Vyxal

lyxal

01:10

Ginger

:3c

what could he have meant by this

lyxal

try not to make planet name sound alike joke challenge (gone wrong)

it had no right making me laugh as much as it did

(I was looking for symbols and found that)

lyxal

02:02

Codepage almost done

Then it'll be making sure key stuff is there and filling out overloads in planning

user

03:01

@lyxal Huh, didn't know uranus had an arrow sticking out of it

lyxal

05:01

The filled codepage, subject to change

@emanresuA that fork list definitely helped with ideas

fairly certain I did everything on the list if it wasn't already done/planned anyway

github.com/Vyxal/Vyxal/blob/v3-new-3.5-ideas/documentation/… is the best source of what does what

it still needs a ton of overloads added, but it's got the key things that can't be overloads

(e.g. I need to add a lot of (any, fun)* overloads)

emanresu A

Here's the rest of that list although a lot of it is only relevant to v2

lyxal

I feel like some of those things aren't actually things to add :p

emanresu A

"int returns int[]" means a builtin that parses a89sdd5d6s into [89, 5, 6]

lyxal

convolution I think we already have in v2

emanresu A

Okay yeah two of them are just for the memes and I have no idea what I meant by "ffi"

I think I meant 1-byte convolution

lyxal

05:07

@emanresuA Foreign function interface I imagine

emanresu A

I know, I just don't know how it'd be implemented / what specific purposes it would/could serve

Like sure it'd be funny if I could host a webserver with flask via vyxal but it needs a bit more thought

lyxal

sounds like that'd be a standard library thing though

emanresu A

¯\_(ツ)_/¯

lyxal

with (at least in v3) a scala backed implementation

(that's how I was thinking of doing it at least)

emanresu A

"reverse join" is [a, [b1, b2, b3... bn]] -> [a, b1, a, b2, a, b3, a, ..., a, bn, a] which is actually pretty useful

lyxal

05:09

two different stack items or in a list?

emanresu A

Two different stack items

lyxal

you're gonna laugh

j

Try it Online!

emanresu A

Note the a at the start and end

Yes you could do the same thing with ~jpǏ or whatever but it's long

What I want is e.g. 4, [1,2,3] -> [4, 1, 4, 2, 4, 3, 4]

lyxal

Digraph or monograph?

emanresu A

I was planning to make it a digraph since it's already so much shorter than the alternative

Could be bumped up to monograph if it's useful enough

lyxal

05:15

@emanresuA all rotations sounds like a good use case for quantum values. A digraph that returns a value that's all the rotations at once

emanresu A

lol

lyxal

That's actually a serious suggestion and thing I'll probably implement

Because quantum values are a new v3 feature

emanresu A

... I see

Is this a nondet thing?

lyxal

A lightweight version

emanresu A

Interesting

Also on the digraph list: I've used yNY like four times

lyxal

05:17

I haven't quite settled on the implementation, but the current idea is that you can quantumify a list of values and it'll be every value at once

emanresu A

So then you can apply builtins to all of those values and/or choose values satisfying specific properties?

Interesting

lyxal

I was originally thinking a modifier but that becomes tricky when you want to do grouped lambdas

@emanresuA correct (haven't done the filtering part yet)

You can collapse a quantum value to turn it back into a list of possible values

The main inspiration was nekomata winning by like 1 byte on some challenges

Which is very mildly annoying and something that needs to be fixed :p

emanresu A

Vyxal really does seem like the jack-of-all-trades golflang sometimes

lyxal

Depending on how far you expand the quote that's either a good or a bad thing :p

emanresu A

We've got jelly-esque modifiers, functions, osabie/keg-esque structures, canvas because osabie was too golfy, (compressed) regex because retina was too golfy, and now apparently nondet because nekomata was too golfy

lyxal

05:22

And don't forget fractional byte compression because of nibbles :p

emanresu A

Right

lyxal

~~I let everyone else flesh out rough ideas and then refine them into a single language :p~~

emanresu A

I'd say almost the opposite: Vyxal is the mildly janky conglomeration of basically every decent golflang ever created. Most of the time it works pretty well but sometimes it's a bit of a mess

lyxal

I don't see what's wrong with that :p

emanresu A

Never said it was a bad thing

Just that there is a fair amount of lingering jank from trying to combine so many paradigms

lyxal

05:26

Well that's what 3.5 should hopefully reduce

The jank

emanresu A

Fingers crossed :p

Also it's 2024 and we still don't have a 1-byte form of ȯt. add that to my list

lyxal

06:00

@Ginger literate to sbcs button when? :p

lyxal

07:34

@emanresuA out of curiosity, what do you mean by jank?

emanresu A

A few interactions between builtins that just don't work nicely - e.g. the second half of this

The Empty String Photographer

Matrix questions are almost always ungolfy

lyxal

@emanresuA what's included in the second half?

emanresu A

vjvw$j

lyxal

Well the v3.5 solution is a map from stack lambda

Although that doesn't reduce byte count

Why not map sum without-popping: prepend append close-map?

Unless that also doesn't reduce byte count

Wait doesn't surround exist as a digraph?

ø.

Feel like that could be useful somehow

RubenVerg

08:13

@lyxal yay, more things to make glyphs for!

oh, you moved some stuff around too, nice

what's eager map?

