The Nineteenth Byte

@user202729 As far as I can tell dyalog uses the MSB of a word for the first bit in a vector. Also I'm not sure why you're bringing popcount into this, are you suggesting that the algorith they use would require it?

Actually you can just shift right instead of left, for a different bit order

@user202729 How does this work?

nvm, I figured it out. Pretty neat

Leaky Nun

1:42 AM

@ASCII-only fixed

:|

what

you improved the score

Leaky Nun

isn't really statistically significant

2:14 AM

@H.PWiz Ah...right. Computing popcnt is not very necessary.

Still, your method seems to me like it would be faster. although I am rather uninformed on the matter

@H.PWiz link

@ASCII-only I'm talking about the xor-reduce of a word

@H.PWiz xor-reduce?

ah

xor

2:25 AM

@H.PWiz sorry typo >_>

@H.PWiz yeah that would be faster

it's basically like the black magic of parallel reduce

except on bits

Context if you missed it

I really meant scan when I said reduce...

Downgoat

4:42 AM

CMC: given a string, output the shortest possible printable-ascii sequence that doesn't appear in that string

they are multiple correct outputs for every input

5:23 AM

@user202729 No, one less step than the number of elements (along the trailing dimension). The first element of the result is always the unaltered first element (along the trailing axis) of the argument, and the function isn't called at all for that one. (Btw, to go along the leading axis, use ⍀.)

ATaco

I have plans to make a Sequel of sorts to Funky.

Called Funkier.

It's going to be so wonderfully terrible.

@Adám That question was referring to the number of steps for a machine word. With popcount being the number of 1s

5:39 AM

Could a mod unfreeze the Brachylog room (chat.stackexchange.com/rooms/34284/brachylog)?

@Mego ^

@H.PWiz Oh, I didn't realise that. So it isn't my department. I generally have no clue what goes on below the surface…

@user202729 My misunderstanding of your question was corrected by @H.PWiz ^^^^ I suggest you leave questions on the blog page or email the author directly: marshall@ (domain should be obvious).

@ATaco Most people call those "updates"

ATaco

Most updates don't change core features of the language.

Or rewrite the language in a different language

Well, any self-hosting compiler had to go through a port at one point

and Python 2 -> 3 was a pretty big change

While I don't know much about Perl, I hear v5 was a pretty big rewrite and changed quite a lot of things

ATaco

Eh, Funky 2 doesn't feel right for it though.

6:04 AM

Then call the original Funky v1 and the new one Funky

that's what TIO does with Brachylog

(The language "Brachylog v1" has url tio.run/#brachylog, and "Brachylog" has url tio.run/#brachylog2)

7:29 AM

@EsolangingFruit However, Seriously v2 is Actually.

7:51 AM

@ATaco I think that's what recently happened with Jelly, but it's still Jelly :)

@EriktheOutgolfer Were those changes not purely additive?

@Adám the code page was changed (so your answer over meta might need to do so as well...), there's now something called "aliases" (i.e. shorter versions of already-existing primitives, e.g. ) -> µ€), and there have been many atom and quick additions...btw, Dennis is in the process of de-aliazing Jelly to some degree :P

btw I'm not sure as to why ) is considered to be an alias, since it's actually a chain separator on its own (µ€ isn't one but two tokens)

@EriktheOutgolfer If the code page has changed, doesn't that mess up base-250 encoding?

@Adám ØJ has changed too

that, since ØJ is set to be the codepage variable and not a constant string

@EriktheOutgolfer Hehe, Jelly adopting ⎕AVU-like functionality.

7:57 AM

@Adám no

that's ØJO¤, not a built-in

and no, you can't assign anything to ØJO¤, because 1) it's not a variable 2) the only variable is the register

@EriktheOutgolfer ⎕AVU is not ⎕UCS ⎕AV. ⎕AV is ⎕UCS ⎕AVU, and ⎕AVU is a settable variable.

@Adám ninja ^^

8:19 AM

@EriktheOutgolfer Thank you for editing. Perfect timing, as I was just about to begin.

@Adám It should be 'actually v2 is seriously'?

@user202729 Oh, got them mixed up then, but the point stands.

@EriktheOutgolfer But they're equivalent? (except in string literal)

@Adám Why?

A: Sandbox for Proposed Challenges

many built-ins can be implemented while not using them, that doesn't make them "aliases", does it?

New Sandboxed Posts

0

Indexing Cha-Cha Slide code-golfnumber Introduction: Although I originally had a Dutch song in my head, where the lyrics are: "Do 'n stapje naar voren, en 'n stapje terug" (which translated to "Take a little step forward, and a little step back"), when I searched for the full lyrics, I realize...

