The Nineteenth Byte

Fun fact: For some reason I didn't know global variables were a thing when I first learned JavaScript, so I'd store any information I needed to keep between functions as invisible text somewhere in the website

Lyxal

2:35 AM

@RedwolfPrograms so someone could have theoretically screwed up your programs with a userscript/clever use of dev tools?

Redwolf Programs

Yes, although that's possible with global variables anyway :p

Also I released an alpha for Ash if anyone wants to mess with it and try to see if you can make it error

Lyxal

@RedwolfPrograms where?

Redwolf Programs

Here, although it's linked in the chatroom (with some explanation)

Lyxal

Online interpreter when?

Redwolf Programs

When I add all the operators in

It's browser JS though so it's not like it's too annoying to run :p

That's maybe the one advantage of programming things in JS

Mitch Schwartz

3:22 AM

wait, rolling back within 5 minutes locks in the edit and creates a new one? that seems a little weird, oh well now i know

oh i did manage to find it actually documented though meta.stackexchange.com/a/21789/154429

ChartZ Belatedly

@MitchSchwartz Rolling back isn't editing, so it ignores the grace period, unfortunately :/

Hi all

howdy

What’d I miss?

@user Ash is almost finished!

user

3:35 AM

Cool!

Congratulations on your first golflang! (or new golflang if it isn’t your first)

Redwolf Programs

At the moment most of operators are implemented, I just need to map them to the correct code points and add in all the constants

ChartZ Belatedly

@NewMetaPosts Reminder to all: don't migrate questions that'll be closed

2

Redwolf Programs

Why's that?

user

@ChartZBelatedly That one could’ve been avoided maybe with a better introduction

ChartZ Belatedly

There's absolutely no point in migrating a question to Meta if it's going to be closed there, and doing so contradicts one of the points in the "When should I migrate posts?" Mother Meta question (on mobile, so no link). Instead, close it on main and direct the user to the relevant meta post(s)

Migration is supposed to be a way of saying "This is on-topic on site X, so we're moving it there", not "This is more relevant to site X, but whether it's high enough quality to remain open isn't important"

@user As is the case with most off-topic, new user posts

user

3:45 AM

The “recommend deletion” reasons on vlq answers may need to be changed too - one is “I’m having this problem too” (note in case we can get it changed in the future)

Redwolf Programs

Does anyone actually use those? The default comments are just annoying.

And what even happens when you "recommend deletion"?

Apparently six of them deletes it

ChartZ Belatedly

@RedwolfPrograms On occasion, when the autocomment actually fits. But not often

Mitch Schwartz

4:03 AM

the "golf languages trivialize everything" misconception seems weird because it doesn't even make sense superficially, like if you look at a golf-language solution to a sufficiently complicated problem, it still looks complicated on the surface lol

5

Redwolf Programs

My suggestion to anyone who believe that about golfing languages is to try using (or making) one.

user

Even though you’re going to have a short answer with a golflang, there’s still a lot of approaches you can choose between to make it as short as possible.

And for longer solutions, it’s very hard to keep track of what you’re doing because it’s so unreadable

Redwolf Programs

To me a golfing language doesn't trivialize a problem, unless it's already trivial. It just abstracts away the implementations of various common operations and lets you focus on the fun part of golfing.

2

Lyxal

4:18 AM

MS Word makes for a very good golfing-language text editor

especially Jelly

With its smart quote replacement and all

And ability to auto-replace sequences with predefined characters

Razetime

5:09 AM

@Lyxal Google Docs based extension for jelly?!

Redwolf Programs

I used to developed Google Docs add-ons, but you don't need an add-on for that. If you go to "preferences" you can make custom automatic find/replace things.

