The Nineteenth Byte

6:00 AM

oh, a register stack? sounds like it could be useful

in Jelly, Dec 13 '15 at 20:07, by Dennis

@quartata I've never written production code before, so sorry 'bout that. Anything in particular you want me to decipher?

this may explain the hacky source code :p

caird coinheringaahing

TIO was launched in 2016 right? That might well explain it :P

Bubbler

Manipulating the register stack could be both implicit and explicit, e.g. explicit push with © and pop with ®, and implicit push on invalid parse

caird coinheringaahing

tbf tho, Dennis is a professional mathematician, not a coder :P

@Bubbler For example?

Bubbler

implicit pop might be annoying at times

hyper-neutrino

@cairdcoinheringaahing yeah, he's just too good at both so it's reasonable to forget xD

caird coinheringaahing

6:03 AM

@cairdcoinheringaahing Nope, TIO is older than Jelly

PyGamer0

i am trying to power on VMWare and it behaves very weirdly; sometimes it works and other times it shouts at me "AMDV IS DISABLED"

caird coinheringaahing

It's very annoying that Dennis had a "Jelly testing" lang on TIO that is now removed, especially as all his links that introduce new jelly features are now on dead TIO links to jelly-testing: jelly-testing.tryitonline.net/…

hyper-neutrino

:/

caird coinheringaahing

@cairdcoinheringaahing Hmm, drop n and take n maybe?

ḣ works as "take n from front" and ṫ is "drop first n", but "drop from end" might be useful

A username

@cairdcoinheringaahing :\

caird coinheringaahing

6:10 AM

@cairdcoinheringaahing Nope, ṫ is "drop first n-1" :/

A username

Which is the creepiest: :} :) :] :D :0

Bubbler

CMC: jelly Produce the array [0, 0, 1, 1]

Jo King

@cairdcoinheringaahing that looks like a pretty old way of formatting the code/input. i'm surprised that part at least still works

caird coinheringaahing

@Bubbler Obvious 4 byter: Ø.x2

hyper-neutrino

@Bubbler 4 lowered range floor divide 2 :p

caird coinheringaahing

6:13 AM

@JoKing TIO is fairly robust with backwards-compatibility. tio.run/#jelly#code=MCsx&input=&args=MTM works fine

(kind of, the decoding is off)

hyper-neutrino

what's the target bytecount

12BU works too

caird coinheringaahing

It should be a fibonacci program, 5 bytes

hyper-neutrino

huh

Bubbler

I knew it was possible in 4, I was hoping to find a 3

hyper-neutrino

just brute force all 16777216 3-byters :p

caird coinheringaahing

6:17 AM

in Jelly, Dec 21 '15 at 3:36, by Dennis

For example, “A”+33 gives the string b, but “A”+“!” gives the integer 98. This is how subtraction typecasts in CJam, and it works pretty well.

Whoever suggested this be removed, I dislike you

hyper-neutrino

wait what does it do now

former errors, latter gives a 2-character-character that breaks everything and isn't usable?

caird coinheringaahing

Try it online!

@hyper-neutrino Basically, yes ಠ_ಠ

hyper-neutrino

cursed

Razetime

why remove da affine space

pain

hyper-neutrino

yuno is a bit wack in that it just stringifies and concats, but that's cuz (i think) i have codepoint shift somewhere else

Bubbler

6:20 AM

Okay, I found a 3 for [0, 0, 1, 1]

wasif

Back again

Bubbler

which makes a 6 for Z matrix :P

wasif

CMQ: hardest golflang

A username

osabie / jelly

osabie has no docs.

wasif

Osabie is easy

Has adequate amount of docs like vyxal

Razetime

6:21 AM

@Ausername what?

hyper-neutrino

docs

@Bubbler wtf

Razetime

@hyper-neutrino docs in the wiki

caird coinheringaahing

05AB1E doesn't need docs beyond a list of builtins

hyper-neutrino

thing is stack based languages just need a builtin list and you're good to go

wasif

You can just string few builtins together in stack basef langs and get jobs done

hyper-neutrino

6:23 AM

every function can be looked at in total isolation based on the state of the stack

wasif

Wdym by total isolation

hyper-neutrino

like, you don't need to look at what's around it in the code

wasif

I see

hyper-neutrino

+ adds pop() and pop() no matter what

wasif

Yes

A username

6:24 AM

Yes

caird coinheringaahing

in Jelly, Jan 28 '16 at 22:15, by Dennis

Jelly also has ones, and [1,0,2,0]O gives [1,3,3]. However, I implemented that because J has it.

what? how does [1,0,2,0] go to [1,3,3]?

wasif

Accidentally typed

Adám

@cairdcoinheringaahing Replicate each index by the number.

