The Nineteenth Byte

12:01 AM

Oh I see

12:30 AM

So today I was parsing a new file in the VSL standard library (that connects to DOM apis) and apparently nodejs needs more than 1.7GB to parse this 2K+ line file so it is now crashing from out of memory ;_;

probably memory leak but wow

@ASCII-only do you have any potential idea on how to address

Unless nearley creates megabytes worths of objects I don't think a memory leak there could be the issue

and additionally they are no reference cycles until the transformation passes so given the simple tree-like data structure it should be pretty easy on the GC

New Sandboxed Posts

2:47 AM

0

A: Sandbox for Proposed Challenges

Loop with no loop code-challenge restricted-source Challenge: Create a loop without your language's loops features. The point is to print to stdout this text: Hello World #0 Hello World #1 Hello World #2 Hello World #3 ... Hello World #97 Hello World #98 Hello World #99 ...but without lo...

ASCII-only

3:10 AM

@Downgoat it does create gigabytes worth of objects if there are ambiguities, yes

Downgoat

@ASCII-only ooooooo ic

why would this be ambiguous:

b = a()
(b)

ASCII-only

wait that is?

Downgoat

yeah

ASCII-only

hmmm

maybe something to do with tuples possibly

Downgoat

tuples are disabled

or at least I changed them to { ... }

ASCII-only

3:13 AM

oh. well

Downgoat

gist.github.com/vihanb/4f225d38ad31236d019e3bc0ed76dee6

ASCII-only

do they parse differntly? or the same

Downgoat

looks like they parse differently no idea how

oh

it does two function calls... hm....

maybe disallow newline between identifier and (

ASCII-only

not a bad idea, would make things a lot more readable

well, would make it harder to accidentally make unreadable code

Downgoat

hmm lets see if that works

@ASCII-only how would that be implemented given this: github.com/vsl-lang/VSL/blob/inheritance/src/vsl/parser/…

ASCII-only

3:20 AM

@Downgoat what's _ again

Downgoat

newline(s)

ASCII-only

oh

so here

is the line you need to change

i guess just return fail if you find a function call node on that line?

ASCII-only

3:39 AM

@Downgoat ^

@Downgoat &

Downgoat

4:02 AM

@ASCII-only :|

oh ok so what I did is the only prop you can prefix with newline is foo \n .bar

this would normally mean .foo would be ambiguous but planning to only allow these in InlineExpressions

ASCII-only

:D

@Downgoat ok that works too

@dzaima @Adám literally just add a userscript with the sole line of code location.href='github.com/notifications'

@dzaima oi lazy. btw which divider

feersum

4:50 AM

Does anyone have a VPS recommendation that would be easy to set up for a short period

I have like $2 of bitcoins

And only need it for a few days.

Downgoat

@feersum DO is really easy to get going in 30 seconds and should only cost you like $0.35 for two days

ASCII-only

5:52 AM

@Downgoat they have 2 day plans???

Downgoat

they charge by hour

but bill at the end of the month

Also does anyone know if there is a WebIDL index for WHATWG HTML living standard. The source doesn't have a full index

ASCII-only

8:48 AM

hmm. if the goal of the builtins is to be as minimal as possible, what kind of conditionals/branching should it have?

just branch if? (i.e. goto_if?)

@DJMcMayhem which one is which though? (pls answer seriously) (wait, don't actually answer seriously - i'm just saying i'm actualyl confused which is which)

feersum

9:08 AM

This retarded DO reseller has a public key upload form but does not specify what format it should be in.

Any guesses?

I tried the one like ---- BEGIN SSH2 PUBLIC KEY ----

Then I tried without the dashed lines and with base64 padding added.

Because there was a screenshot of a DO form using that format.

ASCII-only

@feersum gpg?

feersum

WTF, I suddenly got a chat ping sound

The chat ping sound was broken for years.

ASCII-only

lol

feersum

@ASCII-only What does that look like?

ASCII-only

similar to ssh2, i'm guessing possibly maybe they only accept gpg???

feersum

9:13 AM

Well they linked to DO documentation that wants you to use a tool called ssh-keygen

feersum

9:25 AM

Hmm, the DO instructions for Windows (instead of the "Mac/Linux" ones that I looked at) describe yet another format, maybe that's the one they want

That's it, turns out instead of dashed lines it requires the string ssh-rsa to be placed at the beginning.

