The Nineteenth Byte

lyxal

00:00

@ATaco it's still yesterday for them

ATaco

Oh, I should make the ISO folks happy and return 2023-03-15

Bbrk24

> BUILD SUCCESSFUL in 7m 46s

took you long enough

lyxal

@Bbrk24 what does PIE stand for here?

Bbrk24

Proto-Indo-European

Execution failed for task ':compileKotlinJvm'.
> Kotlin could not find the required JDK tools in the Java installation. Make sure Kotlin compilation is running on a JDK, not JRE.

@Seggan help

Seggan

aaaaaaaaaaaaa

run java --version

Bbrk24

00:08

Unrecognized option: --version

ATaco

O.o

Bbrk24

javac --version responds "javac 18.0.1.1"

AviFS

@ATaco We never stopped! Did we? I hope not.

Seggan

@Bbrk24 ehhh

you might have a borked java install

Bbrk24

The full output from gradlew build

Seggan

00:13

do you mind installing java 17? adoptium.net

Bbrk24

Oh, java and javac live in different folders O.o

one is in Program Files and the other is in Program Files (x86)

Yeah, I can try installing Java 17. I already have at least two javas though

Seggan

yeah ur install is borked

Bbrk24

The two Javas I know I have are Java 1.8 and Java 18

@Seggan I installed it, but my apparent Java version hasn't changed. I must need to change my default Java, but where did it put itself?

Nevermind, I found it

Seggan

mine is in C:\Program Files\Eclipse Adoptium\jdk-17.0.1.12-hotspot

Bbrk24

Yep that's where it is

Seggan

00:26

also change ur JAVA_HOME

user

If you're on Windows, where javac should help you figure out where it's coming from

Bbrk24

@Seggan How do I change it in powershell?

@user That still points to the old javac

Neil

@AviFS works for all n-dimensional spheres, because the delta volume is the surface area times the delta radius. can be made to work for n-dimensional cubes, but you have to correct for the radius being half the length of the side of the cube

user

You don't have to do it through Powershell, you can search "Environment variable" in start menu and just do it through there

5

Q: how do I permanently set a system variable ([System.Environment]::SetEnvironmentVariable) in windows 11?

As above. I tried (inside powershell terminal with run as admin) [System.Environment]::SetEnvironmentVariable('ResourceGroup','AZ_Resource_Group') and the variable is only available for that window. If I open another powershell window or restart my PC, the variable is lost. Thanks

powershell

Bbrk24

Ah, I knew you could edit them there but I didn't think to add it there

@user I set it there but echo $JAVA_HOME still prints nothing

(yes, even after restarting powershell)

user

00:32

Did you try in a new shell?

lol

Does echo $env:JAVA_HOME print something? idk how Powershell works

Also, did you set it in your user's variables or system variables?

Bbrk24

No, and it doesn't show up if I DIR env:\ either

I may have to restart my computer :P

Seggan

@Bbrk24 search for "environment" in the taskbar and it should show up where you can change it

Bbrk24

Okay, after restarting my computer, it is available as $env:JAVA_HOME

let's see if .\gradlew build works now

Yay! It worked

It, uh. Created a new file. kotlin-js-store/yarn.lock

Seggan

it will create zillions of new files :P

get used to build tools

Bbrk24

yeah but the others are gitignored. This particular one isn't. Is that expected?

Seggan

00:40

yeah its fine

hasnt been pushed to master yet

Bbrk24

got it

How do I run it?

Seggan

you cant (yet)

master is currently a skeleton project

i literally set this up a few hours ago :P

Neil

@Adám worst bit is that you can actually rotate around two planes (e.g. xy and zw) independently at the same time

Bbrk24

Well, at least I can build it now

ATaco

I love building anything and having /obj and /bin created and filled with junk

Seggan

00:58

@user well if its just numbers

user

Fair enough

@Seggan What were you envisioning?

Seggan

then releasing it separately is a good idea

bc its just a general number lib for kotlin multiplatform

user

We might be thinking of two different things so I just want to make sure we're on the same page before continuing

OldSandboxPosts

Sandbox posts last active a week ago: All languages have their strengths, right? (Cops' thread) cops-and-robbers code-golf, Convert imperative code to functional, Shortest Code to Implement a Simple Encryption Algorithm, Guessing on straws

Sandbox posts last active a week ago: Calculate the p-adic square root of -1

Seggan

but cartesian product is not just numbers, so we could just make another module in the same repo for it

user

01:00

@Seggan Well, also various golflang builtins, possibly even an infinite lazylist type, but yeah

Seggan

the problem with making a separate repo is that you have to make a new release for small changes

user

Stuff like matrix diagonals, all permutations, etc.

