The Nineteenth Byte

DialFrost

00:10

@JoKing does the 31 byte solution have something to do with walrus operator?

Jo King

how would that be shorter

DialFrost

Good point

Gahhh

Bubbler

01:44

feedback?

DialFrost

@Bubbler none, post it :3 I wanna see someone's <9999 solution

Seggan

02:18

@DialFrost yeah

DialFrost

@Seggan howd u find 80 dart leh? :p

Seggan

i didnt

watch who you pinging bro

were not a hivemind anymore :P

New Posts

0

Q: 50 digits of π in HQ0-9+-INCOMPUTABLE?

Background HQ0-9+-INCOMPUTABLE?! is a half-joke programming language introduced in Internet Problem Solving Contest 2011, Problem H. HQ9+ is an esoteric programming language specialized for certain tasks. For example, printing “Hello, world!” or writing a quine (a program that prints itself) cou...

code-challenge

Seggan

02:35

@DialFrost what you talking abt anyway?

@DialFrost Kotlin

DialFrost

@Seggan 80 byte dart quine, i got 85 lol

Seggan

where

DialFrost

@Seggan code.golf

Seggan

ah

DialFrost

wait i meant to ping @Steffan not seggan

:/

earlier lol

me when the issue why my code wasnt working in JS was cuz of " vs '

DialFrost

03:53

for this is the "binary fraction" jsut a binary?

im confused

int(binary,2) doesnt seem to work

Jo King

integers are, by their nature, whole numbers

DialFrost

my poor attemp that works partially

x=y=0
a=""
while-43516<y:a+=str(1+x%2);x^=~y&~-y;y=~-y&x//2
a="."+"".join(str(int(x)-1)for x in a[1:])
print(int(a[1:],2)/2.**999)

it doesnt give enough decimal places unfortunately

Bubbler

try to not use floating point numbers

Jo King

^

DialFrost

@Bubbler ?

why tho

Jo King

04:03

like the other constant holes, you should try starting with 10**1000 representing 1

floating points have limited precision

Bubbler

because a float can't accurately represent 1000 digits of decimal

DialFrost

@JoKing ???? I can't solve the other constant holes though

@Bubbler oh

@JoKing wdym 10**1000 representing 1

Jo King

@DialFrost the classic 0.1 + 0.2 != 0.3

Bubbler

more like: print 0., and then print a 1000-digit integer

DialFrost

@JoKing why does it print 0.3000000004?

Jo King

04:05

> floating points have limited precision

DialFrost

:/

Jo King

0.3 is not a power of 2, or an easily representable sum of powers of 2, so it gets mangled to the closest a float can manage

Bubbler

or in python, you can use decimal module I think, though it won't do a good job at short code

DialFrost

@Bubbler sigh ok thanks (not trying any more constant challenges anymore lol)

Jo King

try looking at the pi or e challenges here on CGCC for inspiration

Bubbler

04:08

you really need to try to solve things first before even trying to golf them

5

this is why I recommended codewars the other day

DialFrost

@Bubbler ye I can't lmao

Jo King

you can only get so far with googling the problems and modifying other people's solutions

DialFrost

@JoKing I tried but don't understand

@JoKing tru..

@JoKing the attempt above was all mine except the sequence part

Jo King

has a few redundant parts, like that whole fourth line

you add one to each digit, only to use a whole line to remove it again, and then do the same again with the .

Bubbler

also you need more than 3000 binary digits to get 1000 digits in decimal accurately

DialFrost

04:15

@Bubbler OverflowError: int too large to convert to float whoopsie

Jo King

that magic number, 3320

DialFrost

@JoKing ?

Jo King

that's roughly how many iterations you need for most constant holes

DialFrost

this error aint allowing me to get any larger iterations, lemme find another way round

Jo King

just checking, do you understand how a fraction works in binary? otherwise you're probably not going to be able to get it

DialFrost

04:18

@JoKing no

oh, err okey then

Jo King

0.111_2 = 2**-1 + 2**-2 + 2**-3 = 0.5 + 0.25 + 0.125

DialFrost

@JoKing huhhhh???

how does 0.111 = 0.5+0.25+0.125

Bubbler

that's how binary works with a decimal point

in base 10, 0.111 = 0.1 + 0.01 + 0.001 = 10^-1 + 10^-2 + 10^-3

you get this?

DialFrost

errr

Jo King

so when i said you can represent 1 as 10**1000, you can then use 10**1000//2 to represent 2**-1 = 0.5 and then represent all the rest of the digits with integers

DialFrost

04:22

so 10 is 2 in base 2 right

Bubbler

yes

Jo King

10_2 = 2_10

DialFrost

@Bubbler god this is so confusing

Jo King

is there some standard notation for bases without using Mathjax?