ASCII-only

9:47 AM

CMP: In a strictly typed language where an undefined/void value is returned from non-returning functions, what should the type be called?

Leaky Nun

@ASCII-only don't be so strict man

feersum

egg

ASCII-only

also, how bad of an idea is it to make 1. boolean literals, 2. strings and 3. integers behave like identifiers (i.e. you can assign to them) (returning their value by default of course)

@feersum egg?

feersum

I'm suggesting the type be named "egg".

ASCII-only

yeah, but why egg though :|

feersum

9:53 AM

An egg is shaped like, and is a common idiom for, zero, which is the amount of information your type contains.

Neil

@feersum yeah, ssh has a different public key format to openssl

ASCII-only

10:08 AM

hmm. what should i do if generics are required for core features (lists)

because e.g. .NET has the System.Collections.Generic namespace (and possibly others), but for them generics aren't core features

so... i guess the real question is: should generics be handled specially? or... if it's a core feature, it might be even better to treat all normal types as essentially zero-parameter generics??? (i.e. all types are treated as (well, kinda) generic types - this would ~~reduce~~ eliminate the need to specialcase generics from normal types)

ngn

10:23 AM

@flawr a French Fry

flawr

@ngn why are they wearing onions?

ngn

10:46 AM

@flawr just it case they suddenly feel the urge to cook :)

Adám

11:55 AM

I have this challenge idea: Given a Boolean matrix compute the next generation which has each row be the XOR of the row above it and the row below it. How does that sound? Does that have a name?

This is for main, but here is an example:

[[0,1,1,1,0],[1,0,0,1,1],[0,0,1,0,1],[0,1,0,0,1],[1,1,1,0,0]]
gives
[[1,0,0,1,1],[0,1,0,1,1],[1,1,0,1,0],[1,1,0,0,1],[0,1,0,0,1]]

Adám

12:07 PM

This is for main

ngn

oh... ok. deleted

Adám

Maybe it isn't a good challenge as it is could be seen as merely a cumbersome input format.

ngn

it took me more time to format the tio link than to solve it :)

Adám

Yes yes, it is trivial in array languages

XOR isn't important here, the idea was some kind of cellular automaton, but not just in a 3×3 Moore or von Neumann neighbourhood.

Erik the Outgolfer

12:28 PM

@Adám wouldn't a 1×3 neighborhood on a torus be enough?

Adám

@EriktheOutgolfer Yes, it would. :-(

I'd like something where every row is dependent on the entirety of the neighbour rows, instead of every cell being dependent on its neighbour cells.

Erik the Outgolfer

@Adám then XOR isn't enough

Adám

@EriktheOutgolfer No. Hence my posting the idea here rather than in sandbox. It needs brainstorming.

Hm, maybe cyclically rotating the middle row as many steps left as indicated by the row above, and as many steps right as indicated by the row below?

With neighbours being interpreted as integers.

Erik the Outgolfer

@Adám like, neighboring rows interpreted as binary numbers?

Adám

@EriktheOutgolfer Yes. ^^

New Main Posts

12:35 PM

0

Q: Implement an advanced expression evaluator

You don't need to code golf this puzzle. The goal of this is to input a string (containing +, -, * , /, %, ^, ==, !=, >=, <=, >, <, !, &, |, parenthases, and integers). You must implement a function that returns the integer result without using a function that evaluates an expression(like eval()...

math

Adám

Ah, even more fun. A 3D Boolean array. Process layers. Rows of layer above indictate per-row how to rotate. Columns of layer below indicate per-column how to rotate.

Erik the Outgolfer

@Adám hm, can't do that in 9 bytes in Jelly for sure...

Adám

@EriktheOutgolfer Why 9?

Erik the Outgolfer

@Adám because that's what I can do for the rows :P (no, won't reveal)

Adám

@EriktheOutgolfer Just transpose, and repeat, no?

Erik the Outgolfer

12:47 PM

@Adám and also account for the difference between the layer above and the layer below

Adám

Sounds fun.

DJMcMayhem

1:14 PM

@ASCII-only Brain-flak: incredibly easy. V: incredibly hard

Erik the Outgolfer

wait... wasn't ^ posted before...?

