The Nineteenth Byte

Sandbox Posts

00:00

0

Can you swap? code-golf Today's problem is easy. You're given two strings A and B of equal length consisting of only the characters ( and ). Check whether after a few operations you can make both strings balanced or not. The only operation allowed is swapping A[i] and B[i], where i is an arbitrar...

RubenVerg

@Simd if I understand what you want, I think (</¨⊢⍤/⊢)⍤,⍤(∘.,)⍨2÷⍨2+/0,⊢,1+⌈/ works it

I take all pairs of the 0,list,1+max(list), average them, then take Cartesian product with itself and filter for the ones where the first is strictly less than the second

would that be a correct algorithm?

"all pairs" means consecutive pairs, so 0,list[1], list[1],list[2], ..., list[n],1+list[n]

New Posts

04:20

0

Q: how do i solve this version of the goat problem

A goat is tied to a triangle pole with side length 4m, the rope is 12m long. a) show that the distance the goat walks in one revolution is 16pi meters b) find the area of the shaded region c) what is the area of grass that the goat can graze d) repeat parts a,b,c for a pole with a square base of ...

math

Neil

06:59

@NewPosts are there any resources for default string compression for the various golfing languages?

lyxal

øD, øc and øC in Vyxal for dictionary, alphabet and number compression respectively

Other languages have information on their tips questions

Simd

07:24

@RubenVerg I can’t see it run sadly

RubenVerg

@Simd ato.pxeger.com/…

Simd

@RubenVerg perfect!

RubenVerg

I think it can be golfed much further, the real issue was figuring out what the problem wanted

so I didn't want to spend much time optimizing a potentially wrong solution

Simd

@RubenVerg that is always the problem :)

Simd

07:53

which language are you using?

Neil

08:50

@lyxal how do you use them? the first one gave me a longer string and the other two threw an exception

Neil

09:02

actually it wasn't longer I was just measuring the wrong thing

lyxal

@Neil øc expects a string of letters in [a-z ] and returns the base 255 compressed version

øC expects a number and returns the base 255 compressed version

Neil

@NewPosts I went through all the string compression generators I could find. They were terrible. It seems that in most languages you'll actually need to manually compress the string, e.g. by repeating the tail. The one exception is Charcoal, which naturally compresses the string to 320 bytes.

(in other words, trying to do manual compression actually results in longer code)

lyxal

09:18

Should the challenge be considered a dupe of the rickroll challenge?

Given it's just compressing a large chunk of text + some repetition

Feels like it would be a challenge that has been asked before

The Thonnu

@lyxal other similar challenges have in the past

So yeah you could

(I'm so tempted to use my dupehammer rn)

lyxal

That's why I asked here first :p

@TheThonnu it can always be reopened if need be

Go on, give it a crack

The Thonnu

No way

That's so cool

lyxal

Unilateral decision making can be fun sometimes :p

~~although sometimes you wish that some things weren't unilateral and or binding~~

The Thonnu

Show-off :p

Don't forget we're the ones who voted for you lol

lyxal

09:30

I'm just saying it'd be nice to have an option that wasn't binding

Like a toggle switch

The Thonnu

Yeah fair enough

lyxal

I'm not the only one who's said things like dupe hammers would be good as opt in

The Thonnu

Jul 24 at 4:09, by The Thonnu

ik it means that you have to be completely sure before you cv

lyxal

It's a good idea 99% of the time to have it automatically close, but there's always 1% where you want to get more feedback while still performing an action

@TheThonnu well not completely

It can always be reopened

Well, a dupe hammer at least

Ooh you can add additional posts a post is a duplicate of if you have a hammer

I wonder how long that's been a thing

The Thonnu

09:48

@lyxal you didn't know that?

I found out when that happened to my question on MSO last August

But it's definitely been around a lot longer than that

mousetail

@lyxal For many years at least

lyxal

10:39

@TheThonnu I've only had a dupe hammer here, and only posted like 1 or 2 dupes back when I didn't care about moderation tasks (like mid 2019)

I also don't typically go back to duplicate questions :p

So I've never really had a chance to look at the modal closely lol

user

12:56

Anyone have an invite for the CGCC Discord?

lyxal

one seocnd

discord.gg/WgTEPvP

careful, that link enables PvP on your account :p

user

PvP?

lyxal

(look at the link url)

user

Oh it's a joke

lyxal

you expected anything less? :p

user

13:00

From you, I guess not

mousetail

PvP = Player versus Player

user

Yeah ik what PvP is, just was confused because I don't know what it would mean in the context of Discord

mousetail

It means you can fight people in real life

New Posts

13:15

0

Q: Can you perform swaps?

