The Nineteenth Byte

caird coinheringaahin g

00:08

Well, I have officially been on CGCC for 6 years \o/

Unrelated String

🥳

congrats

Neil

well, I might as well join the bandwagon and mention my 9th year as of the end of last month

caird coinheringaahin g

Congrats!

That is 75% of the site's lifetime :P

Seggan

00:25

@user *jvm bytecode, not jvm

@user challenge accepted /s

@user yeah it is

Unrelated String

oh yeah i missed my 4 year anniversary a month ago :P

user

00:37

@Seggan Why the "/s"? This would actually be pretty cool, even if it's not particularly useful or practical

Seggan

00:59

yes, it would, but i dont have the time ~~or hard drive space~~ for it

OldSandboxPosts

Sandbox posts last active a week ago: The dating game king-of-the-hill

user

01:48

@cairdcoinheringaahing ok boomer

caird coinheringaahin g

@user Quiet, child

user

/srs Congrats to you and Neil, I hope y'all stay on

Also, 6 and 9, nice

emanresu A

I only have 2 years

user

To live?!

emanresu A

Yes totally

lyxal

01:50

@emanresuA you get to escape the life sentence of code golf addiction :p

I've served 5 years of my sentence thusfar

xkcd says I'll be serving at least 5 more :p

in The Sand Trap, Mar 8 at 0:51, by lyxal

see?

user

Have some hope, perhaps that's your grave and someone put a flag on top of your dead body :P

3

lyxal

That's a golf course, you dingus

It means I'll still be golfing in 5 years

user

@lyxal Yes. Perhaps you dropped dead of your love for golf while playing golf, and the in your honor, someone buried you right there and made a new golf hole right above your body

A Nineteenth Hole

lyxal

@user you seem to think that the forces that keep us here would let me off that easily

New Posts

03:53

0

Q: Write the smallest possible code in x86-64 to implement the "<=>" operator

Languages, such as C++, have introduced a new comparison operator: <=>. This operator returns -1, 0, or 1 when comparing two items together: -1 means the left handside is smaller, 0 when both values are equal 1 means the right handside is smaller. Here is what I have been coming up with in x86-...

mousetail

06:03

@user This seems like it would be usefull

user

I feel like a Python/Java library chock full of builtins would probably suffice

It would certainly be cool, though

If anyone here finds such a project or wants to start work on something like that, ping me please

doug

Is this a good spot to ask questions about Jelly? If not can someone refer me to one? Thsnks.

user

This is as good a place as any

Jelly Hypertraining

Practice your Jelly :) Rules and stuff are here: golfingsucces...

idk how active that room is

doug

06:18

Thanks. I’ll just post it here for starters. I’m making my way through the tutorial and trying to understand the example in Multi-chain links: github.com/DennisMitchell/jellylanguage/wiki/……

user

Cð+×µH?

doug

Ha! I think I get it. After C is processed in the chain C+×H then only + is processed next because it’s a monadic chain. I thought it might pull off both +× because I was looking at the wrong chaining rules when referencing the rules above.

lyxal

@doug it depends on how many arguments you give it

doug

Do those chain separators act as brackets or can you use them arbitrarily?

In this example it appears to be monadic.

lyxal

@doug iirc it takes everything until the next separator (or link end, whichever comes first in the program) and makes it a single link evaluated with the arity rules of the separator

It's as if you wrote [C, {arity: 2, code: +×}, H]

That +× is treated as if it was a single dyad in parsing the larger link, and the +× is parsed with dyadic chaining rules when evaluated

doug

06:33

I think the point of that section is that you need the chain separators to get that behavior in a monadic setting. In a dyadic setting I think it might parse as you describe it. That being said, I started reading the tutorial only a couple of hours ago.

lyxal

Arity of larger links is determined by how many arguments you call it with

If you call it as link(x), it parses monadically. If you call it as link(x, y) it parses dyadically

doug

Right. In this example it’s implied this is a monadic chain.

lyxal

Correct. The chain separator is needed because of the context of the link in the example

doug

Cool. I think I’m up to speed.