Adám

@EriktheOutgolfer NMP? Yeah, twice.

Erik the Outgolfer

@Adám no, DJ's reply... I can swear that question regarding syntax highlighting has been asked before

dzaima

oh huh i have no idea what exactly this is but it looks somewhat cool

too bad it's forever expanding

Adám

@dzaima Why? You could just scale it. Looks fractal-y

dzaima

1:25 PM

@Adám means nothing like gliders/stationary patterns

Adám

@dzaima Right. It is more akin to Sierpiński.

dzaima

also the fact that every next iteration takes more time to execute

@Adám sierpiński doesn't seem very random though

Adám

@dzaima Neither is this.

dzaima

@Adám i don't really see a pattern even in the middle rows ## ####### ### #### #### ##### ####### ##, even less so in the whole thing

Adám

@dzaima Contratulations, you've invented (found? discovered?) the dzaima carpet!

dzaima

1:29 PM

it can result in more chaotic & non-symmetric things with a different starting setup

hmm this looks interesting

Adám

@dzaima I like the harpoon.

@dzaima You're having way too much fun with ⌺. There should be a tax on that ;-)

Erik the Outgolfer

> Character 9018: $10,00 per item.

Adám

@dzaima Here you go.

dzaima

@dzaima oh fixed the asymmetry of that and it's back to forever-expanding randomness :|

feersum

1:46 PM

Why do I even have a Discover card?

Only like 0.002% of places accept it.

I have spent like 3 or 4 hours trying to set up a VPS, no progress so far...

dzaima

yay symmetric and non-forever-expanding :D

oooooh this is a moving thing :D

Adám

@dzaima And cyclic. (Kind-of has to be…)

dzaima

@Adám why has to be?

Adám

@dzaima If it is non-expanding, there are a finite number of states, which means it eventually must go into a loop. Could be with period 1, or more.

dzaima

@Adám it can expand (I'm pretty sure), i was just happy that it didn't expand constantly infinitely

@Adám also thanks for giving me the idea to try to find non-normal cellular-automaton-ish thing, now i have something do today that doubles as a backup plan for a homework :D

hmmm for a fancier editor/display do i try to implement the thing in straight up processing java or do i try to do it in my APLP5 and deal with the slow speed

Adám

2:02 PM

@dzaima Do you want to animate (i.e. show it in the same position, but evolve over time?)

dzaima

@Adám yes, and have the ability to pause, single-step, and modify it live in the display

El'endia Starman

For people who don't read APL (me) what exactly is the rule/algorithm for going from one step to the next?

dzaima

kind of like this

Adám

@dzaima I made such a thing for GoL.

@El'endiaStarman Wait, you don't read APL‽ We can fix that.

dzaima

@El'endiaStarman take the 5x5 neighborhood, sum the on cell relative positions to the current cell, and take the value of the cell at that relative position, cycling as a torus in the 3x3 neighborhood

El'endia Starman

2:07 PM

@Adám Pretty sure that'd take quite some time. :P

dzaima

oh right, before the cycling step you negate the summed coordinate. pretty arbitrary, but otherwise you get infinite expansion

El'endia Starman

Very interesting.

Adám

@El'endiaStarman Nah, I'd say 24 hours.

El'endia Starman

@Adám Still longer than the ~60 seconds it took dzaima to explain the algorithm to me. ;)

Adám

@dzaima Maybe you could write a blog post about your adventures in ⌺ land?

Erik the Outgolfer

2:10 PM

@El'endiaStarman give someone a fish, they'll be fed once; teach someone to fish, they'll have food forever ;)

dzaima

@Adám that'd require making a blog, and writing stuff :p

Adám

@dzaima You could maybe write a guest entry on Dyalog's. The second part, well, yeah…

El'endia Starman

So, it occurs to me that this CA is like an elementary cellular automaton (e.g. Rule 110), but instead of a 1x3 neighborhood, it's a 5x5 neighborhood.

All 2^25 possible states of the 5x5 neighborhood have a single outcome. It's just conveniently described by a shorter algorithm.

dzaima

@El'endiaStarman hmm, wouldn't GoL technically qualify for that too?

El'endia Starman

In a sense, yeah.

People just decided the birth/survive descriptions (and similar) were more convenient.