Same repo then, just a different module

Seggan

perhaps the number stuff could be a different repo (for those who need big math stuff), and then the rest a different module

user

Hmm

I was going to lump everything together but I guess a math library might be a good idea

Although there already exist math libraries for Kotlin so idk how useful a library solely for math would be

Seggan

but a big math lib plus complex numbers does not exist

smth like this, but pure kotlin (multiplatform)

← 17 messages moved from Vyxal

lyxal

01:05

:+1:

Bbrk24

I was wondering what all the arrows were for, and then I saw that last one

ATaco

CMC: Show a message box. EG: alert() or #include <windows.h>¶int i(){MessageBoxA(0,"","",0);}

lyxal

01:21

Vyxal E, 5 bytes: `ƈ⌈()

ATaco

Is that just basically "Eval literal: alert()"

lyxal

yes

but it's "eval as javascript"

n̥werttl̥h2tós strongʰis

for the online interpreter, couldn't the empty program with E work? :P

lyxal

well depends on whether you want a customisable prompt or the default warning :p

ATaco

I mean, that would definitely show a message box.

Hard Mode: Create and show a new Window entirely.

(Rather than just a messagebox)

lyxal

01:26

Vyxal E, 8 bytes: `×¥.¬⅛()

ATaco

0/10 opened as a tab

lyxal

`×¥.¬⅛("","","›„=1,¨→=1")

if you insist on a window

E flag obviously

@ATaco actually, just `↓Ẏ €ĿḞ»`E should work offline

Good old import antigravity :p

Bbrk24

I don't think it's possible in Trilangle, unless there's a massive oversight leading to an ACE bug

lyxal

Those are always fun

Especially when they're caused by the math library you're using

Bbrk24

01:44

@Seggan I should have push permissions to CGR, right? I want to put up a PR adding .vscode to the gitignore, since I need to put something there to make it default to powershell instead of bash. (As mentioned above, gradle doesn't work well in MinGW bash.)

(Also, I'm sure that once we do this, there will be someone answering CGCC questions in CGR bytecode.)

Seggan

@Bbrk24 pr

everything through pr

Bbrk24

Right I wasn’t going to push directly to master

that’s for personal projects only /hj

> ERROR: Permission to Common-Golfing-Runtime/common-golfing-runtime.git denied to bbrk24.

To be clear, that's not pushing to master, that's with --set-upstream

Seggan

@Bbrk24 should be gud

Bbrk24

Yep, I can now, thanks

Seggan

@user right, im abandoning this big math lib

even writing a bigint is hard. cant get addition to work :P

well stuff everything in the separate module

ATaco

02:00

Surely there is a way to represent a 3 dimensional Direction vector without wasting memory... rotations don't work as there are infinitely many points which have more than one (Infinite?) many rotations which reach it, and direct vectors waste a lot of space with invalid values that do not appear on a surface of a unit sphere.

In 2d it can simply be "256 equal division rotations"

Seggan

doing it specially tailored for cgr allows me to use platform libs for big decimals and the like

instead of writing my own

ATaco

Oh! A UV Sphere would work! 256 values around the yaw, and 256 values mapped from -90 to positive 90. Bonus points if you reserve 0x0000 as the top of the sphere and 0xFFFF as the bottom and shift the other values so they are never the same.

I suppose it could be simplified and allow for more even distribution to say, "The index of a point within a 256 vertex Icosphere" but converting to and from that would suck.

ATaco

02:18

We need a new Hypotenuse challenge, there's way too much overhead in the current one :(

user

@Seggan lmao

Do you want me to do it?

I can make a new math-lib repo or something

I'll just do it, easier to ask for forgiveness than permission

user

02:32

@Seggan Can you convert me to a normal Member instead of an Owner for CGR?

I don't need Owner powers

Seggan

02:52

@user i meant as in writing a whole specialised math lib aint worth it

we can just use the platform specific types and get a better result, faster

@user what if you will

user

03:27

@Seggan Oh ok

ATaco

03:43

I always love that the Q operator of HQ9+ is not a valid Quine, but Hello, World! is.

user

It isn't?