Razetime

@RedwolfPrograms can you share those?

like some sort of document template probably

Redwolf Programs

No, it's per-user rather than per-document

Razetime

ah then that doesn't help much does it

Redwolf Programs

You can just make the strings things you wouldn't usually type by accident, like j_xd for ẋ

Razetime

5:19 AM

yeah i guess it's more personally customizable

I'd keep it similar to lb

ais523

5:46 AM

Ubuntu's default X11 configuration has overdot on altgr-right shift-/

as long as you start holding altgr before you start holding right shift

(underdot is just altgr-/)

actually almost all Jelly's characters have some way of typing them like that, which is probably intentional

Redwolf Programs

For me (Chrome OS with US EXTD layout) it's right alt and dot

ais523

(the main exceptions are ⁴ and friends, for which I need to use the compose key)

Redwolf Programs

Wasn't Jelly's codepage designed specifically for the US INTL layout?

ais523

I think so, yes

mine's a standard UK layout, but there are a surprising number of altgr-combinations defined that most people don't know about

(on Linux/X11, that is; Windows' AltGr is impoverished by comparison)

Unrelated String

@ais523 I think I did read somewhere that Dennis designed the codepage to be easy for him to type

ais523

5:49 AM

it took me ages to work out that ɠ and friends use an altgr-j prefix

Redwolf Programs

The only thing annoying for me to type is the hooked letters like ɠ

I still haven't found a way to on Chrome OS's layout

ais523

although I'm not sure it always picks the same hooked version that Jelly does

it took me so long to figure out why my Đ wasn't working to create two-character quicks

(the actual character is Ð)

in case you still can't tell them apart (and I can't, by looking at it!), the first is a capital đ, the second is a capital ð

ais523

6:06 AM

hmm, so something I've been struggling with: when writing a golfing language where builtins don't line up neatly with single bytes, what sort of codepage should be used (if any)?

I think it probably makes the most sense to just define a program as a sequence of bytes, and assign them arbitrarily, but it's nice to have something more than a hexdump to show

I considered making them all emoji (on an SBCS, of course) to have the greatest possible chance of complying with restricted-source restrictions, but it rather feels like something like that shouldn't matter

Unrelated String

one approach is using an existing and established code page like cp437, possibly with minor alterations for some matter of convenience

I've also seen something use Braille, since it ends up being able to visually represent the binary representation of each byte

Redwolf Programs

Yeah braille's cool

Sledgehammer uses it IIRC

Unrelated String

yeah

ais523

CP437 doesn't work great for C0 control codes

Redwolf Programs

You'd probably want to rearrange the Unicode range the braille charactera are in, since it's in some weird order by default

ais523

6:13 AM

there are assignments defined but I'm not sure they're standard, and they don't work on all CP437-based systems

also, aren't Braille characters technically letters? so not ideal for bypassing source restrictions

Redwolf Programs

They represent syllables, right?

I doubt they'd be considered letters for most restricted source challenges

ais523

most Braille characters are single letters, a few represent pairs of letters, or words

real-life Braille tends to be fairly abbreviated in weird ways, though, it's like words are intentionally misspelled to save on raised dots

(I must be one of the few people who learned to read Braille by sight; I'm a bit rusty at that nowadays, though)

Redwolf Programs

Oh cool!

ais523

either that or it's just that nobody proofreads it

(prooftouches it?)

Bubbler

Is it really that default I/O does not mention "an array terminated by a sentinel value" as a valid I/O format at all?

ais523

6:20 AM

there are so many things that could be controversial, and often nobody thinks of posting them

I added a very late answer to the post to cover a borderline situation, people seemed to think it was OK

also one that was uncontroversial but had been left off by mistake

fwiw, I'm actually not a fan of permissive I/O (in particular, it means that golfing languages get to cheat a few bits on every program by changing the way they take I/O), but I'm comprehensively outvoted on this

(my favoured solution would be to fix the I/O conventions per-language, so that each language could take I/O in the way that was most meaningful for it, but not cheat bits by varying it per-problem; come to think of it, defining different I/O conventions to be different languages might help to solve the problem)

user41805

oh hi @MitchSchwartz, cool sed 99-byter to fizzbuzz btw. i'm some 30 bytes behind (just checked fizzbuzz on ppcg now and saw your 94b one, wow love that 00..99 generation, to make things interesting i started a bounty which will go to the shortest sed solution in a few days from now)

