The Nineteenth Byte

Zacharý

00:03

There, fixed it with a hack.

> #Hackedy hack hack hack

Ugh, it half works. It works when calling the resulting function, but not when actually parsing RAD code.

ASCII-only

@Zacharý :||||||||||||||||||||||||||||\\\

@ngn yes definitely

charcoal has a lot of tests

Zacharý

I think I know why.

ASCII-only

and that's only for like half the functionality

but it's still insanely helpful

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Self-Identifying Coordinates Given an integer n, construct an array with n dimensions of n length, where each element is an identifier of its own coordinates. (todo: make this clearer) For example: n=1 ["1"] n=2 [ ["11", "12"], ["21", "22"] ] n=3 [ [ ["111","112","113"], ["1...

Zacharý

00:20

I totally forgot about a f¨ b

DLosc

00:39

Anyone interested in a game of Codenames over at Puzzling.SE?

Zacharý

I ... just broke ~~addition~~ RAD (didn't push though, thank god)

ASCII-only

@Zacharý :| see this is why you need tests...

Zacharý

I think I need a break.

I'm getting confused which arguments are supposed to be Segments and which ones aren't...

ASCII-only

@Zacharý :|

Zacharý

:|||||||||||||||||||||||||||||||||||||||||||||:

Leaky Nun

00:57

7 hours ago, by Leaky Nun

CMC: Print the last n digits of the 10-adic number that ends ...5781249 that squares to 1.

ASCII-only

@LeakyNun yeah almost done :P

Leaky Nun

:o

ASCII-only

(i'm cheating off wikipedia, but still)

Leaky Nun

sure

Zacharý

01:21

@LeakyNun Third time's the charm?

Downgoat

If anyone's interested, a short article on how VSL does interop

ASCII-only

@LeakyNun crap. it's kinda working now but 5th digit (and above) is off

Leaky Nun

@ASCII-only :(

Conor O'Brien

@Zacharý do you know a nice way of getting the number of characters in a string other than str.byDchar.array.length?

(In D)

Zacharý

nvm

Conor O'Brien

01:25

Try it online!

no

we all make mistakes, you don't need to purge them :P

Zacharý

Didn't see you said "characters"

Just make everything a dstring.

ASCII-only

@ConorO'Brien ~~kinda works~~ ninja'd

Zacharý

import std.stdio;
void main() {
dstring u = "┘";
writeln(u.length);
}

I think there's a suffix as well

Conor O'Brien

ugh

it's a string in D why is it not a dstring by default :U

ty anyway

Zacharý

Glad to help!

ASCII-only

01:30

@ConorO'Brien reads it's actually utf32. which is a really bad idea (i.e. you're supposed to normally use string)

Conor O'Brien

yeah.... so I can casually get cast type misalignment errors. I'm going with the byDchar approach :P

Zacharý

@ConorO'Brien If you need a raw string, "..."d should work

ASCII-only

@ConorO'Brien :P

:|

ninja'd again

@ConorO'Brien MCVE pls?

Conor O'Brien

no

:P I've fixed the error with the other approach

Zacharý

At least with D, we don't get ninja'd by a guy who actually programs in it (solely) for a living :P

Conor O'Brien

01:34

@ASCII-only Try it online!

ASCII-only

@ConorO'Brien fixed

Conor O'Brien

oh

cool

turns out I need the dchar array anyhow but thanks anyway!

Zacharý

You beat me ... to it

Conor O'Brien

ohoho

Downgoat

wow people are going crazy

Zacharý

01:44

How many times has someone mentioned this?

Conor O'Brien

wait why tho e_e

is ms trying to be google

ASCII-only

@ConorO'Brien yes

Conor O'Brien

ahahaha good luck

ASCII-only

02:09

@LeakyNun wait. is the process different to the 7-adic one here since 10 isn't prime

Leaky Nun

right

well the 7-adic case is boring

there's only 1 and -1

ASCII-only

> Suppose that p is an odd prime number

oh.