ATaco

It's an illegal quine, reading its own source code.

user

Interesting

ATaco

Although Hello, World!, only the H instruction, (Which happens to print "Hello, World!") does anything.

user

06:49

old.reddit.com/r/rustjerk/comments/vkdizx/…

mousetail

of course that sub exists

user

Oh wait it's fake

mousetail

Yea OFC, good meme though

user

@mousetail It's Rust, what did you expect

I thought it was an Easter egg of some sort

Just found it today, prepare for me to flood you with memes from there

mousetail

I'm braced, bring it on

user

06:55

Well, I guess I'll flood TST

Sandbox Posts

10:00

1

A: Sandbox for Proposed Challenges

Is it a Frog Number? Tags: code-golfnumberintegerdecision-problem Challenge: A Frog Number is defined by certain rules: Each frog within a Frog Number has an unique positive digit as id (any 0 in a number is basically an empty spot, which can mostly be ignored). A Frog Number will always have ex...

4

mousetail

No onebox?

lyxal

@mousetail Must be one of those times where it posts the message before chat has a chance to make the one box

It's happened before

mousetail

Can the bot edit the message a few ms afterwards to fix the issue?

lyxal

10:49

It doesn't know it didn't onebox

mousetail

You could just edit it every time after a short amount of time

If it oneboxed the first time it would do nothing

New Posts

12:11

0

Q: Is it a valid chemical?

Non-metals typically* 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 single molecule from the given chemicals. Test Cases Falsey: [1] A single element has nothing to bond with...

code-golf number array decision-problem chemistry

Ginger

12:28

time for rabbit!

mousetail

@NewPosts I see this got a close vote, anything unclear?

Bbrk24

"but you can ignore the complexities in this challenge.". Good, I didn't plan on dealing with [1, 1, 1, 1, 0] => true.

mousetail

Yea there are probably more exceptions to this rule then chemicals that actually follow it

Bbrk24

I think the one that's most confusing to me is B₂H₆. You'll never guess how it arranges itself if you don't already know it

mousetail

That's a nice shape, could be a corporate logo

Seggan

12:42

@Bbrk24 wait what

i expected H-(H-)(H-)B-B(-H)(-H)-H in SMILES

Bbrk24

My high school chemistry teacher used that as an example of why we'd never work with anything boron-related in that class

@Seggan Ethane but boron instead of carbon?

Seggan

yep

actually it should be [H3]BB[H3] im forgetting my SMILES

lyxal

@mousetail did someone say corporate logo?

mousetail

That looks more like the logo of a 4-year olds juice stand

lyxal

lol

Seggan

12:48

@Bbrk24 boric acid is really confusing too

id expect it to be an inorganic alcohol, not an acid

Bbrk24

Nah that's what I expected from seeing "B(OH)₃"

lyxal

@mousetail guess we sell juice now

Seggan

> There is no such thing as an inorganic alcohol

oh well

mousetail

I was thinking more this:

Like in the early 2000nds when logos where 3d

Seggan

13:08

@user not when everyones owner for some reason :P

Ginger

how should Rabbit imports work?

like internally?

or for that matter in syntax?

like in Kotlin you can't import packages, but in Python you can

Bbrk24

I like the way Swift does it: import Module is the same as Java's import module.*, or to mimic Java's import module.SomeClass you say import class Module.SomeClass. (You can also import enum, import struct, etc)

What I don't like about Swift's imports is that there's no way to rename something (import Foo as Bar)

Ginger

I'm probably going to do import module and from module import whatever, but the implementation is what I'm not sure about

Bbrk24

13:24

I want to say that imports must be resolved at compile-time, not runtime, but Rabbit isn't a compiled language is it

mousetail

I prefer import module.x instead of from module import x just so you don't need to change the whole sentence when changing the import type

Ginger

@Bbrk24 nope!

mousetail

(I'm aware that was not your question)

Ginger

I think the way I'm going to do it is import module links the target module's context to the importer's context somehow and from module import whatever copies whatever into the import's context

but that's bad too because then I have two copies of whatever

mousetail

You shouldn't copy modules ever really

Ginger

13:25

the problem is the way I'm doing contexts

each context, currently, can have one (1) parent

and if a variable lookup fails for a context it'll ask its parent to do the lookup

mousetail

In python when you do import it assigns the module object to sys.modules and subsequent imports copy the value. If you do from a import b it basically does import a; b=a.b

Importing shouldn't create a parent context

Ginger

in the current (unfinished) system each module has its own GlobalContext

mousetail

That makes sense

Then importing simply creates a object of type GlobalContext

Ginger

but how is that connected to the importer's GlobalContext?

should GlobalContexts have multiple parents?

mousetail

It shouldn't be at all

Ginger

13:29

what

mousetail

Why would they be connected?