@seshoumara (pinging your last chat message, since you may be interested in the following) did you see codegolf.stackexchange.com/a/220633

Mitch Schwartz

hey, yeah i actually submitted because i searched the chat and saw some discussion about it, and it's a pretty old solution anyway :) i looked at it again and found a small variant, taking out ` *` and using 00Buzz and changing ` .*` to ` ..` but haven't found anything promising (idk if backticks work in chat but i'll find out after pressing enter)

ais523

backticks are implemented on chat but the parsing rules are different from on main

with a consequence that nobody can remember what they are

user41805

honestly, such a big difference between anagol's sed skills and ppcg's

Redwolf Programs

SE really just needs to switch the chat markdown renderer to the same one comments use, it works better (and supports markdown and looks better)

user41805

6:36 AM

@MitchSchwartz i recall submitting submitting 160-170 byters to fizzbuzz on anagol, and then a few days hence i saw you submitted yours, i guess you noticed my spamming of golf.shinh.org/recent.rb :)

ais523

has anagol added many golfing languages yet? I recall that it was missing many golfing languages popular on CGCC last time I checked

Razetime

anagol only has GS and GS2 afaik

Mitch Schwartz

yeah the recent submissions page is pretty low tech but it gets the job done :) tails also made a twitter bot named anagolfer that filters for rank 1 or > 10k points submissions

it's been a long time since new languages were added afaik

ais523

wait, how do you get >10k on anagol?

are there bonuses for something other than byte count?

user41805

it's if you beat the top solution after a post-mortem

Mitch Schwartz

6:39 AM

post mortem solutions are scored by comparing with the top score before deadline, if there was one

Bubbler

@Bubbler I/O method proposed. Please check out and vote.

Unrelated String

7:12 AM

@ais523 That would be pleasantly consistent with how flags are handled

ais523

right, that's what made me think about that (assuming we mean command-line flags, not post flags)

Unrelated String

yeah

I feel like broadly speaking permissive I/O isn't a big issue but it always does feel a bit like cheating when I do something funky with Brachylog

generators and reversed variables

ais523

reversed variables is weird, but the only difference from languages like Jelly is that when you do it in Brachylog, it feels like cheating (often a Jelly program will change length if you swap ³ and ⁴ but nobody thinks of it as cheating there)

generator vs. first output is interesting because of how Prolog's meant to be used interactively

if you're running Brachylog in the actual command-line interpreter (rather than on TIO), if you run a predicate, it'll give you the first output and ask if you want more

and this has been standard Prolog for ages

so the standard is to put both the first-output and generator interpretation on everything (unless it's deterministic and thus can't generate more than the first output)

there is probably some rule of programming languages that, for any rule you come up with to try to generalise behaviour across multiple languages, someone can create a language that makes the rule look absurd

2

Mitch Schwartz

the only thing i know about braille is that it has all the pentominoes :) ⠹⠶⠳⠗⠏⠞⡇

ais523

those are tetrominoes

Mitch Schwartz

7:21 AM

right, sorry

ais523

also, I think Braille characters using the fourth line are nonstandard/extensions

the original was just a 3×2 grid

although there are probably lots of incompatible versions nowadays

Unrelated String

I suppose that's why the fourth row renders in descender-space here

Mitch Schwartz

you probably shouldn't trust me with any holy hand grenades

@user41805 i DM'd tails about sed fizzbuzz

New Main Posts

7:41 AM

2

Q: Hello, Permutations!