Bubbler

@cairdcoinheringaahing [i-th element] copies of [i]

hyper-neutrino

CMC: Replicate each index by the number (that is, given a list of non-negative integers, for each index i, i should be repeated a[i] times). 0- or 1- indexing allowed.

Adám

6:29 AM

@hyper-neutrino APL: ⍸

hyper-neutrino

oh so there is a primitive

i was trying to work it out with /, ⍳, and ≢ but couldn't seem to get it to work

Adám

Where.

@hyper-neutrino That's the classic idiom: {⍵/⍳≢⍵}

Bubbler

@hyper-neutrino APL: ⍸, J: I., K: &:

wasif

All APL kids go together

hyper-neutrino

ah. i was trying to use trains even tho a dfn(?) would be shorter but ¯\_(ツ)_/¯

Bubbler

6:30 AM

Absolutely zero reason to omit the same thing in Jelly :P

Adám

@Bubbler You forgot BQN: /

hyper-neutrino

Jelly: Jx I think, unless there is a single atom for this

Adám

@hyper-neutrino /∘⍳∘≢⍨

Unrelated String

pretty sure jelly does not have a single atom for it

hyper-neutrino

ah. thanks. yeah i had some combination of those characters but in the totally wrong order then :p

caird coinheringaahing

6:32 AM

@hyper-neutrino Jelly, 5 bytes ḟ0x@T

hyper-neutrino

you don't need to filter 0s from both sides because repeat by 0 just skips it anyway

so you can golf that to x@J

which is hopefully obviously golfable again :p

caird coinheringaahing

@hyper-neutrino I'm surprised that doesn't output [1, [], 3, 3, []]

Try it online!

hyper-neutrino

well, [1, 2] ; [] is just [1, 2]

caird coinheringaahing

x is implemented as

Unrelated String

i feel like gracefully replicating by 0 is a very big part of why we even have x

caird coinheringaahing

6:34 AM

lambda x, y: rld(zip(x, y if depth(y) else [y] * len(x)))

Adám

CMC: Replicate each multidimensional index by the number (that is, given a multidimensional array of non-negative integers, for each index (i,j), (i,j) should be repeated a[i][j] times). 0- or 1- indexing allowed. E.g. [[1,0,2],[0,1,0]] gives [[0,0],[0,2],[0,2],[1,1]] or [[1,1],[1,3],[1,3],[2,2]].

Unrelated String

in the sense that i vaguely recall the apl/j equivalent of x being the main way of filtering things

Adám

Indeed.

caird coinheringaahing

@Adám I might just be very tired rn, but I have no idea how [[1,0,2],[0,1,0]] maps to [[1,1],[1,3],[1,3],[2,2]]

Adám

@cairdcoinheringaahing 1 (1,1), 0 (1,2), 2 (1,3)s, etc.

caird coinheringaahing

6:39 AM

@Adám Jelly, 8 bytes ŒJWẋ¥"FẎ

Unrelated String

I was just about to post my 7 byter :P

caird coinheringaahing

Ooh, nice use of ŒĖ

Unrelated String

thanks

it was my first thought, then i realized i could sidestep W with it as well

caird coinheringaahing

Shame that Jelly doesn't smash-print [[1, 1], '', [[1, 3], [1, 3]], '', [2, 2], ''] tho :/

@UnrelatedString 6 bytes in my fork ($€ -> ɱ) :P

Bubbler

That's very irregular to my eyes, I didn't imagine a built-in could reliably convert it to a list of pairs

Unrelated String

6:42 AM

smash printing would give 11131322 which doesn't quite exhibit the structure we have to show

Adám

Btw, in APLs, the built-ins for the simple list case of course works for the multidimensional cases too.

caird coinheringaahing

Yeah, I was hoping it would collapse the []s into [[1, 1], [1, 3], [1, 3], [2, 2]] (basically, "hiding" the [] and collapsing the nested lists into a single one)

Bubbler

@Adám ⍸ works in APL and &: works in K, but J's I. has rank 1 and therefore fails to do so :/

caird coinheringaahing

CMC: What should 1+1× do for an input of 3?