Today's problem is easy. You're given two strings A and B of equal length consisting of only the characters ( and ). Check whether after a few operations you can make both strings balanced or not. The only operation allowed is swapping A[i] and B[i], where i is an arbitrary index. Tests: ( ) -> f...

code-golf

Simd

I am running an optimization where I have to find the optimal set of intervals between 0 and 100. I thought it would be fun to see the intervals change live graphically but I code in python. Anyone know if this is plausible in python?

I use matplotlib to plot things but never for anything animated

The Thonnu

@NewPosts second challenge in a row that was in the sandbox for less than a day

lyxal

13:47

I've been waiting to use that one for a while :p

noodle man

14:12

anyone have the link for lumenspire or whatever it was called? can't find it

Sandbox Posts

0

A: Sandbox for Proposed Challenges

Pack the sequence! You have a list of sequences of numbers (1 byte max, unsigned) as an input, e.g.: 4 6 4 9 3 15 150 3 2 1 9 6 4 1 … Your task is to pack as many of them as possible into the square composed from the numbers from this list - the size of square is also an input. The square of siz...

noodle man

@lyxal i thought i was in the CGCC discord for so long and only just realized i was only in the gaming one...

@noodleman nevermind, found it

joyoforigami

@TheThonnu This is regarding this question: codegolf.stackexchange.com/questions/263479/… I would argue in fact that while lots of string compression questions have been asked, some of them are somewhat interesting and we shouldn't just close all of them just because "string compression" has been asked before. But this is kind of a matter of personal opinion on what things are interesting and different.

noodle man

14:28

@joyoforigami generally, the reason challenges like that are closed is because they are all very similar and rarely have much challenge to them

if we let every string compression challenge stay open it would get pretty boring since submissions can use very similar techniques in all of them, which is basically how duplicates are chosen on this site (can submissions from that challenge be trivially changed to work for this challenge?)

question: should an image of a table be replaced with a markdown table? the table is not very large, but the challenge cannot be solved without the table.

i'm going to go ahead and edit it but not entirely sure.

The Empty String Photographer

@Simd optimized how?

@Simd You might want to consider doing 3-dimensional graphs instead: geeksforgeeks.org/…

c--

@Simd did you look into matplotlib.animation?

The Empty String Photographer

@Simd if you need a higher dimension then you can use graph colouring to represent higher dimensions.

Simd

14:53

@c-- I didn't know about that!

I want it to update as the code runs

c--

I don't know how fast your code runs, maybe the default interval isn't good enough for you, maybe you need to run the computation in a different thread and skip frames?

I'm only familiar with FuncAnimation(TimedAnimation)

Simd

I will check it out, thanks

c--

this may be overengineered but for real-time computed values your best option might be to subclass Animation providion your own event_source

I haven't looked into that though

maybe first try a simple solution like this, and if that doesn't work look into more complicated stuff

Simd

15:59

@c-- thank you!!

Bbrk24

17:35

bbrk24@cdlemon:~/workspace/Trilangle/wasm$ gzip index*js -kS .gz
gzip: index.js.gz already exists; do you wish to overwrite (y or n)? y
gzip: index.min.js.gz already exists; do you wish to overwrite (y or n)? y
bbrk24@cdlemon:~/workspace/Trilangle/wasm$ yes | gzip index*js -kS .gz
gzip: index.js.gz already exists;       not overwritten
gzip: index.min.js.gz already exists;   not overwritten

You have defeated the entire point of yes :facepalm:

gzip has -f doesn't it

Neil

You have defeated the entire point of stdin :facepalm:

gzip by default compresses stdin to stdout, unless it sees stdin is a terminal, in which case it guesses you want to do something else

Bbrk24

-kS .gz means "keep the original files and create new ones with the .gz file extension"

anyways, gzip index*js -fkS .gz is what I wanted

@Neil It only wrote to stderr, not stdout

Anyways, I'm trying to determine whether minification is even worth it

$ wc index*js*
  446  1408 14164 index.js
   15    97  3981 index.js.gz
    0   110  6897 index.min.js
    7    55  2532 index.min.js.gz
  468  1670 27574 total

Right now I'm leaning towards "no"

mousetail

18:13

I have 2 numbers in rust, I want a tuple or a array containing the smallest first and the largest second. What's the golfiest way to do this?