Thank to you and @user.

doug

06:51

For any newbies reading this: This has a better description of how the chain separators work: github.com/DennisMitchell/jellylanguage/wiki/Syntax. Long story short, they’re not interchangeable.

user

@hyper-neutrino Didn't you have a site with a better Jelly tutorial?

hyper-neutrino

@user jht.hyper-neutrino.xyz

for chain separators, jht.hyper-neutrino.xyz/beginners#separators

Unrelated String

@doug although they can be used for recursive structure with quicks, they fundamentally just create a single additional level of divisions

The Thonnu

07:09

Just got -60 from "voting corrected" :(

lyxal

That's silly those were legitimate upvotes

The Thonnu

Aiden Chow

@TheThonnu rip dw, ive gotten similar before as well

lyxal

First time I've seen votes on fresh answers reversed

The Thonnu

I don't think it takes into account how old the answer is

It's just if one user votes on another user's posts a lot of times in a short space of time it will reverse all the votes

lyxal

07:14

Let's see if they reverse them this time

The Thonnu

Thx

lyxal

Frickers undid my votes and made me waste my vote quota yesterday

Like literally made the button gray again despite me having clicked it when you answered

It's not vote fraud because I would have upvoted the answers anyway. That's silly that they got reverted just because it was all at once instead of spread out over several days

The Thonnu

It's an automated process: it doesn't understand that

It's fine though, I don't really mind if I lose 60 rep

lyxal

@TheThonnu I know, but it's silly that the automated system punishes voting on a rapid-fire series of quality answers

The Thonnu

I guess, maybe it would be better if it went to the mods first before it reversed the votes

Then the mods could check if it was actually vote fraud

lyxal

07:21

Also, if it decides not to reverse the re-votes, then that's just as silly

Spreading votes between other answers shouldn't be enough to throw it off

doug

@hyper-neutrino Thanks!

user

08:33

@mousetail one problem with a language agnostic vm is implementing eval

now im really tempted to write a rust golfing library with python bindings

mousetail

08:56

800 KB :/

mousetail

09:10

The color grey is 200 bytes now wtf

lyxal

Brainfuck encoded colour code

mousetail

Must be

lyxal

That, or it's got like 38 shades of gray defined in the file

mousetail

Yea there seem to be 13 shades defined, but then I'm confused why there needs to be both blue and lightBlue if each contain multiple shades

lyxal

Wait actually it's multiple shades?

mousetail

09:15

const grey = {
  50: '#fafafa',
  100: '#f5f5f5',
  200: '#eeeeee',
  300: '#e0e0e0',
  400: '#bdbdbd',
  500: '#9e9e9e',
  600: '#757575',
  700: '#616161',
  800: '#424242',
  900: '#212121',
  A100: '#f5f5f5',
  A200: '#eeeeee',
  A400: '#bdbdbd',
  A700: '#616161'
};
export default grey;

I looked it up

lyxal

That was supposed to be a joke. The hell kinda library needs to give you that many options?

And then why not just use css

mousetail

Yea this library is kinda terrible, gonna try to find a better one. 50 KB to load 2 icons that are only 200 bytes each

It only allows specifying named colors which are hard to make fit my theme

and only named sizes which is also a problem because none of them match exactly what I waant

lyxal

@mousetail The library or something else?

@mousetail named sizes?

mousetail

Yea, when you load a icon you can specify the size to be the string "small", "medium", or "large". That's it

lyxal

Holy hell

mousetail

09:20

It depends on some global theming package that allows you to exactly specify what sizes those mean but I don''t want to setup a entire theming context for 2 icons

lyxal

Can you use css manually? Or does it have to be dynamically loaded?

mousetail

Maybe, the icon generated gets like 6 css classes applied. I could try to find the one that corresponds to the size then use !important

lyxal

So the icon is dynamically generated?

mousetail

Yea, I dynamically choose which SVG to draw. The SVGs are from the library

lyxal

And there's no way to do it with vanilla js?

mousetail

09:26

Yes there is but it's convenient to have a library of icons available without needing to download them one by one

I might just do it manually though because this library is worse than nothing

(This is React BTW)

emanresu A

10:24

~~rewrite it in rust~~

mousetail

I would like to, I tried Yew before but it seems a bit too immature still

Also even the smallest webassembly builds are orders of magnitude larger than even bloated JS bundle sizes

mousetail

10:53

Q How do I most concisely create a array of pointers to array literals in C. Like [&[1,2],&[3,4,6]] in rust?