Bubbler

@cairdcoinheringaahing Assuming LTR, it's natural to me to evaluate to (1+1)×3

wasif

6:48 AM

CMQ: What might be a nice overload for multiplication operator when operands are both strings

Adám

@Bubbler Oh, I didn't realise that. Always strikes me as strange when J fails to generalise (right) to higher-rank cases. Another example is i. which is just $i.&(*/) (i.e. ⊢⍴∘⍳×/) rather than mulltidimensional indices.

caird coinheringaahing

in Jelly, Feb 2 '16 at 14:49, by Dennis

The monadic link 1+1× used to return 6 for input 3. 1 set the return value to 1, +1 incremented (yielding 2), and × performed a hook multiplication (yielding 6).

Adám

@wasif Apply the case of one to the letters of the other.

caird coinheringaahing

in Jelly, Feb 2 '16 at 14:50, by Dennis

With the change, it returns 4 instead. 1+ increments (yielding 4), and 1× multiplies by 1.

wasif

@Adám nice

hyper-neutrino

6:49 AM

wait, what

i thought a nilad at the start of a monadic chain set the value

is that only for LCCs?

Unrelated String

wait what was the original parsing rule that caused 6

hyper-neutrino

yep it's only for LCCs

caird coinheringaahing

in Jelly, Feb 2 '16 at 4:54, by Dennis

This came with a minor change though, to keep things consistent: a chain of atoms with arities 0 2 0 2 is now parsed as (0 2) (0 2), i. e., two atops, instead of 0 (2 0) 2, set return value, atop, hook.

Unrelated String

yeah i guess it either special cased nilads at the beginning of any chain or it

yeah that

parsed 2,0 before 0,2

reminds me i need more ideas on all the things lone nilads could do other than just replacing the running value

caird coinheringaahing

in Jelly, Jun 1 '17 at 15:47, by Dennis

I am considering altering the behavior of unparseable nilads though. The implicit print is rarely useful, and instead of printing these, these values could be stored on a stack, and pop'd to be used instead of the rightmost argument for coming dyads.

A username

6:54 AM

CMC: Interpret S10K

wasif

what is S10K

Bubbler

Ignore everything and output 10k copies of the character S

A username

A programming language where any program prints 'S' 10000 times.

ninjad

wasif

oh nice

A username

@lyxal Hi!

caird coinheringaahing

6:57 AM

5 bytes in Jelly

wasif

@Ausername Vyxal 5 bytes: \sk2*

hyper-neutrino

probably the same thing as Jelly but with yuno's builtins: ”Sxᴇ4

wasif

Python: 's'*10000

I tried 10e3 first

A username

Vyxal H, 4 bytes: ²\s*

wasif

but it returned a float

A username

6:59 AM

@wasif Python...

wasif

@Ausername hey what does the H flag do

Bubbler

Yeah, Python's e always gives a float

A username

Puts 100 on the stack'

wasif

lol

hyper-neutrino

oh finally that flag actually sees some legitimate use other than cheesing FB

6

wasif

6:59 AM

flags for everything

A username

@hyper-neutrino I've used it several times

lyxal

@hyper-neutrino except I don't use it for FB lol

Bubbler

7:35 AM

BMG starts in 30 minutes!

Razetime

mark as announcement pls

AviFS

7:53 AM

BMG is an hour, right?

Hey @LeakyNun!

Razetime

not exactly

AviFS

Oh?

Razetime

it's supposed to be an hour i think

but it goes about as long as we have drafts to answer

AviFS

Ah, true

Just hoping for a sense. Anyone know how long was it last time?

Bubbler

It goes like between 60 and 90 minutes

AviFS

7:56 AM

Thanks!

Sandbox Posts

0

A: Sandbox for Proposed Challenges

Is this an ordinal transform? code-golf math decision-problem integer Related: What's my telephone number? which asks to calculate the terms of A000085, the number of possible ordinal transforms of length n. Background Ordinal transform is a transformation on an integer sequence. For a sequence \...

AviFS

I'll try and be here after all, then! Though I really should be sleeping.

I'm in the Nineteenth Sound Byte in the CGCC Discord, if anyone would like to join!

pxeger

BMG o'clock!

Bubbler

Welcome to the 5th Biweekly Mini Golf! During this event, we'll post some CMCs (Chat Mini Challenges) for you all to solve. A new one will be added every 5 to 10 minutes. If you have any questions, don't be afraid to ask. Good luck!