Bbrk24

Is there no minmax builtin?

C++ has std::minmax

mousetail

Not that I can find, maybe under a different name

Bbrk24

If you have exactly two numbers you could probably do sth like [a,b].sorted()

mousetail

Sorted does not exist either :/

Bbrk24

That exact code snippet would work in Swift :P

mousetail

18:18

(This is where I want to use it btw, codegolf.stackexchange.com/a/263499/91213 I'm need it twice so an efficient method would save a significant number of bytes)

Bbrk24

ah I see

The Empty String Photographer

Mousetail-Inspired-CMC: Given an array of numbers or multiple inputs, output the smallest item first and then the biggest item second. You may also output as an array containing these outputs.

mousetail

Currently I do (a.min(b),a.max(b)) but it's long especially if a and bare expressions rather than single variables

Bbrk24

@TheEmptyStringPhotographer C++: std::minmax :P

(tbf that does take either two numbers or a std::initializer_list so not an array properly, only an array literal)

mousetail

I guess python would be, for an arbitrary list lambda a:sorted(a)[::len(a)-1]

Bbrk24

18:26

{($0.min()!,$0.max()!)} in Swift

Simd

If I add a bounty to my fastest-code how quickly do I need to time the code on my PC? Should I time any new submission or would be ok to time them all at the end of the bounty?

mousetail

I'm sure people would appreciate if you timed them fast so they have an idea if they are set or need to improve it more

Simd

@mousetail ok thanks. I ask as I am going away for a week. I won't put the bounty on until I am almost back in that case

@mousetail another question... does anyone care how large the bounty is?

The Thonnu

18:42

@Simd any sized bounty will probably attract more answers, since it will put the question onto the "bountied" tab. But ofc, the bigger the bounty, the more people will be encouraged to write up an answer.

Simd

@TheThonnu thanks. I wonder if there is data to support that hypothesis

mousetail

I don't think there are many people on this site who go for bounties in general, it's too easy to gain rep normally

Simd

@mousetail I do! :)

mousetail

You have 0 answers so clearly not

Simd

@mousetail I mean I give away bounties

mousetail

18:46

I mean people who would be more likely to solve a challenge because of a bounty

If a challenge looks fun and interesting, I'd solve it regardless. Otherwise even a bounty wouldn't persuade me, I'm here for fun.

Simd

ah ok

I'll add a bounty in a week and I hope it will encourage answers!

mousetail

There is not that many posts here that I need to plan which ones to solve, I solve all that interest me

Simd

@mousetail you could win easily currently :)

mousetail

I don't enjoy solving fastest code challenges

Simd

:(

mousetail

18:49

I'm really interested to see what the best answer will be though

I hope they find an interesting algorithm and not just basic optimizations

Simd

me too!

sometimes there is a clever method to make things go fast that isn't provably better

those are cool too

mousetail

If it's algorithm specific

If the optimization is just "I used threads" then I'm not that interested

Simd

codegolf.stackexchange.com/a/197576/116117 is genius for example

full of very clever tricks of different sorts

mousetail

Intrinsics :/

Simd

and a clever algorithmic speedup!

it's got everything

mousetail

18:54

I'm still salty about the one time I tried to solve a fasted code challenge and I used platform agnostic SIMD and then another answer had basically the same but used intrinsics for SMD and it was like 4 times faster

Simd

Argh. Which challenge was that?