>_>

Leaky Nun

hi @PhiNotPi

PhiNotPi

hallo

Leaky Nun

9 hours ago, by Leaky Nun

CMC: Print the last n digits of the 10-adic number that ends ...5781249 that squares to 1.

a mathy CMC that nobody has solved yet @PhiNotPi

PhiNotPi

02:21

Sorry, I'm not really feeling like putting in effort right now.

Leaky Nun

ok

ASCII-only

<- putting in too much effort but can't understand anything

like newton's method is easy but i have no idea how to do it with p-adic numbers

PhiNotPi

Right now I just feel like browsing reddit, listening to music on youtube, contemplating my finite place in this infinite universe... you know, the usual.

ASCII-only

@PhiNotPi wow this is too normal for someone on PPCG

PhiNotPi

Even the most irrational numbers belong on the same number line.

2

ASCII-only

02:30

@LeakyNun :|

ngn

@PhiNotPi "the most"? are some more irrational than others?

Conor O'Brien

why not

ASCII-only

@ngn transcendental :P

Leaky Nun

some say that $\sqrt 2$ is the most irrational number

because its continued fraction has the smallest possible integer all over, i.e. 1

[1; 1, 1, 1, ...]

Conor O'Brien

that to me seems to be the least irrational number

PhiNotPi

02:33

That's not sqrt(2), that's phi.

as in, correct fractions

Conor O'Brien

its not pi? :o

PhiNotPi

Lochs's theorem

In number theory, Lochs's theorem is a theorem concerning the rate of convergence of the continued fraction expansion of a typical real number. A proof of the theorem was published by Gustav Lochs in 1964. The theorem states that for almost all real numbers in the interval (0,1), the number of terms m of the number's continued fraction expansion that are required to determine the first n places of the number's decimal expansion behaves asymptotically as follows: lim n → ∞ ...

You can "measure" irrationality by looking at how close an approximation is vs. length of continued fraction. By that measurement, phi is the most irrational. This basically means that phi has the worst tradeoff between size of denominator and accuracy of the fraction.

Leaky Nun

right, phi = [1; 1, 1, 1, ...]

ngn

@PhiNotPi right, in APL that's 1+∘÷⍣≡1 = 1.618033989

sqrt(2) is [1;2,2,2,....]: 1+÷2+∘÷⍣≡2

PhiNotPi

355/113 = pi * 1.0000000849, so pi is pretty rational in some sense. But 183/113 = phi * 1.0008869

Conor O'Brien

02:41

there was a numberphile video on this

The Golden Ratio (why it is so irrational) - Numberphile

►

ngn

@LeakyNun so how can the square of that p-adic number be 1? aren't the only square roots of 1 just 1 and -1?

Leaky Nun

@ngn 10 isn't a prime :)

there are 4 square roots of 1 in the 10-adic land

ngn

@LeakyNun ah...

user202729

02:57

@LeakyNun +1.

Well...

Can I brute force?

Leaky Nun

sure

provided that the answer is correct, that is

and your algorithm must provably halt

just the usual rules

user202729

⁵*Ḷ²_5781249o⁵Ṁ’ in jelly should work.

The literal can be compressed more, and hopefully I didn't misremember any atoms.

Hm...

No...

⁵*Ḷ²’o⁵<³CT’ ...

... _5781249o⁵ <Drei> ÞṪ .

Destructible Lemon

03:16

the wikipedia page for the jackhammer calls it a groundbreaking invention

FrownyFrog

3 hours ago, by DLosc

Anyone interested in a game of Codenames over at Puzzling.SE?

Conor O'Brien

user202729

@ConorO'Brien Who answered 0 ...

Conor O'Brien

idk. someone answered infinity too

I suppose I should trim that 0

user202729

@LeakyNun 4? Should there be 8?

Leaky Nun

03:25

@user202729 nah, just 4

user202729

Hm...

Something wrong with my brute forcer prpgram.

tio.run/##y0rNyan8/z/…

So...