8:23 AM

I think "alias" is a term for single-byte tokens which have taken the place of deprecated double-byte ones

@user202729 Well, not if just glyph-swaps, and the Unicode-to-Jelly converter considers both original and new chars the same. But I didn't know that.

@Adám also, the "Unicode-to-Jelly converter" is kinda messy in a few places (e.g. ¶)

I had to open an issue so as to make ⁾¶¶ behave as it should

8:50 AM

I have no idea why the "What..." message is starred...

4

9:01 AM

CMC Design an algorithm to solve Downgoat's CMC above in linear time.

Bonus if you don'y use suffix tree.

Q: ASCII art uncompression from a base-n number

10:17 AM

@user202729 Marshall says that that was their previous algorithm, and that the one outlined in the blog is substantially faster. Part of the problem is that the operations can't be pipelines. I emailed him after having previously corrected a typo in the post

New Main Posts

11:59 AM

2

This is inspired by an 05AB1E answer by Magic Octupus Urn. Given two arguments, a positive integer and a string/list of characters: Translate the number to base-n, where n is the length of the string. For each character, replace every appearance of the index of that character in the base-n num...

Anonymous

1:13 PM

@EsolangingFruit Done

New Main Posts

2:18 PM

4

Q: The Unique Padlock PIN List!

Introduction In a private chat, a friend of mine apparently recently stumbled across a security system which has the following two restrictions on its valid pins: Each digit must be unique (that is "1" may only appear once) The order of the digits doesn't matter ("1234"="4321") So to illustr...

code-golf combinatorics

SEJPM

^ I'm around if the need arises

2:31 PM

@NewMainPosts CMC Formula for the length of output n

⁵c

Hat Wizard

3:03 PM

@DJMcMayhem Haskell 22 bytes

Poke

3:22 PM

did you guys ever settle on the difference between bigO and bigOmega

bigO is an upper bound while bigOmega is a lower bound

that's all

so if i wanted to say that an algorithm takes "at least" some amount of time

that's bigomega

4:16 PM

How does /usr/bin/read work? How do you get the read result?

read var;echo $var?

@Cowsquack The shell builtin works like that. /usr/bin/read doesn't , try it.

$ which read
/usr/bin/read
$ /usr/bin/read q;echo $q
asldhjasljll

what, why do they behave differently despite pointing to the same thing?

@Cowsquack Witchcraft, says I.

@Cowsquack Doesn't work for me

4:24 PM

which part doesn't work?

[pavel@localhost ~]$ which read
/usr/bin/read
[pavel@localhost ~]$ /usr/bin/read q; echo $q
foo

[pavel@localhost ~]$

that's how it behaves on mine too (notice the trailing empty newline)

Yes, exactly

@Pavel And it can't. Calling an external command cannot modify a variable.

(Sorry I said that a bit weird)

@Dennis Exactly. So what's it for?

Same with /bin/alias

4:27 PM

Beats me. openSUSE doesn't have it.

$ cat `which read`
#!/bin/sh
builtin read "$@"

wat

mine is slightly different

$ cat `which read`
#!/bin/sh
# $FreeBSD: src/usr.bin/alias/generic.sh,v 1.2 2005/10/24 22:32:19 cperciva Exp $
# This file is in the public domain.
builtin `echo ${0##*/} | tr \[:upper:] \[:lower:]` ${1+"$@"}

Should we panic?

Mine is the same as Dennis'

(TIO's)

Oh yeah that would be the Fedora one.

4:30 PM

Apparently, it came with the bash-4.4.19-2 package.

@Cowsquack this file is the same for other builtins, like alias

Wait, FreeBSD?

bugzilla.redhat.com/show_bug.cgi?id=698670

@Pavel macos

Seems to be the result of an (overly literal) interpretation of POSIX.

4:36 PM

@J.Sallé We can still do CMCs in TNB, so we'll survive.

I guess /usr/bin/read has at least some use cases, even if none of them involves storing user input in a variable. /usr/bin/cd could be used to test if we may cd into a directory.

@Adám yeah actually it happened for like 40 seconds on Boardgames.SE

You could run a debugger on /usr/bin/read in order to retrieve the parsed input :P

That's insanity, not a use case. :P

> You can use GDB to debug programs written in C, C@t{++}, Fortran and Modula-2.

What is this "C@t{++}" language?