Julia is pretty good for these things these days but we don't seem to have any Julia lovers here

mousetail

You won't beat those answers that are half assembly

The Empty String Photographer

19:16

@mousetail lambda a:(min(a),max(a)) is shorter and more naïve.

noodle man

i don't like answering fastest-code since the only praclang i'm good with is JavaScript

Bbrk24

JavaScript is on par with Java in terms of performance, which surprises me

noodle man

my only fastest-code answer was just ported to C by somebody else and demolished

Bbrk24

but modern V8 is really well-optimized

@noodleman yeah that's one thing I don't like

Seggan

@Bbrk24 not in general tho

Bbrk24

19:21

plus you can just throw -fno-exceptions -fno-rtti -flto -O3 on any C++ answer to instantly triple the runtime speed

@Seggan What do you mean by that?

Simd

@mousetail I guess that is true if the assembly coder is really an expert. It's hard to write good quality assembly though

The Thonnu

19:44

CMP: Who thinks codegolf.stackexchange.com/q/263479/114446 should be reopened?

I'm starting to think that because of the repetition it is slightly different from the other challenges

Unrelated String

dupe target also has repetition, to be fair :P

but yeah it's definitely structurally different

but also a ton of it has barely any structure

also speaking of structure i'm beginning to wonder if someone's already cracked this and it's just taking this long to actually write the next answer :P

mousetail

@TheThonnu If gzip is the shortest answer it's a dupe, in this case gzip seems to be the shortest

Bbrk24

zlib is shorter than gzip :P

Unrelated String

also we haven't seen any golflangs with built in dictionary compression take a crack at it yet

a boring challenge for praclangs might still have some merit there

mousetail

20:04

That's true

user

@Bbrk24 Multithreaded, perhaps?

Bbrk24

new Worker

user

?

Bbrk24

You can do multithreading in JS via workers

user

Yeah, but is the performance better than multithreaded Java?

Bbrk24

20:11

no idea

Rydwolf Programs

20:49

I think they mean Java's better in ways other than performance? But not really sure how that'd be relevant if JS is just a compilation target

Bbrk24

We were talking in the context of fastest-code

and obviously neither Java nor JS will beat anything compiled natively

Rydwolf Programs

Altho JS can get close :p

Bbrk24

I was saying that JS and Java will be similar to each other

but yeah asm.js is a thing

Rydwolf Programs

Nah I just mean V8

It's slower than C/C++, but only by a single digit multiplier, which is pretty damn impressive for what it is

Bbrk24

yeah modern V8 is impressively well optimized

Unrelated String

20:55

wow

Rydwolf Programs

I mean I guess with the whole internet basically depending on JS being fast, there's a lot of incentive to make it that way :p

user

@RydwolfPrograms I know I should be happy V8 is that good, but it just makes me sad because of all the interpreted languages that exist, JS got chosen to be the backbone of the internet

It's not an awful language but it's an awful language

@Adám You gotta make Dyalog APL target WASM so it can take over the internet :P

Bbrk24

wasm still can't manipulate the DOM

user

I don't want to live in a world where every website uses JS on the frontend and backend and every mobile and desktop application also uses JS

@Bbrk24 Wait, it can't?

Bbrk24

nope

Rydwolf Programs

21:03

IMO a sort of machine code should've been the scripting language of the internet

user

That's pretty limiting

Unrelated String

wasm applications can but only through a stupid js api

Adám

@user We're already looking into it.

user

Nice

Rydwolf Programs

Maybe like a super-CISC that's a superset of most of the common architectures (x86_64, arm64, RISC-V, etc.), where every instruction can be broken down with a lookup table into an instruction, set of instructions, or library call for the platform it's being run on

user

21:04

@Bbrk24 and @Seggan, re: the Common Golfing Runtime, is it going to target WASM or just JVM, JS, and native executables?

Bbrk24

Are we doing native?

user

idk, I assumed we were given Kotlin compiles to it

Bbrk24

We only have Complex impls for JS and JVM

user

(also can you add me back to the GitHub org? I left it)

Bbrk24

sure

user

21:05

Thanks

Rydwolf Programs

Ooh can you add me too

user

@Bbrk24 Oh I thought you just needed a single impl for all the platforms

Bbrk24

Well we're using different BigDecimal libs for the two

user

Ah

Rydwolf Programs

If JS is already a target, Wasm's probably not that important right?

user

21:06

Yeah

Bbrk24

@user What's your GH username again?

user

ysthakur

Bbrk24

Got it. I knew it started with ys but I couldn't remember the full spelling

user

But I hear WASM's going to be used for more than web stuff in the future (heard something about how it would've made Docker unnecessary if it'd come out sooner, although I have no idea how it would've done that)

@Bbrk24 Here's a handy way to remember: it's rukahtsy backwards

Bbrk24

Huh, GH has a special dialog for "This person used to be a member, are you sure you want to reinstate them?"

user

21:08

Thanks!