Ginger

so that the importer can access the importee's variables

to clarify, GlobalContext is an internal object

it's not visible to the Rabbit code, but the interpreter uses it to manage global variables

mousetail

There is no parent child relation here. The context is now just converted to some container that is just a mapping of variables

Of course, functions inside the mapping can refer to the global context but the contexts are not linked

Ginger

hmm

mousetail

Like how python works

The Thonnu

13:32

Why did this get downvoted?

It's a valid answer as far as I can tell

mousetail

FractByte?

The Thonnu

Do we have an actual consensus on log_256?

I can write up a meta question if we don't

Ginger

@mousetail hmmmmmmmmmmmmmmm

Bbrk24

My opinion is that if your measuring system isn't convertible to integer bytes, just don't measure it in bytes (e.g. if it's machine code on a base-3 machine, measure it in trits or something.) I'm not aware of a meta consensus though

Ginger

so, like, when I import a module all of the things in its GlobalContext get copied over to the importer's GlobalContext?

mousetail

13:34

^^

The Thonnu

@Bbrk24 I didn't start the system, @Seggan did with Fig

Seggan

@TheThonnu we do

frac bytes are always allowed

Ginger

(well, all the things that are public)

mousetail

log bytes are dumb because they are compared inside the same language

I'm all for fractbytes but logbytes are dumb

lyxal

@Bbrk24 relevant consensus

Seggan

13:35

@TheThonnu i suggest rounding up, not down for fairness

mousetail

@NewPosts Why did this get closed?

The Thonnu

@Seggan alr sure

Seggan

@mousetail logbytes would be necessary if measuring in trits

mousetail

Then use integers trits

not log bytes

lyxal

@mousetail mod hammered

The Thonnu

13:37

@mousetail I can't see anything wrong with it

mousetail

Yea I know but what's unclear

The Thonnu

I'm VTRO'ing

lyxal

@mousetail (cc @cairdcoinheringaahing)

Ginger

@Ginger I guess this'd work, but then I need to figure out how to preserve the imported module's GlobalContext for use by any functions/structs/whatevers in the module

mousetail

@Ginger You'd only copy it after the functions have been initialized

Ginger

13:39

@mousetail what?

caird coinheringaahin g

I have no idea why [1, 1, 1, 1] is false, but [1, 1] is true

Ginger

FunctionDeclaration objects require a Context to be run

caird coinheringaahin g

The challenge should include some explanation of what makes it possible to construct a single molecule

mousetail

@cairdcoinheringaahing It's explained, 1, 1, 1, 1 is only possible with 2 separate molecules

caird coinheringaahin g

Okay, why is it only possible with 2 separate molecules?

Seggan

13:41

1-1, 1-1

you cant have 1-1-1-1- or anything else

mousetail

@cairdcoinheringaahing The first one needs to connect to a 1. Then both are "full", but there is left over atoms so it's impossible

lyxal

Quick question: what does the bonding for 4, 1, 1, 1, 1, 2, 2, 2 look like. I'm having difficulties visualising it and don't know where to look on the internet

caird coinheringaahin g

@mousetail Then add in something along those lines as an explanation

mousetail

@lyxal Not sure if that's possible in real life, but it works for the purpose of the challenge

lyxal

@mousetail I just want to know what connects to what, that's all :p

Because this reminds me of that puzzle where you're given a bunch of islands and you have to connect them with bridges

Bbrk24

13:46

@lyxal Would that be CH(OH)₃?

mousetail

Bbrk24

or that i guess

lyxal

@mousetail ah so two of the 2 connect to the 4

Thank you

Ginger

how in the frick do imports work

mousetail

@Bbrk24 This makes a lot more chemical sense though

Ginger

13:47

this is confusing me

mousetail

Basically, since a O has 2 sides it's "free" and can be placed in between any other bond without effecting the validity, as long as at least 1 bond exists beforehand

Seggan

@lyxal the koingsberg bridge problem?

Bbrk24

@Seggan relevant xkcd

lyxal

@Seggan Different puzzle

That kind of puzzle

The numbers are how many bridges a circle can have

mousetail

That's basically the same yes, except you can cross bridges

lyxal

13:51

Fair enough. I just thought that's what the challenge reminded me of

Seggan

@lyxal ah

@Bbrk24 lol. paying with aluminum is genius

lyxal

@Seggan it's one of my favourite puzzles along side signpost

I've played it quite a fair bit without knowing what it's called lol

Ginger

right, I think I've figured out how to do imports

import module will set module in GlobalContext to an object containing the module's members, and from module import thing will set thing in GlobalContext to the value of thing in module