Unrelated String

11:03

iirc c doesn't have a concept of "array literals"

you just have array initializers

but in any case... i think just a normal multidimensional array should work? decay could be weird depending on what you need it for but

emanresu A

Don't you need something special for non-rectangular arrays?

Unrelated String

oh yeah good point

in that case, you probably just want to pad it to rectangular, unless it's really ragged lol

okay it looks like you don't even have to pad it manually

tio.run/##S9ZNzknMS///XzkzLzmnNCVVwaa4JCUzXy/…

anyways the fucky thorny bit here is that you don't have a real array of pointers, but dereferencing the array still gives you an array that decays to a pointer

mousetail

11:39

@UnrelatedString Thanks

Unrelated String

also looks like you can omit the outermost=leftmost dimension

lyxal

hey hi hello so say I have a javascript file I've imported as type=module, how would I pass that to another file I've included without having to mark the other file as type=module?

<script src="./vyxal.js" type="module"></script>
<script src="./main.js"></script>
<!-- I want to be able to call the stuff in vyxal.js in main.js without having to make it a module -->

main.js can't be a module because it relies on other included, non-module files

Huỳnh Trần Khanh

11:55

can you modify vyxal.js? if yes why can't you just set properties on the window object to expose your vyxal stuff

lyxal

I can't modify it

it gets automatically generated, and exports an object

it has an import in it meaning it needs to be a module

mousetail

Can you add a third module script that exports the members as window attributes, that then can be loaded by the non-module script?

lyxal

I'll try that

just gotta rebuild everything first lol

ooh that works

Ginger

12:17

I've come to make an announcement

Rabbit will not have a system for selectively importing modules or including them in dependencies

you either import a module or you don't

lyxal

@Ginger ~~shadow the hedgehog is a b*tch ass mfer~~

Ginger

if you have a module that causes a program crash if imported on Windows, that's bad module design

</callout post on my twitter dot com>

oh, also: casts in Rabbit

so there are two kinds of casts: safe and ~~unsafe~~ spicy

safe casts look like foo as Bar and spicy ones look like foo as! Bar

what's the difference? simple

you can safely cast a struct to a type iff the struct implements a cast method

they look like this:

struct Foo:
    as Bar:
        return Bar(<some parameters>)

you can also implement cast methods on structs you didn't create, using this even better syntax:

Unrelated String

are unsafe casts possible between types of different sizes

Ginger

as Bar for Foo:
    return Bar(<some parameters>)

I don't like this syntax, but it's consistent and I do like that

Unrelated String

ooh, for is a nice way to do extension methods

Ginger

12:23

@UnrelatedString what do you mean "different sizes"?

Unrelated String

i'll just let you get to how unsafe casts work :P

Ginger

@UnrelatedString yeah, I'm also doing trait implementations using trait X for Y

@UnrelatedString well, when I say ~~unsafe~~ spicy I mean spicy

Unrelated String

oh boy

Ginger

an unsafe cast literally switches the type of the target to whatever it's being cast to

that's why they're unsafe

Unrelated String

so it reinterprets the underlying representation

?

