Vyxal

@hyper-neutrino Yes

Because flux_string is a list

is this meant to be flux_string[1] += "".join(character)?

No

7:31 AM

also why would character even be a list

Because strings are grouped as lists

Bubbler

Mixed types D:

@hyper-neutrino if anything, it should be += character[0]

`@Hy∧‛NeꜝṘ⋏⇩ ÷… λǐ!`
'`'
'@'
'H'
'y'
'∧'
['Ne', '`']

tf?

do you have a two-char literal syntax that you're not parsing properly

Depends

Which function is that returned from?

7:35 AM

i just inserted a print into group_strings

Where?

at line 157 print(character) (github.com/Vyxal/Vyxal/blob/…)

Oh then it's perfectly fine

It's behaving as expected

oh, okay

@hyper-neutrino it's supposed to parse like that

2 mins ago, by lyxal

@hyper-neutrino if anything, it should be += character[0]

7:37 AM

['`', '@', 'H', 'y', '∧', ['Ne', '`'], 'ꜝ', 'Ṙ', '⋏', '⇩', ' ', '÷', '…', ' ', 'λ', 'ǐ', '!', '`']

this is what the full source is

Where's that from?

inserted print(source) at line 156

(which function)

@hyper-neutrino once again, as expected

← 36 messages moved from The Nineteenth Byte

hm, okay

so what's the expected result then?

Golly, stop making me think everything is on fire

7:39 AM

lol

2 mins ago, by lyxal

@hyper-neutrino if anything, it should be += character[0]

I think A username wants it to print @HyperNeutrino Welcome Back!

but if you do ^ that won't work

2 mins ago, by lyxal

@hyper-neutrino if anything, it should be += character[0]

That's how you fix it

That's what's causing the bug

mfw localhost link

Is that with the fix I suggested?

7:41 AM

yes

Hmm.

I don't have my computer at the moment, so I'm going blind here btw

(mobile debugging ftw)

source = group_two_byte_strings(source)
source = group_strings(source)
source = group_digraphs(source, variables_are_digraphs)

this order feels wrong

It really isn't

i think you need to run group in parallel

two byte strings should block ` from starting a string

but ` should block two byte strings

@hyper-neutrino it does

7:44 AM

wait, no, two-byte-strings has a switch to catch strings

@hyper-neutrino Try it Online!

so why doesn't it work then

@hyper-neutrino wait no it doesn't

Tell me what the return value of group_strings on `This shouldn‛tn‛t work` is

just group_strings?

Ah, sorry

group_strings on the result of two bytes

gs(gtbs(...))

7:48 AM

>>> from vyxal.parser import group_strings as g
>>> from vyxal.parser import group_two_byte_strings as b
>>> b("`This shouldn‛tn‛t work`")
['`', 'T', 'h', 'i', 's', ' ', 's', 'h', 'o', 'u', 'l', 'd', 'n', ['tn', '`'], ['t ', '`'], 'w', 'o', 'r', 'k', '`']
>>> g(_)
[['This shouldntnt work', '`']]

this is with your proposed fix

without it, this would error

Well frick

        temp = components.append([temp + character, "`"]) or ""
        in_string = False

ngl i don't get what this is supposed to be trying to do

@hyper-neutrino I can kind of see a solution to this, but I need to stare at the results for a bit longer

I think I have an idea

okay i guess the important part is

is ‛ supposed to indicate that the next two characters form a literal

Yes

7:50 AM

and is `...` supposed to be a literal

and is ‛`h supposed to be a two-byte literal and not start a string

Yes

@hyper-neutrino yes

and is `...‛...` supposed to insert the ‛` directly into the string

wtf that breaks markdown

Depends on the flags used, but yes

wait which flag changes this

D turns off dictionary compression

7:51 AM

oh you mean that

Yes

well i still consider that inserting the character, it just gets decompressed after

Oh okay

Su

> escaped = components.append(character) or True

why must you hurt me in this way

a) just do two lines

b) escaped = not components.append(character)

math

I don't want to interrupt, but I have a Vyxal question again: Where's the code that translates the Vyxal command (eg. +) to its helper function (eg. add)?

7:54 AM

@math you're not interrupting

math

half

i can just add a check in GTBS such that when it sees a backtick, it disables the two-byte group until it sees another backtick

@hyper-neutrino I'm currently writing a half-fix into a gist

aight

@math remember that non-structures are have the NONE structure type

So if you go to vy_compile, you'll see that the command dictionary is accessed under the if statement that checks if token_name == Structure.NONE

math

7:56 AM

oh right

@hyper-neutrino gist.github.com/Lyxal/1787f447a25e284af16c619ed4ad07bf

That would require a few changes in vy_compile, but I think that would work

why else would type(character) be list

> there's probably a need for an elif for the other string delimiter types because I think there was something funny with strings being parsed inside other structures or something

Honestly, I don't know

opened #148

thoughts on a github bot here?