Leaky Nun

4 gives [1, 1249, 3751, 4999, 5001, 6249, 8751, 9999]
5 gives [1, 18751, 31249, 49999, 50001, 68751, 81249, 99999]

there's something inconsistent there :P

ngn

@ConorO'Brien trim it to 37 :)

Conor O'Brien

was that you?

ngn

@ConorO'Brien the infinity - yes

Conor O'Brien

03:29

ah

ngn

03:52

"КУ!" - the most meaningful answer to "oaisdjf?" :)

ASCII-only

04:15

@LeakyNun probably doesn't work?

Zacharý

05:07

He sorted it!

ASCII-only

@Zacharý wait what

Zacharý

docs.google.com/spreadsheets/d/…

Holy crap, 7 15 year olds... jeez

ASCII-only

O_o

~~kids these days~~

4

Mr. Xcoder

And I'm among the 2 14 yos o_O

Zacharý

@ASCII-only Totally agree with you ... DOUGLASSSSSSS (Student Competition)!

@Mr.Xcoder And I'm among the 2 17 yos.

ASCII-only

05:10

@Zacharý you mean 5

Mr. Xcoder

@ASCII-only ~~What the ... are the rest doing with their time?~~

ASCII-only

@Mr.Xcoder --- (lol)

Zacharý

Oh, looked at the wrong one

TFW you get beat by an 11 year old >_<

ASCII-only

@Zacharý wait, what

Mr. Xcoder

@ASCII-only That's what I wanted to type lol and tried to correct it 4 times before I finally suceeded... ~~mobile keyboards these days~~

ASCII-only

05:12

also: i was here two years ago so the insane number of 15-year-olds seems reasonable

Zacharý

I lost to an 11 yr old

dyalog.com/news/112/420/…

I think I was here when I was 15 (PPCG, not TNB)

Mr. Xcoder

@Zacharý lol nice

Zacharý

This time, I'm going to try much harder. (As in, complete all three sets of problems)

Wrong link: dyalog.com/news/119/420/…

AN ELEVEN YEAR OLD.

Mr. Xcoder

I forgot how mind-bending implementing a Python 5-liner or 12-byter in Jelly is in C++... I literally have about 60 lines and am not done yet...

ASCII-only

@Zacharý ~~only a junior last year, disappointing~~

@Mr.Xcoder well at least it's not an interpreter in C++

Zacharý

05:17

WTF is Pythob ?

Mr. Xcoder

@Zacharý ~~Whaaaa?~~

ASCII-only

@Zacharý github.com/topics/pythob

Zacharý

._.

Mr. Xcoder

@ASCII-only No, it's a simple CS olympiad problem which is designed to be very annoyingly simple in nornal languages (eg Python) but you're forced to compete in C(++) :DD:

ASCII-only

@Mr.Xcoder solution: do project euler. you'll never complain again

Mr. Xcoder

05:19

I am doing PE...

ASCII-only

you'll magically find that you'll have created a bunch of macros to make programming PE in C++ easier than in Python

(disclaimer: I have not tried this yet)

offtopic: it's always fun when your laptop has been processing this for a few days straight and it's still not finished

Zacharý

@ASCII-only Macros are fun

Mr. Xcoder

And it's great. I love using the language which is more suited for the job. I've used all of these while solving PE: Python, C, Cpp, Jelly and Pyth

ASCII-only

@Mr.Xcoder ~~use Pypy because it's basically as fast as C++ for a lot of algorithm-focused programs~~

Mr. Xcoder

Yeah of course I use PyPy

Zacharý

05:22

PE?

ASCII-only

Project Euler

Mr. Xcoder

Project Euler

Bubbler

Jelly for PE? Wow.