I'm having difficulty finding any literature on 2D elementary cellular automata. Maybe because there are too many, perhaps?

Adám

2:19 PM

Shortest GoL in Dyalog APL (that I've found so far) is U∊⍨⊢∘⊂⌺3 3 where U is a constant.

dzaima

@El'endiaStarman exactly. though i wonder if it's possible to take this idea further and make that picking part not only in the 3x3 neighborhood but the whole board, without making almost nothing possible

Adám

@El'endiaStarman ^^ just looks up 3×3 neighbourhoods in a look-up table.

El'endia Starman

@Adám Huh, interesting.

I'm guessing U is what encodes the lookup table?

Adám

@El'endiaStarman U is the table. The lookup is ∊ (⍨ just commutes its arguments).

So All it does is for each cell ask whether the 3×3 neighbourhood is a member of the table of 3×3 neighbourhoods that lead to life in the next gen.

El'endia Starman

Neat.

feersum

2:35 PM

All right new plan: create a paypal account and try to get a company that owes me money to send it to the paypal.

Then use that to buy VPS service.

dzaima

2:45 PM

@El'endiaStarman hmm really any cellular automaton which evaluates everything at once can be called an elementary cellular automaton with an ∞×∞ neighborhood :p

El'endia Starman

@dzaima An exhaustive accounting of possible rules would be...... exhausting, methinks.

I had better words but couldn't resist the "pun".

@dzaima Also works with a 2x2 square without one corner.

(Which is the smallest asymmetric configuration you can get.)

El'endia Starman

3:11 PM

A bunch of coding memes: imgur.com/gallery/TA76KUT

4

DJMcMayhem

3:21 PM

@EriktheOutgolfer Yeah, it was but ASCII was asking about it

Erik the Outgolfer

yeah, maybe somebody else asked it before, memory can be fuzzy but weird déjà vu is weird

New Main Posts

3:44 PM

0

Q: Connect 4: Spot the Fake!

The bank has been broken into, and all the local mafia thugs have an unusual alibi: they were at home playing Connect 4! In order to assist with the investigation, you are asked to write a program to validate all the Connect 4 boards that have been seized in order to check that the positions are...

code-golf board-game

trichoplax

4:03 PM

@El'endiaStarman Thanks! I found the comments more amusing than the post...

Particularly:

> Wait, so most people DON'T call it "c plus plus plus plus"? But it's four plusses!

El'endia Starman

Took a while to get that one but that's good. :P

trichoplax

Yeah, C jokes are faster to get

Erik the Outgolfer

@NewMainPosts faulty premise, why can't Y start?

El'endia Starman

> For example (with `R` starting)
> you must specify which player goes first

Erik the Outgolfer

4:36 PM

@El'endiaStarman um, yeah, but my point is that the backstory doesn't really stand like that, detectives can't know who started or if they switched the starting color between different games

yes, I understand the challenge

(one more reason to avoid backstories, I guess)

Khuldraeseth na'Barya

That Android Studio one is too relatable. I've just spent 36 hours at a hackathon (and accomplished very little :(). Struggles with Andriod Studio and with Qt Creator, as well as with trying to write C++ like Haskell.

MilkyWay90

5:00 PM

Who would have a suggestion for golfing languages?

What builtins should it use

Khuldraeseth na'Barya

All of them. Perhaps even an "execute string as Jelly code" one.

Anonymous

@MilkyWay90 If you're scared to click it, why post it here?

Anonymous

6 messages deleted

MilkyWay90

5:15 PM

@Mego Sorry

ngn

5:51 PM

@MilkyWay90 i clicked

it looks like one of those cicada challenges

hm... i thought it's an old challenge but it seems really fresh. i didn't go too deep down the rabbithole, yet only 7 people have been there

@Mego i think you should restore that link

MilkyWay90

@ngn Thanks, it was posted on my website (the clues). I'm currently decoding a ceaser cipher

oh, maybe we should make a PPCG challenge for a ceaser cipher

I'm worried it may be closed as too broad

Maybe the user would give the substitution, and then the program would take another input as the input to the substitution input, and output the ceaser cipher output

I think that would work

Unless if it's a duplicate

ngn

@MilkyWay90 tell me how many have watched the rick roll when you get there :)

Cicada 3301

Cicada 3301 is a nickname given to an organization that on three occasions has posted a set of puzzles to recruit codebreakers/linguists from the public. The first internet puzzle started on January 4, 2012, and ran for approximately one month. A second round began one year later on January 4, 2013, and a third round following the confirmation of a fresh clue posted on Twitter on January 4, 2014. The stated intent was to recruit "intelligent individuals" by presenting a series of puzzles which were to be solved. No new puzzles were published on January 4, 2015. However, a new clue was posted on...

