Vyxal

Vyxal Bot

00:03

Lyxal approved on PR #4 (Vyxal/vyxapedia)

hyper-neutrino

00:14

so I have an urge to create a language but also not the motivation to fully carry through an entire new programming language

so instead I will try to start contributing to vyxal :P

lyxal

add the ideas to version 3 planning :p

Also, #474

hyper-neutrino

my ideas are just a mix of python and JS (but like, actually not cursed, but more so mixing the parts I like about both of them) so

probably nothing I can add to vyxal ideas :P

anyway lemme see if I can recall how to use the dictionary

I think I have it written out in my vyxapedia tutorial

oh I do. I'm unironically going to use a tutorial that I myself wrote to learn something ¯\_(ツ)_/¯

Unrelated String

hell yeah

hyper-neutrino

hmm

Unrelated String

if i ever make a praclang it'll probably be something functional and homoiconic with python interoperability

hyper-neutrino

00:23

i just did compression with a trivial DP right? IIRC

lyxal

Sounds about right

Unrelated String

or whatever i started designing to do advent of code in

think so yeah

hyper-neutrino

i think i was gonna make a trie and then realized I was thinking about it the wrong way lol

Unrelated String

lol

hyper-neutrino

also what is the short dictionary even

lyxal

00:24

When you have an Unicode char without a second one

Unrelated String

i think jelly also has a short versus long dictionary thing going on so it might be the same as there

hyper-neutrino

@lyxal ah

Unrelated String

ah

hyper-neutrino

@UnrelatedString jelly's compression thing is completely different lol

Unrelated String

oh yeah that

hyper-neutrino

00:25

it can also be implemented with DP tho lol, which is how I did it for the compendium's compressor

lyxal

2

A: Tips for Golfing in Vyxal

Utilise the Short Dictionary Note: This is a 2.6.x+ feature only Newer versions of Vyxal have a neat little feature where 1 character SCCs are indexed into a special "short" dictionary of overlyspecialised "words". Here's a list of all the current entries: \d+ λ -?\d+ ƛ \d+\.\d+ ¬ -?\d+(\.\d*)? ∧...

hyper-neutrino

ah

yeah this should be fairly easy

well, if I remembered how to do it :P

Vyxal Bot

hyper-neutrino created branch Vyxal/hyper/compressor-short-dictionary

hyper-neutrino

what version should I be running vyxal on

lyxal

python 3.9+ iirc

> Requires: Python >=3.9, <3.11

from the PYPI page

hyper-neutrino

00:40

okay thanks

uh, no utility for the small dictionary? like there's dictionary.word_index which automatically formats it as the encoded pair of chars

is small dict just a list lol

lyxal

yes

it's just a list

hyper-neutrino

hmm

okay time to do some scuffed stuff then ¯\_(ツ)_/¯

oh wait vyxal.encoding.compression means I don't have to generate the set of non-ascii codepage chars myself

poggers

Vyxal Bot

Steffan153 created branch Jyxal/factorial

Steffan153 opened PR #8 (Vyxal/Jyxal) (Vyxal:factorial → Vyxal:master): Add factorial/sentence case element

hyper-neutrino

00:56

@lyxal does dictionary.max_word_len account for the short dictionary

lyxal

no

hyper-neutrino

hmm

lyxal

for index, word in enumerate(contents):
    lookup[word] = index
    max_word_len = max(max_word_len, len(word))

hyper-neutrino

ah

okay thanks

lyxal

that doesn't enumerate over the short

it only goes over the long

hyper-neutrino

00:58

alright cool. i will use a separate method for scanning the small dict cuz the inefficiency from just using the trivial approach is probably significant enough that I should actually optimize

since the max len is 18 but the short dict goes up to 113

lyxal

gotta love that website regex

((www\.|(http|https|ftp|news|file)+\:\/\/)[_.a-z0-9-]+\.[a-z0-9\/_:@=.+?,##%&~-]*[^.|\'|\# |!|$|?|,| |>|<|;|$])

Steffan

lol

big fat regex

though i dont see it being used in vyxal

lyxal

there was this site with like 6 or so "important regexes you should know"

and it was one of them

Steffan

but who needs to check that when code golfing

lyxal

honestly it could probably be removed

Steffan

01:01

max_word_len is gonna go way up

emanresu A

Honestly

Stuff in the short dictionary should be usable as part of something else

For example, the deadfish stuff and the website regex should probably be constants