ASCII-only

@Bubbler Jelly does have an insane number of builtins

Mr. Xcoder

Anyway, I gtg do something I absolutely dislike :( Cya o/

ASCII-only

05:23

i managed to do this in <1s on Pypy IIRC

@Mr.Xcoder and what would that be :P

Mr. Xcoder

@ASCII-only Literature preparation for an exam :c

(Don't ask)

ASCII-only

@Mr.Xcoder rip. good luck, I guess

Mr. Xcoder

Thanks

Zacharý

@Mr.Xcoder Same as well (sleep)

Bubbler

Jelly isn't even on the list of languages on PE

ASCII-only

05:29

@Bubbler why would you restrict yourself to such a small list though

@Bubbler also. what list

Bubbler

The one on the statistics page

And yeah, that's insanely small list, I agree with that

ASCII-only

@Bubbler yeah. 99 languages vs 191 on TIO :P (well maybe only 150 excluding variants)

ASCII-only

05:54

:P
https://tio.run/##zTztctvIkb@lpxjTiQhKECXbe8mFlOSiTNrLKpnykdTtOraXBYFDEjEIsABQEuPozz1Dft7T3YOcr7vnA4MPkpLtrUu2YgEz0z09Pf09A7qLxeHUdb9@feoFrr8cc3bihXEScWd@tpu2zZ2F@RonjvvZbLgOAejI53ee6/gj14mT@myxKI5YJp7vJasjmMELpqOIT8rHOf40jLxkNpcjjwAlH7lhcKPGT8Z8wn7p9tqXvwwM8F@8YBzexvUZjOHB2JvsPl0GOLR3dXGx@xSevICzq3cj667GaK5GIwlHy8WCR4B/sYK1jRsNMeuZZbwAQM1AgPBm76Nx1XZ3j47Y8LJ92WDXTuy5DJYHjA0SNgn9MYxiVhxG0cpmQZgweqylIOOQdRlsk@@vWMD5mCUhW8Ycnm9fMqBxzMJlwryAJTNsvEsYX3hxCDwKJ2wcOdMwgGl9f3c3cOY8XjguZ/9xTnR82d3d4cFyzlzfiWM2SJyEQ@POTu@y/7Z1YcPT66veq2H3sjd61bqghs6v7zqvhoMR/O13BgPoKgwbFcfs7tw3s5MNVws515VAPBi@HdoM…

ASCII-only

06:13

wait. wrong emote. *:O

ASCII-only

07:01

:D shunting yard works too

this is amazing

user202729

08:33

@Mr.Xcoder Then it should be simple in C++ too, except you have to type more.

In en.m.wikipedia.org/wiki/Cornacchia%27s_algorithm :

Where is the "primitive roots" defined? (I understand that it means coprime root, but I cant see that bein defined anywhere

@Mr.Xcoder ... What builtin is that?

How can it be THAT bloated?

Sure C++ is more verbose but not THAT.

ASCII-only

@user202729 i know right >_>

@user202729 maybe he's trying to implement like matrix ops and bigints from the bottom up :P

Mr. Xcoder

08:56

It was about finding the level of similarity between two DNA samples (strings), according to some definitions, and then finding the largest submatrix of neighbors that all have similarity over some treshold to some given sample.

I don't have my Jelly code anymore, though, I could have shown it to you

New Main Posts

10:18

-3

Q: PAPR reduction in ofdm

pls suggest the latest technique or some new algorithm for papr reduction in ofdm needed for my M.tech research work. Thanks

matlab

ASCII-only

10:45

@betseg +1 it's paper (what theme though? ~~clearly anything other than adapta-nokto is bad~~

ASCII-only

10:57

@user202729 it doesn't mention primitive roots though?

ASCII-only

11:17

@LeakyNun interesting (well, I guess it is kinda expected since it's 10-adic)

i.e. they're basically negatives (right?)

Cows quack

@ASCII-only how work

ASCII-only

@Cowsquack idk. i'm assuming I accidentally stumbled on some kind of manual version of Newton's method

@Cowsquack (newton's method for quadratic being n' = (n^2 - square) / 2 I think?)

so what I do to get the next digit is (n/2 / 10^pow)%10 (and then raise it back (* 10^pow) to the appropriate power)

and for some black magic reason it works

Cows quack

11:36

is there a wiki page you used?

ASCII-only

@Cowsquack no. i gave up on that after too many wikipedia pages and SE questions

Cows quack

how is n*n/10**p/2 parsed

ASCII-only

@Cowsquack n*n/(10**p)/2

(* and / are same precedence)

user202729

@ASCII-only Oops bigint...

