do someday succeed in getting that done I can just run a certain script to pull the latest version of flax

Well, o/

o/

11:07 AM

@UnrelatedString and i made this Try it online!

Adám

11:27 AM

@PyGamer0 Huh, will he now work on ngn/k2?