For this challenge, you will be required to write 3 programs: The first program should be a quine, i.e, it should print it's own source code. The second program should read the input and print all of it. The third program should simply print the string "Hello, Permutations!", with or without quo...

user41805

i guess clock would be interested too, i once 'met' them on a hn post on a maze-solving sed program by xsot, where clock brought up a tip to golf.shinh.org/p.rb?maze+solving, so i actually outgolfed clock's submission tying tails, then clock outgolfed us both. turns out after this encounter, clock started golfing again at anagol after not having 'golfed in years'

though i don't know how you anagolfers keep in touch, #anagol's pretty dead last time i checked

Mitch Schwartz

twitter is the best way i know of, but i'm not the most outgoing person these days

wow that's cool running into clock, i didn't know what hn was but figured it out, looking at the thread now

i was in the irc channel for a few years consecutively i think, i probably switched computers and just never set it back up, but yeah it was pretty quiet, especially after the bot stopped announcing

Mitch Schwartz

9:10 AM

(i just realized that my last message could be easily misinterpreted to mean that before the bot stopped announcing it was mostly just bot traffic and no human interaction, but what i actually meant was that i think people talked more when the bot was working)

Adám

9:51 AM

CMC: Given some JSON in "fully expanded form" (each node on a separate line), uppercase all object member names. E.g.

{
  "be": {
    "dragons": 1,
    "b\\e\"ar\ns": "two"
  },
  "here": "are"
}

becomes

{
  "BE": {
    "DRAGONS": 1,
    "B\\E\"AR\nS": "two"
  },
  "HERE": "are"
}

user41805

10:08 AM

@Adám does "\u0065" becomes "\u0045"? what about "é" → "É"?

Adám

@user41805 Yes and yes.

I should either add a more comprehensive test case or maybe post on Main?

@user41805 \u0065 can also become e.

I should also add that your output can look whichever way you want, as long as it encodes the equivalent value (save for uppercased names, of course).

user41805

@Adám i'm not sure if 'fully expanded form' is standard terminology, but if it isn't this should be explained (say, can there be whitespace between the key and the colon?)

or just have input be normal json

Adám

@user41805 Good points. (And good that I made this a CMC.) Making it just "valid JSON" would remove all doubts.

pxeger

how about saying that the JSON will be something that could have been outputted by JS's JSON.stringify(data,null,4)

and is the top-level value guaranteed to be an object (not an array or string etc)?

Adám

10:25 AM

@pxeger That would require intricate knowledge of JS spec/impl. Just valid JSON according to json.org is simpler, no?

@pxeger No.

user41805

shell, using jq and sed, 38b jq .|sed 's/.*[^\]":/\U&/;s/\\./\L&/g' which i think works (i was about to post a 48 byter, then i remembered mitchs might be watching, so now i shaved 10 more bytes)

@Adám oh wait, this invalidates my answer, i forgot about arrays

Adám

@user41805 Maybe you shouldn't keep working on it now, if I am to Main it.

user41805

okay, are you going to sandbox first or go directly to main?

Adám

Probably sandbox.

Razetime

11:24 AM

@Adám can member names be of any datatype?

Mitch Schwartz

12:03 PM

late reaction because i hadn't signed into this site since before the site redesign, but - i'm digging the green color for when you hover the mouse over your own "message actions" arrows

it took me a while to figure out it was related to the redesign, i thought maybe it was a global thing

or, i guess it's just an assumption, but there's a lot of green everywhere, so

pxeger

@Razetime do you mean the keys in the object? if so, the answer is no - in JSON, all keys must be strings

Razetime

1:23 PM

@pxeger oh that make it simpler

New Sandboxed Posts

3:10 PM

0

A: Sandbox for Proposed Challenges

Uppercase JSON member names Given a single valid JSON value, uppercase all member names. That is, you must also return valid JSON that encodes an equivalent value, except that all names of members in all objects, have been converted to uppercase according to Unicode standards. Note that the give...