RubenVerg

08:31

+1 for new space character constant

lyxal

08:57

@RubenVerg map, but not lazily evaluated

So you can do things like print or things with side effects without lazy evaluation messing things up

It can sometimes save a byte without needing a flag

If flagless is something you're into

@RubenVerg I needed a join on spaces glyph and I decided that the old space constant symbol would probably be the best

So I needed a new space glyph

:p

RubenVerg

@lyxal hmm, is there a force eager element or do i have to do map-eager id

well actually there's no id either so it'd be like map-eager lambda close or something

lyxal

09:17

@RubenVerg map-eager close

No need for the inner lambda

But typically speaking you don't need to force evaluate a single value as much as you need to force evaluation when mapping (in terms of golf)

Meaning a force evaluation element would probably be a digraph anyway

Thinking of repurposing the escape hatch modifier, because I can't see it having a use outside of things that should probably be digraphs anyway

No clue what I'd make it do

RubenVerg

10:14

@lyxal oh it's a syntax thing not a modifier

lyxal

11:16

@RubenVerg yeah

Logical and with dot in v2

Ginger

12:41

@lyxal wooprs

Ginger

13:41

@lyxal how did you want it to be implemented?

I've added a button that converts all three fields to sbcs

lyxal

Just code

Ginger

but uh

header/code/footer are concatenated together

they can't mix literate and non-literate code

lyxal

Doesn't need to replace code

Just print to terminal

Ginger

ohhhhhhhhhhh

alright, one sec

lyxal

Yeah as a preview

Ginger

13:44

should it clear the terminal?

lyxal

Sure

Ginger

ok

lyxal

It's nice to have it as a sort of preview to check what you're writing is on the right track

Ginger

ig it'll be disabled while code is running

lyxal

Yeah

Well...

Ginger

13:46

I can probably also make it queue the message to be displayed when the code finishes?

lyxal

I sometimes click it mid program execution

Because I want to see sbcs and then have output

@Ginger I mean, disable it during execution if it's easier

Ginger

alright

lyxal

I'm just thinking about how I use it lol

Ginger

would be cool if the output had syntax highlighting...

lyxal

Ginger

13:54

eh, that can be a later thing

lyxal

Ginger, we have to talk

Ginger

what about :3c

lyxal

I'm very sorry to say this

but we don't make bikes

I'm afraid the handlebars are useless

Ginger

seems like a missed opportunity

lyxal

it's not economically viable

programming languages typically don't turn into bikes very well

when was the last time you saw anyone riding Java?

@Ginger about the discord link on the website apparently

it's...uh...invalid

Ginger

13:58

me when I somehow failed to set the runner state correctly when the worker spawns when I set it up half a year ago

@lyxal ugh

I'll fix it

lyxal

use the one in the room description

it seems to work

Ginger

button added and invite fixed

also fixed permalink changes spamming the browser history

lyxal

where's the button?

I would like to click it

Ginger

it only appears if literate mode is on

lyxal

it is

Ginger

14:04

uh.

that's broken

what's the commit hash in the settings dialog

lyxal

788d171

Ginger

okay, firefox is probably doing something weird

what's the console say

lyxal

Ginger

whar

hard refresh?

lyxal

found it

Ginger

14:06

webpack should be including cache-busters on the scripts

lyxal

that's not the kind of button I was expecting

Ginger

evidently, it is not

lyxal

it appeared after I cleared cache

Ginger

yep

fixing that

lyxal

why must enter do autocomplete?

Ginger

14:08

that's how CM does it (and how vscode does it)

lyxal

very inconvenient when doing literate mode

e.g. I type:

record Name =>

and press enter, and it changes it to assign

Ginger

I'll see if I can add a setting to make tab do autocomplete

lyxal

Very happy to have my favourite button back

slight problem when running a program after clicking:

it seems to keep the last line in the output

Ginger

hmm

Ginger

14:34

I have changed some webpack settings, it's generating a lot more files now :p

Ginger

14:55

@lyxal should be fixed now

assuming my webpack changes didn't blow everything up

lyxal

works good

(late reply because bug fixing inputs :p)

speaking of which, is it fine if I make a new release?

Ginger

yep

lyxal

That's kinda sorta another major bug fixed

cool

Ginger

(hopefully)

lyxal

Vyxal 3.4.7

2

Ginger

15:05

hopefully the webpack changes didn't bork the archive

lyxal

fun how that bug has been around since launch last year :p

seems to be doing fine

oop uh nope

Ginger

mfw

nope, it's github's fault

serving an old version of versions.json

lyxal

well that's a good thing at least :p

it's not an us problem

which is the best kind of problem

anyhow before I break anything else I'm going for the night :p

Ginger

hey uh why is there other stuff in the pages output

github.com/Vyxal/versions/tree/main/v3.4.7/vyxal

lyxal

@Ginger which stuff?

Ginger

15:11

everything that isn't Vyxal and theseus.json

it looks like an old copy of the legacy website?

lyxal

seems to be

Ginger

it's copied from github.com/Vyxal/Vyxal/tree/version-3/pages

lyxal

probably the workflow needs updating

in main repo

tomororw

ir, well, todya

Ginger

oky

lyxal

toayfa

toafy

today

given its' 1am

goingr noww

o/

Ginger

15:13

\o

user

21:29

@lyxal Ooh wait is this like Curry? Non-determinism?

Excited

lyxal

23:36

@user I don't know curry, but if it has a feature where values can be made to be multiple different values at once, then yes

Transcript for