Tips: Just use InfInt when you can.

ASCII-only

@user202729 infint? but bigint = arbitrary precision?

user202729

11:44

@ASCII-only C++ header only library for bigint arithmetic.

It's just a name :\

ASCII-only

@user202729 ~~clearly gmp is way faster~~

user202729

@ASCII-only And 100% harder to use. Most onlune judges require the src to be in 1 file.

...

About LeakyNun's CMC...

ASCII-only

@user202729 ah. fair enough. there are a lot of exceptions though. e.g. GCJ/PE don't even check the code

user202729

For those which dont check the src just use Python...

GCJ? Not now.

ASCII-only

@user202729 hmm?

@user202729 what if Python isn't fast enough? also yeah. there's about an even split between Python and C++

user202729

11:49

Then Python can't solve it "in 20 lines", and you should not complain anyway.

... the number is = 1 mod 2^n and = -1 mod 5^n.

@ASCII-only I think GCJ start verifying the src.

ASCII-only

@user202729 no...

Cows quack

@ASCII-only shuoldn't it be / (2*n)?

ASCII-only

@user202729 what does length have to do with speed...

@Cowsquack yeah >_> so it's not exactly newton's method then I guess?

betseg

12:12

@ASCII-only adapta-nokto-eta with papirus-adapta icons

ASCII-only

@betseg :D

betseg

use eta, gnome's default window border sizes are too huge

ASCII-only

@betseg I'm not risking bricking my tablet again thank you very much

betseg

oh, eta is a different version of adapta with more acceptable window border sizes, not like a different DE or anything

ASCII-only

@betseg no i think my laptop is allergic to linux

betseg

12:18

tablet or laptop 🤔

ASCII-only

@betseg it's a spectre x2. so i'm using it as a laptop but technically it's just a tablet with a keyboard

betseg

by the way you can copy emoji from gnome's overview mode :D i.sstatic.net/pjPx1.jpg

ASCII-only

@betseg hang on. ew gnome

betseg

helps when im coding in emojicode 🤷

ASCII-only

@betseg i wonder if it's any good? on my old laptop I had unity but unity's overlay sa insanely slow to load

like win10 slow

betseg

12:21

unity is bad

even canonical is discontinuing it

ASCII-only

@betseg ok. great. ~~but gnome is worse~~

betseg

unity is based on gnome

wait, just noticed that that message can be interpreted in both ways

ASCII-only

~~xfce is the only acceptable de~~

betseg

de? pfft

wm is the only way

Bora Calim

guys

im writing a programming language

and it has stacks

infinite stacks

and how should i handle loops

should i only peek at the top of the stack

ASCII-only

12:30

@BoraCalim quite a few languages have that

Bora Calim

or pop that value

which one should i use?

ASCII-only

@BoraCalim some languages have both variants, but if i had to say then i'd go with peek

Bora Calim

ok thx

maybe i may do both

ASCII-only

(it doesn't really matter because the difference is potentially a pop for one vs a duplicate for the other)

Bora Calim

is it turing-complete in both cases?

ASCII-only

12:33

@BoraCalim as long as you have pop and duplicate (or commands that do something equivalent), because with those you can have the exact same loop either way

Bora Calim

i have pop but it does not store the value

it just discards it

ASCII-only

@BoraCalim yeah, that's the version of pop i meant. it's not like there's any other type of pop in a stack-based language anyway >_>

Bora Calim

i dont have duplicate but the language is way too primal rn

ASCII-only

@BoraCalim *primitive

Bora Calim

yes thats the word, thx

Erik the Outgolfer

12:46

another design change? >_<

Anush

hi

ASCII-only

@EriktheOutgolfer wait what

Erik the Outgolfer

codegolf.stackexchange.com/users

Anush

if you have a bit string of length 8, how many distinct bitstrings can you make by inserting 3 bits anywhere you like?

it depends on which string you start with

ASCII-only

@Anush so what's the point of asking >_>

Anush

12:52

CMC Given a bit string of length 5, count how many distinct bitstrings you can make by inserting 3 bits wherever you like into the string

@ASCII-only it's a coding question :)

user202729

13:06

(for LeakyNun's CMC) Good thing is Jelly have modular inverse builtin. ...