Perhaps it's time for a third rewrite of the string grouping system

@hyper-neutrino yes

8:02 AM

what's wrong with making GTBS ignore the delimiter if currently in a `-string?

Probably nothing

But I mean, it's kinda clear that there's some confusion around the way strings are handled internally

there are two easy ways to fix this

the first is to just look for GTBS-style things inside strings and unpack them appropriately

the second is to make `-strings stop GTBS

the issue with the former is that if the two-byte-string delimiter is too close to the end, it cuts off the closing backtick and causes the string to not terminate (being the one your gist has)

do you want to maintain that strings can be parsed within strings?

?

like, do you want two-byte-strings to be parsed within normal strings or is that unintended behavior

It's neither really

It's just an intermittent step in the string parsing process

And this means that the same bug will be present in the experimental branch

Because I totally didn't rip half the code for the parser straight from the experimental branch

8:08 AM

hmm. okay

I think I'll have to redo the string parsing system

what do you expect `abc'x` to do? (where ' is the 2b delimiter cuz i'm lazy)

Just insert the '

okay

It being parsed inside the string is relatively arbitrary

8:10 AM

so as long as the end result works that 'a` does not start a string and `abc'` will close a string, it's fine?

Yes

aight. i think this should be reasonable tofix

I kinda want to rewrite it though

I've hit a lack of momentum with the experimental branch

And I'm not quite sure what's happening with the code clean up

So a challenge where I somewhat know what's happen is greatly appreciated

Plus, I can document it properly as I write it

fair enough. i will leave this to you then

CMQ:. What should something like '\'m be parsed as (let ' be the two byte string delimiter)

8:19 AM

i think it should insert the two-byte delimiter and m into a string

\' (literally backslash and delimiter) or 'm (delimiter and m)?

IMO if scanning a two byte literal, \ should escape the next character iff it is \ or '

and likewise, when scanning a normal literal, \ should escape only \ or `

So '\\? > \?

And '\n > newline

@hyper-neutrino ^ and ^^

no, IMO the second is unnecessary because you can just put a newline itself

use regex to parse these /s

But say it was '\t or '\r?

8:23 AM

oh wait

There are no single byte elements for those

hm, then i guess those should be escaped too

And then what about '\t?

Literally \t or tab followed by question mark?

i think tab makes more sense then

8:55 AM

in Sandbox, 13 secs ago, by Vyxal Bot

The repository was starred by hyper-neutrino

\o/ it begins

@HyNerateElevation You should totally change your name to that

@hyper-neutrino Could be nice.

Also, Vylight....

Can I safely fix this or will there be massive merge conflicts?

9:25 AM

Vy light is pretty much separate

You should be fine

ok

We're mostly cleaning the python code

ok

Turns out I just forgot to check if it's in SCC mode.

PyGamer0

9:45 AM

How is the structure separator | parsed?

I don't really understand the current version but I'll try to explain how I did it:

There are three different usecases:
- statements where it's just a simple branch, like while loops and if statements
- statements where it modifies what came before, such as for loops and lambda arities
- lists.

With the first, you can just parse it as its own token, and let the compiler handle it.

The second is a bit more complicated. If you encounter a |, you basically pop tokens while accumulating what's in between until you encounter the original, push what's in between as a token, then push the branch as its own token.

The third can mostly be parsed as its own token, although it requires a little lateral thinking.

10:03 AM

@lyxal ⁺ is broken

10:59 AM

And ↵

That does sound like a combination of issues with the new parser and the clean up

Hm

I'll fix it shortly

I kinda want to just put the 2.4 snapshot online and leave it be...

11:44 AM

That would have to be on a separate pythonanywhere server

Because my current site has already used 350mb of my 500mb storage allowance

And that's just one version

That's fine

Wait 350mb

How

Idk

(Anyway I can make one and push buttons)

Wonderful

@Ausername i have fixed the bug

poggers

11:56 AM

now to make strings great again

using queues

is vectorisation great again?

probably

who knows?

not me

same here

generators though...

12:01 PM

we don't talk about those

Vyxal 2.6:
- Fixed generators
- Removed Herobrine

3

Vyxal 69.420:

- Added ability to make programs use telekensis
- Fixed generators

And now ‛ is broken in compressed strings

'night

@Ausername we know

that's issue #148 on github

12:22 PM

@AaronMiller hello there!

everything is only half on fire this time

yay!

and now it's not on fire

because I used a fire extinguisher on the code

hello @user

hooray for code fire extinguishers!

Hi

and by fire extinguisher, I mean this song

(not a rickroll I promise - Rick isn't exactly that helpful for vyxal work)

12:38 PM

This failed, methinks group_strings or whatever didn't get refactored/renamed prpoerly

@user heh, imagine not checking the most recent commit

lol

this isn't the first time I've been unhappy with pylint: github.com/Vyxal/Vyxal/commit/…

12:57 PM

CMQ: What kind of decorators would be a good idea to add to make writing Vyxal easier?