3

in BMG Drafts, 7 hours ago, by Bubbler

Draft: First n terms of 1, 2, 1, 2, 3, 2, 3, 4, 3, 4, 5, 4, ... which has a pattern of [1, 2, 1], [2, 3, 2], ...

Razetime

ngn/k, 16 bytes: {x#,/1 2 1+/:!x}

pxeger

8:06 AM

@Bubbler Zsh, 33 bytes: for ((;++x;))eval '<<<$['{,1,}+x]

AviFS

~~APL: Playing around with {0 1 0+¨⊂¨⍳⍵}, but can't get it to work~~ Totally wrong

Razetime

0 1 0∘+¨

and you don't need the ⊂¨

Bubbler

Or (⊂0 1 0)+⍳⍵

Razetime

yeah, that's the more APL way of doing it

Bubbler

{⍵↑∊(⊂0 1 0)+⍳⍵) should work I think

pxeger

8:08 AM

@Bubbler looks like a kaomoji

AviFS

@Bubbler I was enclosing the wrong side!

Bubbler

@pxeger Is there a way to quit execution early?

pxeger

@Bubbler ato.pxeger.com/…

that'll have to do for now

Bubbler

Oh ok

pxeger

I'm working on a stop button :)

It's surprisingly complicated to implement

ovs

8:09 AM

05AB1E, 7 bytes: ƵKSλ£₃>

Bubbler

in BMG Drafts, 7 hours ago, by caird coinheringaahing

Draft (inspired by a builtin I just added to my Jelly fork): Given a list (even length) of integers, split into consecutive pairs, sort each pair by its sum, then flatten. e.g. [2,3,6,4,2,1] -> [[2,3],[6,4],[2,1]] -> [[2,1],[2,3],[6,4]] -> [2,1,2,3,6,4]

att

@Bubbler Mathematica, 30 bytes: Join@@#~Partition~2~SortBy~Tr&

Bubbler

@att I always wondered about ~ parsing. Looks like it's left-associative (kind of)

AviFS

@Bubbler Brainfuck: +<+[.>+.<.+]

We're supposed to be golfing these, right?

pxeger

@Bubbler Python, 52 bytes: lambda l:sum(sorted(zip(l[::2],l[1::2]),key=sum),())

Bubbler

8:14 AM

@AviFS Yes

@AviFS Do you really need <>? :P

Razetime

@Bubbler ngn/k, 20 bytes: {,/{x(<+/'x)}0N 2#x}

AviFS

@AviFS Golfed 1st challenge in BF: +[.>.+<.+]

ovs

@Bubbler APL, 19 bytes: {⍵[⍋2/+/¨⍵⊂⍨2|⍳⍴⍵]}

I feel like there is something shorter for ⍵⊂⍨2|⍳⍴⍵

att

@Bubbler ~ ~ is, yes

Bubbler

@AviFS That's wrong, I think +[.+.-.+] should do

Jo King

8:16 AM

^

AviFS

@Bubbler Oh... that's good. But I don't think mine is wrong. If you look at the section where it prints numbers, you'll see 010121232...

att

@Bubbler Mathematica, 22 bytes: #+2#~Mod~3&~Array~#/3&

Bubbler

@ovs The only other alternative is to reshape, but (.5×≢⍵)2⍴⍵ isn't particularly short either

in BMG Drafts, 8 hours ago, by AviFS

Draft: Output a magic square with the input number in one of the cells.

att

what size square?

Bubbler

@AviFS I guess you mean 3x3 magic square?

pxeger

8:20 AM

3x3 I'd assume?

Razetime

size 1x1 lol

AviFS

@pxeger Interrupting execution may be nontrivial, but allowing the user to change the timeout should be an easy temporary fix.

pxeger

hmm I hadn't thought of that

AviFS

@Bubbler I thought that initially, but then I decided any size should be allowed to make it interesting...

@Razetime Then again... this. Good one!

I say 3×3 or greater.

Doesn't need to be consistent across inputs

@Razetime With Razetime's idea, we do have an honorary 0 byter in the implicit IO langs

Razetime

ok so wait

pxeger

8:23 AM

Polyglot, 0 bytes

Razetime

should it theoretically terminate before the end of the universe

Bubbler

3x3 is still easy: 7 0 5 2 4 6 3 8 1+⊢

Time to compress that digit vector :P

AviFS

@Bubbler Wow... that was clever!