Bbrk24

@RydwolfPrograms I would want to run that by @Seggan or @lyxal first so I'm not unilaterally adding new members. (I forgot Lyxal was even part of it)

Rydwolf Programs

@user Yeah no I think the major advantages of OCI/containers are kind of orthogonal to those of Wasm-not-on-the-web

Bbrk24

Well the docker thing was about WASI, which is similar to POSIX but different

user

Oh

Bbrk24

but WASI was designed for wasm so

user

21:10

I'm still not sure it does what containers do

Bbrk24

More GH actions questions: If I have push: branches: '**' do I need pull_request?

wait I could just see if the workflows ran on the PR you opened the other day :P

they did not

Also, why are my commits showing up as "8 hours ago". Is there a timezone problem

user

I forget, is a debugger on the roadmap for CGR? Because without one, it wouldn't really have a ton of advantages over a plain golfing library

@Bbrk24 Your commits simply time travelled

Make sure to never mess with git history

Bbrk24

It wasn't doing this last night

this is new

Even git log locally shows them at the wrong time so something is up

Seggan

@Bbrk24 in specific cases, yes, but not in general

Bbrk24

247fb01 was in fact committed at 2:30 in the morning. 11acfa2 was committed this afternoon. Sometime in between those commits, git got messed up

user

21:19

Is it a thing with your own computer?

Try making commits from GitHub directly, or some other device maybe?

Seggan

@user WASM kotlin support is still experimental, besides we have JS

user

Right yeah

No native either right?

mousetail

In general WASM is a more suitable compilation target than JS if you have the choice

user

@user @Seggan Also this

Seggan

@user go write a BigDecimal impl in native, then sure, go ahead :P

@user not really, but its a great idea

user

21:21

@mousetail It's going to depend on some JS libs, not sure how easy it would be to access them from WASM

Bbrk24

I committed just now and it showed up as being at 8:32 so it's not even a timezone thing. It's 8 hours 48 minutes out

user

Is your system clock off?

Bbrk24

nope

mousetail

It's quite easy to call JS function from WASM

Bbrk24

user

21:22

ET gang

Seggan

im actually thinking of moving to rust, since it can do wasm too

Bbrk24

@user I thought you were in India

user

No more JVM tho :(

@Bbrk24 I haven't been to India in 7 years lol, I live in Maryland

Seggan

@RydwolfPrograms ~~pass this 1024 question exam firstt then well let you in~~

@user i agree, the JVM is nice

user

Nah but if Rust makes things easier let's go with that

Bbrk24

21:23

~~CLR is better~~

Seggan

NEVER

user

Also, having to target multiple platforms is tough

Seggan

@user it is?

Bbrk24

But yeah I don't have a JDK on my new computer so Rust would be nice

user

@Bbrk24 Less optimized + Microsoft + CLR + doesn't run on a gabillion devices

Seggan

21:24

@user as in the wasm/desktop thing or multiple native targets?

user

No as in JVM and JS are multiple targets for the current Kotlin impl

Bbrk24

oh I figured it out

Rydwolf Programs

Isn't supporting multiple native targets in Rust like, really easy?

Seggan

@user exactly

Bbrk24

$ date
Sat Jul 29 08:36:31 EDT 2023

Seggan

21:24

@user kinda the biggie is with math stuff

mousetail

wasm-bindgen in rust contains convenient wrappers for most JS functions and types making it convenient to cross lanuguage bounderies whenever you want to

Rydwolf Programs

Also, if the goal of this is to be a runtime, it seems like depending on a second runtime would be kind of silly

user

@RydwolfPrograms Was talking about how CGR currently targets JVM and JS, not that

mousetail

Downside is it's quite large in terms of binary size

Seggan

@RydwolfPrograms too meta for you?