But [a-z] can be used as part of something else

Steffan

also like ₁ƛ₍₃₅kF½*∑∴,. that code doesn't even work right in the latest v of vyxal

why in the world is isdo in there. what's isdo

lyxal

deadfish

Steffan

what are the html tags for

lyxal

in case you want to want to render HTML

Steffan

01:09

i can't imagine needing <kbd> and </kbd> tho

hyper-neutrino

oh this is more annoying than I imagined

because of how compression works short and long things can't be put back-to-back

also @lyxal how does it work if i have an ascii char between two non-ascii chars

is that two short forms or does it get merged to a long form like old vyxal does

lyxal

two short forms

hyper-neutrino

poggers

okay, this is still easy by just using two DP arrays

Steffan

!!/issue summate does not work on negative numbers it just returns its input bug difficulty: easy

Vyxal Bot

Vyxal-Bot opened issue #927 in Vyxal/Vyxal: summate does not work on negative numbers

hyper-neutrino

01:19

what is it meant to do

digit sum but negative?

Steffan

i suppose just ignore the minus

idk, maybe it should negate the result

i wouldn't expect it to return the input tho

lyxal

it's "returing" the input because the - at the start makes it use string concatenation

it's effectively the same as summing a string

!!/run 23113N ƒ+

Vyxal Bot

@lyxal -23113

lyxal

def vy_sum(lhs, ctx=None):
    """Element ∑
    (any) -> reduce a by addition
    """
    return foldl(add, iterable(lhs, ctx=ctx), ctx=ctx)

so you'd have to change how addition works with strings and numbers

Steffan

...or just take absolute value.

!!/run 12.56 ∑

Vyxal Bot

01:22

@Steffan 3.56

Steffan

that's also a bug ig

lyxal

not really

Steffan

it depends on what you'd expect

hyper-neutrino

hm

Steffan

it does 1 + 2 + .56

lyxal

01:23

that's how foldl works

hyper-neutrino

i'd say potentially surprising behavior but given the implementation, expected and valid

a bit weird and potentially not useful though

Steffan

but the question is, does 3 + 5 +6 make more sense

imo it does

but idk

flags?

lyxal

@lyxal just realised in pasting this here, I'd accidentally ctrl-x instead of just ctrl-c

Steffan

02:02

lol

Seggan

02:14

night yall

Vyxal Bot

02:31

Seggan approved on PR #8 (Vyxal/Jyxal): "Yeah... sorry, but you could do a lot better"

Seggan requested changes on PR #8 (Vyxal/Jyxal): "Woops hit wrong button"

Steffan

Lists dont work at all in jyxal lol

Unexpected error while computing stack sizes:
Class = [jyxal/Main]
Method = [listInit$0()Ljava/lang/Object;]
Exception = [java.lang.IllegalArgumentException] (Stack size becomes negative after instruction [0] invokestatic #27 = Methodref(io/github/seggan/jyxal/runtime/list/JyxalList.create(Ljava/util/Collection;)Lio/github/seggan/jyxal/runtime/list/JyxalList;) in [jyxal/Main.listInit$0()Ljava/lang/Object;])
Unexpected error while editing code:
Class = [jyxal/Main]
Method = [listInit$0()Ljava/lang/Object;]

hyper-neutrino

02:59

alright so my compressor passed 100,000 random tests for correctness (i believe it should be optimal but i can't test that easily) so I'll PR it now

Vyxal Bot

hyper-neutrino opened PR #928 (Vyxal/Vyxal) (Vyxal:hyper/compressor-short-dictionary → Vyxal:main): Add short dictionary into optimal compression

hyper-neutrino

uh, is this the right branch to PR into

Vyxal Bot

Lyxal approved on PR #928 (Vyxal/Vyxal)

hyper-neutrino

okay I will take that as a yes

i failed a test? wack

how did it fail the black test; I literally have black format on save

lyxal

Format to line length 80?