Bubbler

I don't think so

DialFrost

04:23

@JoKing er wut..oh well just forget it

Jo King

what part don't you understand? you seemed to get the division by 2**999, which shifts the binary integer to a fraction

DialFrost

05:53

this guy requires exactly 10 rep to get 10k :p

@JoKing sorry i just got frustrated so i didnt wanna try it sorry :(

DialFrost

06:40

anyone have any hitns for 55 bytes? JS collatz code.golf

f=n=>n-1&&-~f(n%2?n*3+1:n/2)
for(i=0;i++<1e3;)print(f(i))

mousetail

07:00

@DialFrost On week golf the solutions are public

DialFrost

@mousetail i tried logging in with github and it bugged btw

mousetail

Log in on SO

Tell @TKirishima if it's bugged

DialFrost

@mousetail wdym

mousetail

The solutions are public

So you can see how collatz was done in every language

There is no pure colatz challenge but one that includes a collatz component

DialFrost

i got confused by what the challenge was asking for and gave up lmao

@mousetail u have to do it first right

like post mortem?

mousetail

07:08

No

DialFrost

?

mousetail

This is a old challenge

Answers are secret for only the first week of each challenge

DialFrost

oh

@mousetail really meh

i took a look at the solution, didnt see much interesting stuff

mousetail

Mostly look at t he difference between the best and second best

those are golfs

DialFrost

@mousetail and that will give me 55 bytes somehow? XD

mousetail

07:11

No you do that yourself

This is just hints

DialFrost

@mousetail no

mousetail

you wanted hints

DialFrost

I meant hints

mousetail

yes

DialFrost

since when this challenge is hints lmao, its honestly a ... complicated one

Jo King

08:56

@DialFrost there's a blank space in the for loop you aren't taking advantage of

DialFrost

@JoKing blank space??

Jo King

after the second semicolon

DialFrost

oh

you can put something there???

Jo King

you never seen the traditional for(x=0;x<i;x++) style loop?

DialFrost

ya I've seen it

but I was asking what to put there lmao, cuz I don't see anything of use

UndoneStudios

09:02

@JoKing tbh the first kind of loop I've seen is for each [x v] in ()

from scratch removed blocks

DialFrost

@JoKing the only other 2 things avaiable are the print(f(i)) or the function, both where i dont see any adv

Jo King

you can save on the newline

DialFrost

I tried that, unless I messed up the syntax

isnt JS function style like this? f=n=>...

no alternatives?

Jo King

you can also do f=(x,y)=>... for multiple parameters

or f=x=>y=>... for chaining

Jo King

09:19

new hobby, short-circuiting other people's boolean expressions. Just wait for them to ask something like "Is the meeting in one hour or-" and hit them with the taser

7

UndoneStudios

@JoKing speaking of booleans, it turns out we can use ternaries in python in while loops when declaring the condition

Jo King

wait, like while x if cond else y:? i guess that makes sense

UndoneStudios

exactly

tio.run/##RY3LCoMwEEX3@YpZJq2Ij12q/…

which is a new revelation

~~we can now take over the whole world~~ we can now golf our code even more

Jo King

i feel like there's usually a shorter way to handle booleans than the native ternary

mousetail

[a,b][c]

UndoneStudios

09:28

@mousetail what does that do?

mousetail

What do you think?

UndoneStudios

looks... non-pythonic to me

mousetail

Yes of course, but it's much shorter than a if c else b

UndoneStudios

@mousetail multiplies 2 lists, making no sense

Jo King

multiply? it's just indexing

UndoneStudios

09:29

I mean

(forget algebra already!)

mousetail

there's no algebra here

UndoneStudios

yeah my bad

mousetail

it's literally just indexing a array

UndoneStudios

my math exam is still in my head

Im coming to terms that I didnt get full

although I topped

besides the tio link is an answer to The Thonnu's most recent sandbox challenge

which I think borrows off yours

mousetail

Doesn't seem related at all

UndoneStudios

09:35

you know, the classical "replicate the x function of language y"

mousetail

That's not what my challenge was like at all, those are bad challenges

UndoneStudios

0

A: Sandbox for Proposed Challenges

Implement Python's range function code-golf Your task Given up to three inputs, \$a, b, c\$, output list(range(a, b, c))*. * range doesn't actually return a list in Python 3, but we'll assume it does. Explanation In Python, the range function can take different numbers of inputs: If there is on...