I feel there should be a stackify decorator to turn a Python function into a command that pops from the stack and pushes back onto it

Turning Generator into a decorator would be great, but time consuming

because you'd have to go through each instance and refactor it

which is why in the experimental branch, I'm using LazyList as a decorator right from the start

Interesting, how does that work?

Actually nvm I'll go check it out myself

@user v

def __call__(self, *args, **kwargs):
    return self

example usage:

def map_every_n(function, lhs, index):
    @LazyList
    def f():
        for pos, element in enumerate(lhs):
            if (pos + 1) % index:
                yield element
            else:
                yield function([element])[-1]

    return f()

Nice

because apparently you can use classes as decorators

1:39 PM

you can also decorate classes

decorators can be any callable (cuz it is just syntactical sugar for calling something, anyway)

i try decrat fns wit generator but nat wrk :(((((((

L

eet wurk nau

much hapi yay

2:24 PM

Fun fact: When using the r flag, the $ command is rendered completely useless. I think this should change. Essentially, when using the r flag, $ should either take its arguments in the normal order, or it should return its arguments unchanged and unswapped, which would mean they are swapped since they are taken in reverse order.

lol

You can probably commit to master directly, that's a pretty reasonable change

Btw does anyone know what first_n does?

probably gets the first N integers satisfying some condition for that one element

2:38 PM

@user I'm just not sure which would be better between the two ways of doing it: The flag doesn't affect that command, or that command acts differently depending on the flag

The flag shouldn't affect that command, I guess. Making it act differently would mean extra swapping and unswapping and swapping

it only takes me five keystrokes to get to the vyxal repo lol

g (rightarrow) v (rightarrow) v

(well technically then i need to hit enter)

@VyxalBot should be pretty operational now for reporting activity on github

It used to take me just one: g, but when I was setting up the interpreter for brainbox, I referenced the grok one a lot, so now g tries to go to there, so it takes two: gi

ah

@hyper-neutrino Noice

2:48 PM

bot will report forks, issues opened/closed/reopened/edited, PRs opened/reopened/closed/merged, commits pushed, and branches created/deleted

if there's anything else y'all want just lmk

3:11 PM

aight i'm disabling the bot here for now so i can test other stuff

Hi @VyxalBot! How do you do?

not very well; the bot is failing to send requests to the backend for whatever reason

also i have a somewhat cursed setup; i have two flask servers, one of them is internal and sends messages when it receives a POST; the other is external and receives webhook activity via a POST and then POSTs to the internal server

and to prevent having to log in / out too much (which hits captcha), when the bot receives a message, it just POSTs it to my ext server which POSTs the reply back xD

I would say you should add a rickrolling command that's only accessible to ROs, but pretty much everyone here is an RO, so... :/

*makes it mod exclusive*

Can't you use the internal server directly?

@AaronMiller lol this room probably has the highest RO:participant ratio

3:21 PM

@user idk this should work fine and idk what best practice is lol

i mean probably to pack everything into one file but then if i change it i need to restart the bot and that hits captcha

3:35 PM

CMQ - what commands should the bot have and what syntax?

I want to add a hyperping command but I am worried about abuse

@VyxalBot lol u good? - Check status of bot

or just non-malicious misuse

@VyxalBot epic gamers - Hyperping all users currently in room

@VyxalBot plz give me teh otput - Runs Vyxal code (inputs and flags optional)

Should it be given RO powers?

@VyxalBot is <code> funny? - Execute <code> as Vyxal code and only return the output if it is 69 or 420 /s

3:42 PM

:bruh:

@user doesn't seem necessary

@hyper-neutrino It's not necessary for half the ROs to have RO powers, so you know somebody is gonna give it RO powers

¯\_(ツ)_/¯ true

aww, it needs 100 rep first. dang it

i think lyxal just like RO'd anyone who sent more than 2 messages here lmao

Feeds

hyper-neutrino has added Vyxal Bot to the list of this room's owners.

3:43 PM

not if I commit mod abuse

3

11

@lyxal btw i hate you for making strings delimited by backticks /s

@lyxal btw i love you for making strings delimited by backticks /s

@hyper-neutrino hehe, fenced code blocks ftw

@user Btw this is a completely serious request, since it is in the spirit of Vyxal

tbh i just hate markdown lmao

3:47 PM

Every day, more people die

nice

Someone must be sacrificed to stop this

Feeds

hyper-neutrino has removed hyper-neutrino from the list of this room's owners.

fixed

i don't actually need to be an ro kekw

Thank you for volunteering

3:49 PM

:58598309 If you're gonna add a rickroll, you can't say that you're gonna add a rickroll or everyone will know. You have to put it in in such a way that no one will notice until they get rickrolled, like when A username added a rickroll to Vyxal.

Please don't insult me by using that link :P

→ 1 message moved to Trash

wdym?

its a perfectly valid example

It's the most commonly used link used to rickroll people

I've memorized it by now

Oh wait you changed it lol