Vyxal Bot

03:04

chunkybanana approved on PR #928 (Vyxal/Vyxal)

emanresu A

Waiit...

hyper-neutrino

i do

lyxal

Oh

Oh I see

hyper-neutrino

also damn we're approaching 1000 items

emanresu A

While we're at it, does this escape backslashes and backticks in the result?

hyper-neutrino

03:05

doubt

emanresu A

@hyper-neutrino in what?

lyxal

You're not using the latest version of black are you.

hyper-neutrino

uhh

hm

@emanresuA github "things"

like the latest PR is #928

lyxal

They made a change that removes spaces between ** and other things

And that version is what the ci uses

hyper-neutrino

oh... that explains why my editor changed those back

lyxal

03:06

You need to pip upgrade black

hyper-neutrino

install --upgrade?

lyxal

That one

hyper-neutrino

hm, i did that but atom doesn't autoformat to the new version

Vyxal Bot

Lyxal merged PR #928 (Vyxal/Vyxal) (Vyxal:hyper/compressor-short-dictionary → Vyxal:main): Add short dictionary into optimal compression

Lyxal deleted branch Vyxal/hyper/compressor-short-dictionary

hyper-neutrino

oh I was going to fix the format issue first but aight

lyxal

03:09

CI fixed it

So even if you submit something not nice it does formatting for you

github.com/Vyxal/Vyxal/pull/928/commits/…

Pxeger set that up for us

hyper-neutrino

oh that's really nice

lyxal

It also regenerates all the documentation from yaml if you change the yaml

Including test cases

hyper-neutrino

i need to figure out how to update atom tho so i don't keep breaking formatting and then having ci auto-fix it

lyxal

There'd be a setting somewhere to change the version it uses right?

hyper-neutrino

i'd assume so??

Vyxal Bot

03:14

Steffan153 closed issue #474 in Vyxal/Vyxal: Add the short dictionary to the compression algorithm

lyxal

Anyhow, just in case 474 didn't kill your urge for language design, I've another thing that I'd like help with

hyper-neutrino

oh i'm stupid

i ran pip install in WSL

but of course the atom python-black config points to an executable that is within windows so i need to run it in command prompt as well :P

lyxal

👏

hyper-neutrino

@VyxalBot wait, shouldn't there be a way to auto-close when merging or smth

or does that require extra setup we don't have

lyxal

There is

Closing keywords

Github manages it for you

hyper-neutrino

03:15

ah

where can I read a tutorial on how to use all these fancy things I don't know how to do :P

lyxal

If you include something like closes #474 or fixes #474 it'll link it for you

hyper-neutrino

oh poggers

lyxal

github.blog/2013-01-22-closing-issues-via-commit-messages

hyper-neutrino

oh thanks

lyxal

145

Q: Closing multiple issues in Github with a commit message

I know you can close issues by putting closes #1 or fixes #1 in the commit message. What is the best way to close multiple issues in the same commit message? Also, it appears that using fixes rather than closes doesn't create a link from the commit message to the issue, but it does close the iss...

hyper-neutrino

03:17

@lyxal anyway, sure

lyxal

As well

@hyper-neutrino version 3 element planning

hyper-neutrino

oh v3?? huh

lyxal

I'm having difficulties choosing an appropriate element set

hyper-neutrino

ah

do you have a list drafted of stuff you're considering that you need narrowed down

lyxal

What I have so far is in the version 3 branch

Anyhow, since you have much experience with competitive programming and languages like jelly, I think you're the perfect vyxaler for the job

@hyper-neutrino github.com/Vyxal/Vyxal/tree/v3-planning-documents

hyper-neutrino

03:19

oh we're pruning flags?

lyxal

Some of them, yes

hyper-neutrino

also is v3 still python or are we moving languages

lyxal

github.com/Vyxal/VyxalS

Scala migration is a real thing

user has already made headway in regards to the foundations

hyper-neutrino

oh, okay. but python will still be the language for the primary version?

or are we doing that for v3

lyxal

python v2

Scala v3

hyper-neutrino

03:21

hmm

anyway i don't know shit about scala but that doesn't matter for now i'll just take a look at the element list

lyxal

I don't know much about scala either lol

But it should be easy to use

hyper-neutrino

something I could do is just look through all existing elements and rate them all (in terms of usefulness, versatility, etc)

lyxal

That'd be great

hyper-neutrino

or should I look through the list on the v3 branch?

lyxal

Existing would be better

hyper-neutrino

03:23

alright 👍

lyxal

Combined with the corpus that should give an indication of what to keep and what to drop

hyper-neutrino

what's the best file to look through

lyxal

elements.yaml

It gives description + an idea of what it does with raw data

hyper-neutrino

are .txt and .md fine as well or do they have the possibility to be out of date

or is that the thing you said the auto-regeneration covers

lyxal

Those are autogenerated

The yaml is the only thing manually changed

Everything else is automagic

hyper-neutrino

03:25

ah, okay, thanks. the md is easier to read for me so i'll look at that

lyxal

Sure

Vyxal Bot

hyper-neutrino created branch Vyxal/hyper/element-evaluations

user

I’m kinda sad the md is no longer a huge table :(

Guess it’s better this way tho

lyxal

It's easier this way lol

user

Yeah

lyxal

03:29

Less thinking for me about how to markdown table

Because I do not markdown table very well

user

Getting the escaping to work properly was hell

I think the table was my first contribution to vyxal, sad to see it go :(

lyxal

Hey so @user are there testing work flows for scala like there are python?