John Dvorak

7:02 PM

Anyways, re the semi-recent topic of intentionally bad languages, there is something you could use: floating point style error handling

pxeger

you mean NaN != NaN

John Dvorak

Every operation that should return a pointer but fails, returns a special not-a-pointer value, or NaP for short. Any operation on a NaP returns a NaP and otherwise does nothing. And in case someone decides to check ... indeed, NaP !== NaP

or another idea, !(NaP == NaP) && !(NaP != NaP)

there's nothing better than subtle bugs introduced by careless application of DeMorgan's law :D

Adám

7:34 PM

@JohnDvorak No need for complicated logic. See "0" vs 0 vs "" vs "0"

Redwolf Programs

@JohnDvorak If that fails, it should obviously return a NaNaP

Neil

7:49 PM

Just discovered this: youtube.com/watch?v=BnnmA2klBN8

pxeger

a work of art

New Main Posts

10:38 PM

0

Q: Matching ABACABA-type patterns

(This challenge is related to the challenge "Generate the Abacaba sequence." Zimin words (also called "sesquipowers") are an important idea in the subject of "combinatorics on words". This is because every Zimin word is "unavoidable", meaning that every sufficiently long word over a finite alphab...

code-golf math combinatorics pattern-matching

ChartZ Belatedly

10:58 PM

Any feedback? (changed significantly since I posted to main and deleted it)

rak1507

looks good

I guess the only issue I have with it would be, is there going to be any way of doing it that isn't find fortunate numbers, outer product, is n in that

ChartZ Belatedly

I doubt it, Fortunate numbers are only really interesting in one bit, and that's "does a composite Fortunate number exist?", so it's unlikely it'll get amazing answers

But I want to salvage the deleted main post, and I couldn't think of a better thing to do with Fortunate numbers

rak1507

yeah, it seems like it's basically 'do a few different things chained together' which is not all that interesting

I think the main post could be salvaged by having only strictly increasing sequences

(you could also post that separately as another challenge though, this one is sufficiently different)

ChartZ Belatedly

No, then it just becomes "get the first n elements, product, get the next element, subtract"

which isn't bad, it's just not that interesting

rak1507

true

another issue with that challenge is the numbers will get so big that a few languages might not be able to handle them, but that's not really that bad of an issue

ChartZ Belatedly

11:07 PM

@rak1507 Actually, I think I'll do that and try and properly salvage the original one

rak1507

sounds good

ChartZ Belatedly

Given that the function is a black-box, and that I'm not going to allow people to take a function that takes an integer x and outputs the next integer greater than x in the sequence, finding the "next" value from the product isn't as trivial as just "get the next integer", so the original challenge should be more or less ok if I limit it to strictly increasing sequences

New Sandboxed Posts

1

A: Sandbox for Proposed Challenges

I ain't no Fortunate sum The primorial \$p_n\#\$ is the product of the first \$n\$ primes. The sequence begins \$2, 6, 30, 210, 2310\$. A Fortunate number, \$F_n\$, is the smallest integer \$m > 1\$ such that \$p_n\# + m\$ is prime. For example \$F_7 = 19\$ as: $$p_7\# = 2\times3\times5\times7\ti...

ChartZ Belatedly

3

Q: Generalised Fortunate Prime Sequences

The primorial \$p_n\#\$ is the product of the first \$n\$ primes. The sequence begins \$2, 6, 30, 210, 2310\$. A Fortunate number, \$F_n\$, is the smallest integer \$m > 1\$ such that \$p_n\# + m\$ is prime. For example \$F_7 = 19\$ as: $$p_7\# = 2\times3\times5\times7\times11\times13\times17 = 5...

code-golf math sequence

NMP won't pick it up as it's been edited and deleted, not reposted

@ChartZBelatedly Correction, I just got Curious, despite my tracker saying I still need to post 2 more questions O.o

Newly Featured Questions

11:43 PM

30

Q: Sum of primes between given range

Write the shortest code for finding the sum of primes between \$a\$ and \$b\$ (inclusive). Input \$a\$ and \$b\$ can be taken from command line or stdin (space seperated) Assume \$1 \le a \le b \le 10^8\$ Output Just print the sum with a newline character. Bonus Points If the program accepts m...

code-golf primes

Transcript for