Its just my opinion

so I may be wrong

mousetail

I dislike those types of questions, since any real function in a real language has way too many edge cases

UndoneStudios

leave a downvote then

besides i think the first rule is useless cuz smart guys will use workarounds

mousetail

CMC Given a number, output the base so that the number in that base has the greatest number of unique digits. If there are multiple output the largest. For example f(12)=10, f(27)=4, f(44)=5, f(101)=8, f(156)=11

UndoneStudios

09:46

I'm having a feeling you've asked this before

mousetail

It was different

The sequences are related though, but kinda inverse

but not quite

UndoneStudios

What's the answer for f(10)

just asking

mousetail

8

or 9

My program says 8 but I think it should be 9

UndoneStudios

what about 1234567890?

mousetail

That would take wayy to long to calculate

0003      2      3
0004      2      4
0005      3      3
0006      4      4
0007      3      5
0008      5      6
0009      4      7
0010      6      8
0011      3      9
0012      7     10
0013      4     11
0014      3     12
0015      3     13
0016      3     14
0017      3     15
0018      4     16
0019      3     17
0020      6     18
0021      3     19
0022      3     20
0023      3     21
0024      4     22
0025      3     23
0026      4     24
0027      4      4
0028      4     26
0029      3     27

left column is lowest equal value, right is highest

UndoneStudios

09:50

Ok this is now getting confusing

so i quit

mousetail

Try yesterdays one then

UndoneStudios

nice try, but I only answer code-golf questions from twitter gadflies

inspired by dilbert.com/strip/2022-11-08

DialFrost

@JoKing hmm but i still dont see how a function can fit there

DialFrost

10:15

It seems to give a reference error :/

Jo King

have you tried declaring the function before you use it?

DialFrost

ya?

Jo King

then i don't see why you would get that error then

DialFrost

10:50

@JoKing i probably did it wrong

Jo King

the last byte is a bit cleverer

DialFrost

oh so its supposed to be 56 after putting the func in? (ignoring the last byte removal)

i kinda just plonked it in :/
for(i=0;i++<1e3;f=n=>n-1&&-~f(n%2?n*3+1:n/2))print(f(i))

what'd I do wrong?

Jo King

> have you tried declaring the function before you use it?

emanresu A

The last clause in for is evaluated after each iteration

DialFrost

oh :facepalm:

@emanresuA i tried putting it in the first clause, doesnt print anything :P, second clause gives error

@JoKing I... er did? Or I just don't understand JS lol

emanresu A

11:03

Swap the print and the declaration

DialFrost

??

ah nvm got it at 56 rn

@JoKing cleverer in what way? :3

@emanresuA smart

Jo King

start with the fact that n-1 can also be ~-n and go from there

DialFrost

hm

On first looks it seems useless, does it have something to do with the && in the middle?

TKirishima

12:03

Hi! What is the bug that you have?
@mousetail thanks for notifying

mousetail

@DialFrost said he had a issue logging in with github

DialFrost

@TKirishima i tried logging in with github and it didnt work :(

TKirishima

Did you have any errors ?

DialFrost

ya

gimme a sec lemme find it

darn i cant log out

TKirishima

So you are logged in ?

If you want to log out you can go on settings + Log out at the bottom of the page

From what I see you already have an account: week.golf/profile.php?id=294

DialFrost

12:09

This page isn’t working
week.golf is currently unable to handle this request.
HTTP ERROR 500

this is when I try logging in with github ^

@TKirishima ye

TKirishima

Well, Ill look at it. I have to go for the moment. Maybe log with SO as it seems it worked last time?

DialFrost

yup that works

DialFrost

14:08

@JoKing i give up :(, I alr know that trick and it doesnt seem to help

huanglx

15:58

Hello! I am new here, so sorry if this question is very basic. When someone golfs your answer further is it standard practice to edit your answer to include the new golfed version?

Radvylf Programs

Yes, usually you'll cross out the old score (with <s>/</s>) and add the new one, then update the code and add a short thank-you to the person who suggested the golf (like "-1 thanks to @[person]")

caird coinheringaahing

@huanglx Yes, if someone suggests a golf to your answer, generally, people will edit their answer to include that. The main time when people don't, is if they believe the new version is distinct enough from their original answer as to merit another another

huanglx

thanks

Steffan

16:37

@DialFrost I just wrote it myself?

I wrote a program that read its own source code?

The filneame is /tmp/code.dart

@DialFrost See this hint

Seggan

17:09

argh ofc i have a great challenge idea while falling asleep and i forget now

Unrelated String

18:39

f

also wait when did you unfuse

user

19:28

They never unfused, it's still one hivemind typing those messages, it's just made two socks. See how they're not interacting with each other? It's because they're the same entity. I say ban both accounts