Bbrk24

21:25

I found out why this happens, btw: it's a known issue with WSL2. stackoverflow.com/questions/65086856/…

user

@mousetail Nice, so Rust effectively gets us CLIs and browsers

Unrelated String

oh yeah that

Seggan

@cairdcoinheringaahing (pinging you since ure the most recent one in here) can you unfrz chat.stackexchange.com/transcript/144825?

im so annoyed ROs cant unfreeze

Unrelated String

i used to have to deal with that all the time but... actually i can't remember if it stopped or if i just stopped doing online stuff through wsl :P

Seggan

@RydwolfPrograms whats ur GH

Rydwolf Programs

21:28

Radvylf

Seggan

done

i sadly cant work on CGR today, got other priorities, but might monday

Bbrk24

If we're redoing it in Rust, performance might be a thing

It might be a stretch to say that a golflang will become competitive for fastest-code

Seggan

@Bbrk24 Myxal go brr

Bbrk24

but I can imagine that Vyxal in CGR will be faster than Vyxal in Python, because Python is so slow

Seggan

@Bbrk24 you ever doubted the JVM's performance? how dare you

@Bbrk24 Vyxal 3 is in Scala tho

Bbrk24

21:38

Oh

right

user

Since we're dropping the JVM, we can at least expose Python bindings so that people can use it as a library

@Seggan The JVM is pretty good, but native is even better

@Bbrk24 Jelly programs go from running in 600 billion years to merely 500 billion years

4

Unrelated String

*brachylog

jelly tends to outperform the language that basically has brute-forcing as a primitive :P

Seggan

@user i disagree, in terms of performance

sure hyperoptimized native will be faster, but not if ure the average coder

and ofc, the JVM really shines once its warmed up; cold start JVM is pretty slow compared to native

user

I assume this new version of the CGR won't be coded so horribly that Java would be faster :P

Seggan

erm i think its hard to beat a JIT lol

user

21:46

@Seggan A lot of programs are pretty short-running and will only be run once, so native would be better there

@Seggan Is it?

Seggan

yeah

Unrelated String

don't jits tend to lose to full ahead-of-time compilation

user

^

Bbrk24

depends

Seggan

@user itsnt that the whole point of JITs?

@UnrelatedString i dont think so

user

21:47

You can do some extra optimizations with JIT that you can't with AOT but AOT still compiles everything, right?

Bbrk24

In some cases JITs are better because you can use it to specialize tight loops

Seggan

^^

Unrelated String

jits are way better than straight interpretation but oh interesting

Seggan

@Bbrk24 not just

Bbrk24

heck, someone made a C++ JIT for that reason

github.com/jmmartinez/easy-just-in-time

user

21:48

Whoa, interesting

Seggan

IIRC the reason JIT compilers can be faster is that they have access to the specific machine they are running on + more metatdata than an AOT compiler

Bbrk24

22:05

lyxal

@Seggan haha I have done this myself :p

Refresh the page

Simd

@Seggan aren't JITs slower typically?

Bbrk24

Depends on the application

Simd

Are there any languages with both AOT and JIT compilers where the latter is faster?

Seggan

@Simd Java :P

@lyxal you spy

Simd

22:09

@Seggan does it have an AOT compiler?

I guess it compiles to byte code

lyxal

@Seggan nyeh heh heh :p

Seggan

@Simd Graal Native Images

Simd

graalvm.org/22.2/reference-manual/native-image :)

Is it slower ? I have never tried it

lyxal

@Bbrk24 I forgot I was part of it too lol

Simd

Julia really needs that

Bbrk24

22:12

holy cow lyxal

the message spacing

Seggan

@Simd graalvm.org/latest/reference-manual/js/FAQ/….

lyxal

@Bbrk24 and that's why I fight for you Alberqerque!

Dannyu NDos

Is it possible to propose a challenge that is beyond RE-complete?

Simd

@Seggan I find it hard to work out which is faster!

Seggan

@Simd Also Kotlin (Kotlin/Native being the AOT)

Simd

22:15

Is it the AOT compiled version?

The JIT will have to warn up

Seggan

22:35

@Simd Kotlin/Native is slower than Kotlin/JVM

Transcript for