user

Ye

Check out the vyxals repo

It’s already there (hopefully not failing lol)

lyxal

I see them

Very excellent

user

Oh there’s a pr by steffan, didn’t see it

Steffan

03:45

u didnt? lol

i've been working on jyxal now too

also there's two PRs (by lyxal) waiting in the main repo

user

Ah I’ll review a couple tomorrow

Just took an ap today so i can coast until the end of the school year

Steffan

nice

hyper-neutrino

@lyxal are you just considering freeing up single-byters or do you also want to prune some digraphs

Steffan

copilot does not understand java...

it does a terrible job

lyxal

@hyper-neutrino Digraph pruning would be good

hyper-neutrino

03:49

okay

Steffan

there's plenty of space with digraphs tho

we should also look at the um, what's the name

corpus

we should look at the corpus

lyxal

@Steffan But some could become a single byte

Steffan

yes

github.com/Vyxal/Vyxal/issues/14

mainly look at the lexer one

several common ones

lyxal

29 mins ago, by lyxal

Combined with the corpus that should give an indication of what to keep and what to drop

Steffan

didn't see that sorry

i didnt' look through the whole chat history

lyxal

03:53

No worries

Steffan

!!/'night

Vyxal Bot

@Steffan o/

lyxal

o/

hyper-neutrino

05:32

@lyxal github.com/Vyxal/Vyxal/blob/hyper/element-evaluations/documents/…

can't say I thoroughly read through all of the digraphs cuz there are just way too many of them for me to stay focused enough to give a good evaluation anyway

a lot of them seem like they are probably not used particularly often but also i'm not sure if you pruned them we could necessarily make better use of the space

user

05:46

@hyper-neutrino i think i agree with pretty much all of that

I’d like to keep ranges and triplicate around

Some of the digraphs aren’t that useful but hey, they’re digraphs so whatever

Why does E also evaluate as Python?

I thought the cross evaluated as Python and E was supposed to eval as Vyxal

Vyxal Bot

chunkybanana opened issue #929 in Vyxal/Vyxal: `ẇ` is still cursed

chunkybanana opened issue #930 in Vyxal/Vyxal: `øD` doesn't escape backticks / backslashes

emanresu A

@lyxal My custon corpus should be good

@hyper-neutrino I think × should stay - it's often useful to have a 1-byte non-space string in general. Whether it should be changed to something else, I don't know, but it's been used 22 times which is way above average

₂ can probably stay, and ₃ and ₅ have useful string overlaods

ǒ's a weird one that can probably go

Base compressed strings are useful for certain purposes

Ranges and triplicates are necessary

(it's always fun doing :D when I need to quadruplicate lol)

max/min by tail should've died a long time ago

They're literally only useful with Ċ, and we can add digraphs for those purposes.

Space (58) is definitely necessary, newline (9) is nice to have. Register is very useful (34 / 50), global array is quite useful (11/3/25).

We can leave digraph pruning for another time

Honestly, I've suggested a bunch of digraphs that are very niche and quite unlikely to be used in golfing (graph theory for example)

They do have uses in golfing though

hyper-neutrino

06:35

as long as we're not short on digraph slots we can honestly let some pretty absurd ones stick around

keeping pointless one-byters is bad though

emanresu A

Yeah

full commentary

lyxal

@hyper-neutrino apart from the asterisk literal (well, non-whitespace non-empty literal) and the is even, I think that's fair enough

hyper-neutrino

!!/run 0 2 τ

Vyxal Bot

[@hyper-neutrino: 61067844]
STDERR:
Traceback (most recent call last):
  File "/home/Vyxal/mysite/vyxal/main.py", line 164, in execute_vyxal
    exec(code, locals() | globals())
  File "<string>", line 5, in <module>
  File "/home/Vyxal/mysite/vyxal/elements.py", line 4281, in to_base
    maximal_exponent = int(log_mold_multi(lhs, len(rhs), ctx))
  File "/home/Vyxal/mysite/vyxal/elements.py", line 2386, in log_mold_multi
    return {
  File "/home/Vyxal/mysite/vyxal/elements.py", line 2387, in <lambda>