Ginger

12:24

yes, I think

I'm not sure if it'll copy functions over or not tho

probably not

Unrelated String

what happens if the underlying representation of the target type is larger

Ginger

I don't know what you mean by "larger" :p

Unrelated String

like i assume the underlying representation is, if not a raw array of bytes, some kind of container with a size

Ginger

I... guess? Currently objects are stored as instances of a class, with a MutableMap containing their properties

Unrelated String

ohhhh

Ginger

12:27

hmmmmmmm one sec

Unrelated String

so it doesn't arbitrarily reinterpret fields based on layout, it just... has the same fields as it would have in the original type

whoa

Ginger

so I have two things I can do here: unsafe casts can either 1. change the internal type property of the target, making the rest of the program see it as whatever its new type is or 2. create a new instance of the thing being cast to and copy over all properties from the original object, overwriting any duplicates

I like option 1, because it's the least bad of the two

or 3. I can just not do unsafe casts

why would you use them when you can implement a cast method?

is there any real use for unsafe casts?

Unrelated String

yeah i think #3 wins here

Ginger

(@RydwolfPrograms)

@UnrelatedString same lol

okay, now for annotations

Unrelated String

because honestly making it possible to create a value in a type that doesn't necessarily have any properties of a value of that type is even worse than the c-style of pointer cast bullshit

Ginger

12:30

I have two options with annotations: the Java way, where annotations are basically signals to a preprocessor, or the Python way, where annotations are a fancy way of invoking methods

I like Python more

anyway time to implement variable assignment in Rabbit!

actually, first I have to get my annotation preprocessor working

yay

Unrelated String

i feel like java annotations and python decorators are completely distinct concepts that happen to share syntax :P

Ginger

:+1:

math scat

true

Ginger

13:42

mfw my build environment commits die because I misconfigured something

Bbrk24

14:18

@Ginger At least in Swift, you use the same syntax for unsafe downcasts of class types as for unboxing Any with crash-on-failure

I don’t think allowing a reinterpret-cast with as! is a good idea, personally

Ginger

good thing I'm not adding it then!

Bbrk24

Swift does have a reinterpret-cast, but it's spelled unsafeBitCast(_:to:)

But Swift's as, as?, and as! correspond to C++ static_cast<T>, dynamic_cast<T*>, and dynamic_cast<T&>, respectively

Sandbox Posts

15:58

0

A: Sandbox for Proposed Challenges

Is it a valid chemical? code-golf decision-problem Non-metals typically[citation needed] have a fixed number of covalent bonds in every chemical they are part of. Given the number of bonds every element requires, output whether it's possible to construct a molecule from the given chemicals. Test ...

Sandbox Posts

16:39

0

A: Sandbox for Proposed Challenges

Great Leap to the Right code-golf Given a string s consisting of characters a-z lowercase, generate 5 arrays a1, a2, a3, a4, and a5 of size n (length of s) where the i-th element of each array represents the index of the next occurrence of the i-th character in s, with the following conditions: ...

Bbrk24

Since data in Trilangle is 24 bits, it's possible to attempt to print a value more than 21 bits wide as a character. No such character exists in Unicode. How should I handle this?

user

Make your own extension to unicode :p

Bbrk24

- Ignore it and generate invalid UTF-8, leaving the terminal/browser/etc to decide what to do (current behavior)
- Spit out an error message saying there is no such character
- Extend UTF-8 to handle 24-bit payloads and generate a byte sequence starting with 0xF8

user

Alternatively, only print ASCII. If someone tries printing a 24 bit value, print 3 ascii chara

mousetail

Print it the same as any other invalid character?

Alternatively mod 2^21

Bbrk24

16:49

Currently if you try to print 0x3C0000 it generates the invalid UTF-8 sequence FF 80 80 80 and then what it looks like is implementation-dependent

@mousetail Currently I just send a UTF-8 bytestream to the output and the implementation of that (terminal, JS, etc) is responsible for rendering it

This does not work very well on Windows, where the terminal interprets the incoming bytestream as CP-437

Ginger

cringe windows

Bbrk24

For example, the code "🎈@o prints "🎈" on the web interface but "≡ƒÄê" in my terminal

Apparently the current web interface renders the FF 80 80 80 bytestream as "��", i.e. four of U+FFFD

If I extend UTF-8 and convert that to F8 8F 80 80 80 instead, I expect to just see five of them

DLosc

17:18

@Bbrk24 If it weren't an esolang, I'd say option 2 definitely. Since it is... I'm not sure. But maybe still option 2.