Razetime

cause the general short solution would be generating random 3x3's until one of them is a magic square and contains said number

AviFS

I thought one would have to do something like:

betseg

8:25 AM

@Bubbler Rust, 86: |n|{let mut m=-1;std::iter::repeat_with(move||{m+=1;[m+1,m+2,m+1]}).flatten().take(n)}

AviFS

n+i  1     2
3    4    n+i
5   n+i   6

Razetime

@betseg memelord has arrived

betseg

im a bit rusty

Bubbler

@AviFS That doesn't quite work, but there's indeed a general solution for 2 or 3 givens using a pattern like that

AviFS

No, it doesn't quite work, I know. But I thought something like that. Yours is really clever!

We can move on! Between @Razetime & @Bubbler I think we're good on this, haha

Bubbler

8:29 AM

in BMG Drafts, 7 hours ago, by Bubbler

Draft: Given an ASCII string, double each single quote and enclose it in single quotes. "How'd'n't I see that" -> "'How''d''n''t I see that'"

in BMG Drafts, 7 hours ago, by Bubbler

Draft: Given a rectangular 2D string of odd width and height, place the character X at the center.

Posting two because I'm briefly going AFK now

pxeger

@Bubbler the double-quotes on the outside are not part of the input, right?

Bubbler

@pxeger Correct

pxeger

is I/O as a list of codepoints ok?

Bubbler

Yes

Razetime

@Bubbler ngn/k, 20 bytes: {a,((1+a=)#x),a:"'"}

@Bubbler this can be a list of strings?

Bubbler

8:33 AM

@Razetime yes

Razetime

hm this is not very obvious in K

pxeger

@Bubbler Zsh, 25 17 bytes: <<<"'${1//'/''}'"

Bubbler

@Razetime Likely easier if you take it as a newline-separated string instead

Razetime

@Bubbler V(vim), 4 bytes: Try it online!

AviFS

@Bubbler Vim:i'<esc>qqf'l.q9999@qA'

I'm sure V has some loop thing to make 9999×macro less silly

pxeger

8:41 AM

@AviFS you're missing an escape there

AviFS

True, thanks

@pxeger Fixed!

@Razetime Nice!

Razetime

@Bubbler yeah probably @[;;] but no clue on what to do

pxeger

alternatively, i'<esc>A':s/'/''/g<cr>

AviFS

@pxeger True, but why do that...

pxeger

cos it's shorter and works with more than 9999 apostrophes?

AviFS

8:43 AM

Whatever :P

pxeger

@pxeger also missing an escape, and needs to be done in the other order

Bubbler

in BMG Drafts, 7 hours ago, by DLosc

Draft: Given N (0- or 1-indexed), output the Nth element of this sequence:

in BMG Drafts, 7 hours ago, by DLosc

""
\"""\"
\"\\"""\\"\"
\"\\"\\\\"""\\\\"\\"\"
\"\\"\\\\"\\\\\\\\"""\\\\\\\\"\\\\"\\"\"

in BMG Drafts, 7 hours ago, by DLosc

Brownie points to whoever finds the 7-byte Pip answer. ^_^

pxeger

:s/'/''/g<cr>I'<esc>A'

probably something shorter with ctrl-O

AviFS

@Bubbler Another lame Vim: 3i 1<esc>b<c-a>qqYp<c-a>w<c-a>w<c-a>q9999@q

Razetime

Pip, 8 bytes: Lax:RPxx Try it online!

@DLosc idk what black magic you've done to get 7

betseg

8:53 AM

@Bubbler Rust, 85: |a:&[i32]|{let mut v=a.chunks(2).collect::<Vec<_>>();v.sort_by_key(|e|e[0]+e[1]);v.concat()}

Bubbler

@betseg Isn't v:Vec<_> shorter than the turbofish?

in BMG Drafts, 7 hours ago, by Bubbler

Draft: Given a string, output which-th time each character is appearing in that string. "abracadabra" -> 1, 1, 1, 2, 1, 3, 1, 4, 2, 2, 5

Razetime

@Bubbler 7 bytes: Try it online!

gotteeem

Bubbler

Lol nice

betseg

@Bubbler oh yeah

v:Vec<_>=a.chunks(2).collect() for 83

Bubbler

Under CGCC rules you can also cheat and replace i32 with i8 :P

betseg

8:57 AM

true lol

i misread mb

Razetime

@Bubbler APL, 28 bytes: {a⊣{a⊢←(⍋b)@(b←⍸⍵=a)⊢a}¨a←⍵}

Transcript for