lyxal

@emanresuA they were originally intended to be used with zipmaps when zip maps were a custom data type

emanresu A

06:44

... ಠ_ಠ When was this

lyxal

Way back

Early days and months

The 0.x releases

hyper-neutrino

@VyxalBot @lyxal can I just edit the make-to-base-work-with-negative-numbers branch to patch this, since your branch also doesn't work with base -1 which can be handled with a special-case

lyxal

Sure

hyper-neutrino

aight

lyxal

So long as it all works out merge conflict wise I'm good with that

@hyper-neutrino can't believe adnan wrote non-functional code

emanresu A

06:47

Lol

hyper-neutrino

wait why is τ running line 2387

isn't that •

oh wait no to_base called it

weird

lyxal

That's a comment though

tf?

hyper-neutrino

should τ on 0 return [] or [0]

and what is intended behavior for base=0, and should n=0 base=0 be a special case

lyxal

Iirc 05ab1e just returns the input

Which is what vyxal should do

hyper-neutrino

ah so 0 τ is a no-op?

and yeah should base conversion of n=0 be [] or [0] and what about n=base=0, would that follow whatever you think makes sense for n=0, or just return the 0 since base=0

lyxal

07:00

I would say [0] and base=0 always return 0

hyper-neutrino

wait when you say base=0 always returns 0 you mean regardless of n right?

oh also since the rhs can be a non-number

i assume for n=0 it should just return [base[0]]?

and what about base=[]

emanresu A

More commentary

lyxal

07:31

@hyper-neutrino for edge cases I say return 0

Regardless of the case

Either that or error

hyper-neutrino

hmm, okay

lyxal

07:49

Consistency of unexpected or non-practical edge cases is a good idea

Just like how there's consistency among github webhooks :p

Anyhow, as I know you're going to sleep soon, @hyper-neutrino I sure hope your urge to esolang design sticks around for a while... I'll take all the help I can get :p

And @emanresuA and @Steffan I greatly appreciate your feedback and commentary too!

lyxal

11:49

man looking at other repositories deving on main / committing straight to main looks so weird

lyxal

12:10

!!/issue Sandbox Does bot allow labels properly? I noticed a tiny discrepancy between github docs and the bot code. So this is a test test beans ostrich

huh

interesting

user

@emanresuA wait, why is space necessary?

lyxal

@hyper-neutrino for the github stuff, are you using octokit or something else?

Vyxal Bot

12:23

Lyxal forked Vyxal/SE-Bot into Lyxal/SE-Bot

lyxal

nvm I figured it out

gonna make some juicy changes

Seggan

12:48

@Steffan what was your input code and debug.log? it shouldnt spit such errors out anymore

lyxal

13:02

there's a ever so slight chance i messed something up there

user

Over here, peek() will return null if you call it immediately without pushing anything first. Is it expected that the Dequeue will be empty when you instantiate CallStack?

Seggan

yes

user

If so, you probably don't want to take it as a parameter, but if not, you should probably check if it's null or not

Yeah I think you should just instantiate it inside the body of the class itself then

Seggan

im only taking the delegated stack as a param?

user

Having it be a constructor parameter implies that you should be allowed to pass in whatever you want

Seggan

13:07

im confused

user

Like, if I do CallStack(myDequeueContainingStuffAlready) and then call peek on that immediately, top'll be null and there'll be an AssertionException or something

Seggan

ah

ye ig ure right

user

Hmm, you won't be able to use delegation if you don't make it a constructor parameter

You could make the primary constructor private instead and have a secondary no-arg constructor

Seggan

im just collapsing it into an anonymous object

user

Good idea

For methods like fun isDouble(): Boolean = value.contains('.'), do you want to instead make them vals with custom getters?

So you can do num.isDouble without parens

Seggan

13:13

thatd be fine

user

pog

Seggan

im still in love with kotlins property syntax ever since i wrote

private val mv: MyxalMethod
        get() = callStack.peek()

user

It's definitely one of my favorite things about it, especially since you can delegate properties and have extension properties

Is it no longer possible to have all the builtins return Anys? Kinda sad to see the explicit stack.pushes

Seggan

its only for the ones that do stack manipulation

since it was kinda hard with the new abstract compiler to make exceptions for em

user

Ah ok

hyper-neutrino