#!/usr/bin/env pwsh
param ( [string]$Variable )
Set-Variable -Name $Variable -Scope 'global' -Value ([Console]::ReadLine())

PowerShell read script works :P

PS>read foo
bar
PS>$foo
bar

5:05 PM

Unix shells: We have two output streams. Output, and Error.
PowerShell: We have 6 output streams. Output, Error, Warning, Verbose, Debug, and Informational

@Pavel can't say MS is not thorough with their outputs

The last 4 are supressed by default and you need a flag to enable them, so you don't really have to worry about them generally.

Actually, IDK about Informational. I think that one is shown anyway.

It's not used much.

My knowledge of powershell goes as far as "it's a way to use linux commands on windows"

@J.Sallé For me it's a way to use Windows commands on Linux

@J.Sallé That's WSL

5:15 PM

They've been putting a lot of the "unixy" commands into stock CMD anyway

You've got curl and tar and cat and whatnot

Although nothing has grep yet

@DJMcMayhem no idea what that is

Windows Subsystem for Linux

Windows Subsystem for Linux (WSL) is a compatibility layer for running Linux binary executables (in ELF format) natively on Windows 10. WSL provides a Linux-compatible kernel interface developed by Microsoft (containing no Linux kernel code), which can then run a GNU userland on top of it, such as that of Ubuntu, openSUSE, SUSE Linux Enterprise Server, Debian and Kali Linux. Such a userland might contain a Bash shell and command language, with native GNU/Linux command-line tools (sed, awk, etc.) and programming language interpreters (Ruby, Python, etc.). When introduced with the Anniversary Update...

@Pavel pisses me off every time I type ls into CMD and it throws an error

quartata

Powershell is a shell

@DJMcMayhem sounds pretty cool actually

5:18 PM

@quartata It's a really nice shell :P

@J.Sallé It's like Wine, but backwards

@Pavel That is very subjective.

Exactly what I thought

quartata

ksh is better

@Dennis It has braces instead of fi/done/esac crap, ergo it is better.

Speaking of powershell, I have a new favorite command: w | !powershell.exe "python.exe -u test.py \| tee output.txt"

(from vim)

FreezePhoenix

5:21 PM

hello

Also, thanks @Dennis for informing me of -u

@DJMcMayhem What does w do on Windows?

It's a vim command

@Pavel I think that's a vim "save" command

Oh. I thought it was a shell command.

My bad.

5:22 PM

It's (w)rite (|)and (powershell ...)

@Mego Thanks

@Pavel I'm not a friend of esac, but it's a small price to pay.

@DJMcMayhem I downloaded gvim on a whim and forced myself to learn how to use it

Gotta say it's pretty neat

:D

@Dennis For what? I'm actually curious what anyone likes about bash. I feel like it's just a mess of weird syntax.

5:26 PM

Also, I found out the other day that if you do an s/something/somethingelse on discord it will edit your latest message.

Skype for mac supports full sed I hear

I typed it as a joke to a friend and I got confused when it didn't come through, then I saw my message had been edited

@J.Sallé it's annoying that discord thinks the ending / is a part of the replacement

@Pavel Maybe pwsh is just too different to what I'm used to. Set-Alias is a lot more complicated then alias, yet it doesn't allow you to do simple things like `alias lsl='ls -l' without creating a function.

@Dennis Oh yeah, good point. I have a really hacky solution to the alias problem.

Here's a file called ~/bin/pwsh-aliases/alias.sh:

#!/usr/bin/zsh

COM="$(echo $0 | sed 's/.*\///')"

case $COM in
	"ls" )
		exec exa --git -Fh --group-directories-first "$@"
		;;
	"l" )
		exec exa --git -h --group-directories-first -Flah "$@"
		;;
	# More aliases...
esac

ls, l, etc. are files in the same directory that are symlinked to alias.sh.

And my $PROFILE contains

Set-Alias l  /home/pavel/bin/pwsh-aliases/l
Set-Alias ls  /home/pavel/bin/pwsh-aliases/ls
# More aliases...

So I wrote a function to register a new alias

function Register-Alias ($Name, $Command) {
	pushd "~/bin/pwsh-aliases"
	sed -i "$ d" alias.sh
	"`t`"$Name`" )" >> alias.sh
	"`t`texec $Command `"$@`"" >> alias.sh
	"`t`t;;" >> alias.sh
	"esac" >> alias.sh
	ln -sf alias.sh $Name
	"Set-Alias $Name /home/pavel/bin/pwsh-aliases/$Name" >> $profile
	popd
}

(`t is tab, `" is escaped quote)

5:38 PM

I'll stick with esac. :P