ASCII-only

@user202729 will that help though

user202729

20 hours ago, by Leaky Nun

CMC: Print the last n digits of the 10-adic number that ends ...5781249 that squares to 1.

ASCII-only

@user202729 yeah how will modular arithmetic help here

user202729

@LeakyNun Jelly, 15 bytes: Try it online!

@ASCII-only (do you know Jelly?)

Based on the observation

ASCII-only

@user202729 :| no

user202729

13:16

1 hour ago, by user202729

... the number is = 1 mod 2^n and = -1 mod 5^n.

ASCII-only

@user202729 *-1 mod 2^n, and also -1 mod 10^n :P

user202729

So use CRT (Chinese Remainder Theorem) and the number can be computed as (1-2×(2^n × (1/2^n mod 5^n))).

@EriktheOutgolfer Yes of course...

J. Sallé

Hey guys, anyone having trouble with github today? I'm trying to fetch a repository and fetching through https says the repo doesn't exist, while fetching through ssl says the host github.com is unknown. Never had that problem before...

user202729

@J.Sallé Probably Microsoft's fault.

(seriously. Probably just IP change or something. The website work fine for me.)

Erik the Outgolfer

but it hasn't acquired GitHub yet, has it?

J. Sallé

13:20

@user202729 yeah the site works fine, and the repository is there

I've been working on it for months :p

user202729

@user202729 ... why is my repost starred and not the original message?

ASCII-only

@user202729 1/what

Dennis

@J.Sallé Nothing like that, but I stopped getting their email notifications because their SMTP server ended up on a blacklist.

I know this cannot be related to the current news, but it's still a weird coincidence.

New Main Posts

0

Q: Get the sequence steps

Challenge Given a sequence of numbers, create a function which returns the sequence steps. Assume a sequence will be N >= 3 Sequence will repeat it steps at least once Sequence will only contain natural numbers Example: Input: [1, 2, 3, 5, 6, 7, 9, 10, 11, 13, 14, 15] Output: [1, 1, 2] Ex...

code-golf sequence

J. Sallé

@Dennis yeah I'm just slightly freaked out because my whole project is there and I can't fetch it on this new computer >.>

ASCII-only

13:32

@user202729 :| more languages should have an explanation mode

user202729

@ASCII-only 1/2^n mod 5^n.

Modular inverse of 2^n, mod 5^n.

ASCII-only

@user202729 ah.

@user202729 i wonder how Jelly does that so quickly

user202729

Although...

@ASCII-only sympy igcdex.

ASCII-only

@user202729 looks up oh yeah >_> it was on the wikipedia article

user202729

It's not too hard to compute it without builtin, it's just ((5^n+1)/2)^n mod 5^n. Although it would be a bit slower.

Anush

13:40

maybe I should post my CMC on main with variable numbers of bits?

Cows quack

@user202729 is n ...5781249 or is n the prec?

user202729

@Cowsquack The precision.

Cows quack

also how do you know that it is 1 mod 2^n and -1 mod 5^n?

ASCII-only

@Cowsquack by checking

user202729

First, note that the question can be reduced to computing a square root of 1 modulo 10^n.

ASCII-only

13:44

@user202729 :| oh. this makes a lot of sense

Cows quack

13:56

@ASCII-only but you only know that the last digit is 1 mod 2 and -1 mod 5

user202729

If it is 1 mod 2^n and -1 mod 5^n (for n>7), then it's = 1 mod 2^7 and = -1 mod 5^7.

And of course it's a sqrt of 1 mod 10^n.

ASCII-only

@user202729 >_> oh wait nvm it can't be 1 mod 2^n then

@user202729 n>7?

user202729

@ASCII-only ≥ is enough, but that is not wrong.

ASCII-only

@user202729 it wouldn't work with <7?

Transcript for