Seggan

guys just one more reopen vote

mousetail

And it's back

Seggan

14:04

ive added 2 more

esp since the last one is a bit hard to visualise

mousetail

Tetrahedrane does not seem to fit the chemical formula

Bbrk24

It says "tetrahedrane without the hydrogens" though

mousetail

True, though I'm not sure if it's beneficial to list it then

Seggan

its helpful to visualise

> actual N4 does not link in a tetrahedron

(so doesnt ozone, but you can draw a latex thing with it)

mousetail

Ozone does not actually form a triangle shape either

Seggan

14:08

31 secs ago, by Seggan

(so doesnt ozone, but you can draw a latex thing with it)

Bbrk24

Cyclozone can exist in some conditions iirc? But it's definitely not normal ozone

> Cyclic ozone is a theoretically predicted form of ozone.

ah okay

mousetail

I think images and graphics are helpful but only if they illustrate the graph in terms of nodes/edges etc. otherwise they just add confusion

The Thonnu

14:31

@mousetail your challenge is on 15 revisions in 2 hours!

mousetail

Was just going to say

It's been a absolute rollercoaster

Bbrk24

Meanwhile, in JavaScript: [...text].length

The Thonnu

@mousetail It sort of goes to show how inactive the Sandbox is now

A lot of that should have been fixed there

mousetail

It got 5 upvotes in the sandbox lol

Bbrk24

'𫚣'.length is 2 but [...'𫚣'].length is 1. Because of course it is.

5

mousetail

14:34

Lol the starboard version is broken

The Thonnu

mousetail

UTF-16 is fun

Bbrk24

nah that's a font problem

it looks fine to me

mousetail

oh

Bbrk24

@mousetail This is why String.count in Swift is O(n) rather than O(1)

mousetail

14:35

Only way to make it O(1) and accurate is UTF-32

The Thonnu

4 bytes per character

Sounds fun

Bbrk24

No, because Swift goes by grapheme clusters, not Unicode scalars -- ā and ā compare equal in Swift but have different UTF-32.

mousetail

So today I found out that in python, strings start at ASCII, and are only converted to UTF-16 or UTF-32 when you add a high-byte chracter

This makes a lot of sense IMHO, not sure why it's not more common

Bbrk24

Swift String is basically a wrapper around an Either<_NativeString, NSString> in which _NativeString is UTF-8 and NSString it UTF-16

mousetail

That makes sense, though UTF-16 is kinda terrible

Bbrk24

14:38

You're a few decades late if you want to influence the design of NSString :P

mousetail

You can always delete things

Bbrk24

You call [+NSException raise:format:]. I deregister NSObject from the Objective-C runtime and let the program crash of its own accord. We are not the same.

/j

math scat

15:10

Just realised that the Havel hakimi algorithm errors for inputs like [3, 2, 1] (regarding this question)

And I have no idea why my Erdos theorem implementation returns false-positives like 50% of the time.

good times

Apparently I can't do sigma

Well Erdos won't seem to be work either, cause [3, 2, 1] isn't a simple graph.

Aiden Chow

15:34

ayyyy i got the outspoken badge, finally

Ginger

16:49

I have an exciting new problem with Rabbit, best summed up by this comment in Types.kt:

Seggan

i love exciting new problems

Ginger

> Initialization requires resolved types, but resolved types require initialization!

and I don't know how to fix this

in order to resolve the types for a module, its imports have to be initialized and ready, but getting the imports ready requires their types to be resolved, and so on all the way down

in theory, the modules at the bottom of the import tree should only use builtin/Kotlin-defined types (which don't need initialization), but I don't know how to program that in

Seggan

just resolve the types for a module before packaging it

thats what rol does

Ginger

wat

Seggan

oh wait this aint a compiled lang

Ginger

16:52

maybe I should clarify: "resolving types" here means converting identifiers representing type names into actual type objects

Seggan

ye ik

@Ginger just make the builtin types automagically resolved from the beginnning

Ginger

huh?

how?

Seggan

do it like rol does

i have an UnresolvedType and a ConcreteType

then youve got ConcreteType.NUMBER, ConcreteType.STRING, etc

Ginger

hmmmmmmmmmmmmm wait

Seggan

type resolution converts UnresolvedTypes into ConcreteTypes

Ginger

16:57

I can probably implement the type resolution system as a recursive function, and have it just not recurse if there are no imports!

that makes sense

now

@Seggan how does the Rol type resolver work? ~~(at least 1000 words)~~

Seggan

... is there any other option

@Ginger gimme a few hours :P

Ginger

I was Jo King lol

Seggan

no youre not, youre Ginger

Transcript for