:63169025 Typing with gloves on, lol

Bbrk24

But yeah I'm leaning towards an error message as well

Ginger

18:25

just got my annotation processor working :D

and, in an amazing twist, the Kotlin language server actually detected the generated code

Ginger

18:53

Advanced Kotlin:

this::class.declaredFunctions.filterIsInstance<KCallable<RabbitObject>>().filter { it.hasAnnotation<Wrap>() }.associateTo(properties) { Identifier(it.findAnnotation<Wrap>()!!.identifier) to WrappedFunctionDeclaration(it) }

now I can just decorate functions with @Wrap(<some string>) and have them be automatically added to properties!

Seggan

19:48

@Ginger ive never written one

howd ya do it

@Ginger python

im doing pythonic way for rol too

@UnrelatedString yep

but they both "preprocess"

Ginger

@Seggan suffering

Seggan

(also im back from texas, sry for the long absence)

math scat

20:05

How was texas?

Seggan

thursday was fine, friday rainy and chilly, saturday was really warm

everythings green there alr

here in south carolina the trees are just starting to bud (granted, i live in the extreme upstate; go south 45 mins and its prob like summer there already)

Ginger

did you see Radvylf? :p

Seggan

@Ginger nah he lives a bit far away

in The Sand Trap, Feb 24 at 21:26, by Rydwolf Programs

Pretty close to four hours away, that's where my grandparents live

Ginger

lol

Seggan

the slavic youth had a congress there, over 4k people

Ginger

20:12

huh?

Seggan

there was a big meeting of america's young slavs

(i.e. russians, ukranians, beylorussians, moldovans, etc)

Ginger

I did not know that was a thing

Seggan

~~we were discussing world domination~~

but seriously, we kinda took over tyler

you couldnt go anywhere without seeing slavs

Ginger

waow

~~I wonder if you're in any photos of the event~~

Seggan

i am

it got into local news

~~not that im gonna dox myself~~

Ginger

20:16

well, I go to github.com/advisories and the first thing I see is a High vulnerability for sqlite :|

(CVE-2022-43441)

Seggan

> Apache Log4j 1.x (EOL) allows Denial of Service (DoS)

noice

> builderio/qwik is vulnerable to code injection

great

Ginger

Seggan

awesome

Ginger

I don't know about yall but I think Answer might be vulnerable to Cross-site Scripting

just a hunch

mousetail

@Ginger Seems to just be the JS bindings, not the database itself

Ginger

20:22

Finally, we are become one with technology

Seggan

what

Seggan

20:48

AutoClosable is finally multiplatform in Kotlin

huh TIL [putting a space ](example.com) after words in a link unlinkifies it

Seggan

21:02

@user you got me

ima start building it (i dont think i have time today tho)

Seggan

21:14

CMP/LDQ: should i build it in rust or kotlin?

rust pros: rust stuff, probably more memory efficient

kotlin pros: more interop (if someone decided to build their lang in a jvm lang, looking at you vyxal 3, theyd have an api free)

im thinking kotlin but i cant get the rust option out of my mind

user

Rust is cooler, so if you're tempted by it, go for it

You already know Kotlin, so you wouldn't be learning anything new if you did that

If you use Rust, you can make Python bindings so that people can write golflangs in Python that use your super-duper-hyper-fast-native-optimized Rust functions

Seggan

@user i know rust well too

user

Ah ok

Seggan

@user i can make python bindings in kotlin too

i was thinking of "supply the compiler command as an argument"

also theres a jit option for later in rust

@RydwolfPrograms werent you writing a rust golflang lib once

caird coinheringaahin g

# Note of caution: We can't have functions inside of functions.

I hate my programming module

user

21:24

@Seggan Oh, didn't know you could do that

I think pxeger was writing a golflang lib for Python

Seggan

@user nvm i misunderstood ur msg

kotlin advantage: easy multiplatform

@user i wont have to worry about that anyway if the runtime takes the compiler as an arg

@Seggan so you can make fancy online runtimes too

Jelly Hypertraining

Transcript for