MilkyWay90

6:08 PM

@ngn There's a rick-roll?

ngn

^^ if someone has no clue what i was talking about

MilkyWay90

And I do know about cicada

@ngn Do you know where the ceaser cipher leads to?

ngn

@MilkyWay90 it's right after the caesar septem, you'll get there (i'm giving you a hint)

MilkyWay90

Okay

@ngn I know what a ceaser cipher is, not a ceaser septem

dzaima

6:26 PM

so i put some vectors on a pattern in that cellular automaton of mine

2

El'endia Starman

Cool!

Exploded so individual frames can be studied.

Also, not sure how I didn't notice before, but the first half and the second half are the same, just vertically flipped.

MilkyWay90

@ngn The rickroll has 7 views

ngn

6:51 PM

@MilkyWay90 hm, the same as when i watched it...

MilkyWay90

@ngn Oh, maybe a bug or something

Do you know how to decrypt the rest?

ngn

@MilkyWay90 i haven't tried yet. i'm having dinner now

MilkyWay90

@ngn Okay. The last few were easy, now I don't know how to decyrpt this

ngn

i'm not willing to spend too much effort on this, as at some point there's likely to be a US location-based task, but i was thinking, maybe a combined effort among ppcg, the side channel, and puzzling.SE can go a long way :)

dzaima

7:08 PM

@dzaima here's the thing that made that

MilkyWay90

@ngn Yesss, great idea

New Main Posts

0

Q: I am getting an error in string operation. Please help

s=s+arr[i]; Here s is string and arr is int. s+=arr[i]; why the first statement is invalid but the second statement is valid?

string c++

ngn

@MilkyWay90 the real challenge is, what could be a more motivating reward for people to publicly share their progress than a potential job at whatever super-awesome company or intelligence agency is behind cicada? is stackoverflow reputation worth more? :)

btw, we haven't verified yet it's truly cicada. at some point there should be something signed with the key they published a couple of years ago

MilkyWay90

@ngn I really doubt it's Cicada, although I could post this on Puzzling.SE

And reputation

would be the payoff, and it could also be a collaborative effort, like the Quest for Tetris

But if it is Cicada 3301, people could also be blocked from entry because it was a group effort, right? This is also why people were blocked from Cicada in 2014

ngn

yeah, the tetris effort looks really impressive

MilkyWay90

7:22 PM

@ngn So which SE site would be appropriate for this?

ngn

@MilkyWay90 i should read more about this history of this thing...

MilkyWay90

@ngn I've studied Cicada 3301 a lot

ngn

@MilkyWay90 i don't know. i would bet that puzzling and crypto would have the most volunteers

MilkyWay90

Not a lot, but I've looked at it a few years ago

ngn

@MilkyWay90 and yet, you thought this was a virus??

MilkyWay90

7:26 PM

@ngn Yeah. Since people from Geometry Dash came (and the trolls from there too), I was wary whether it was a virus or not

I decrypted the original clue at my school, so I guess my friends would also be interested.

docs.google.com/document/d/…

@ngn for what I found so far

Khuldraeseth na'Barya

Another team at this hackathon made something they call Tithub.

MilkyWay90

@Khuldraesethna'Barya What's that?

@ngn do you know any puzzling.se chatrooms that I could ask for help for on this puzzle?

Khuldraeseth na'Barya

7:48 PM

I have no clue what it is, and I wasn't about to stroll over and inquire.

All I saw was their logo.

ngn

@MilkyWay90 i can see you're already there :)

MilkyWay90

@ngn Yes, I am

I already asked puzzling

I think I will ask cypto now

ngn

i'm still not sure whether people would have any incentive to publish their findings, but let's see

MilkyWay90

@ngn Most people wouldn't need incentive. I did it because it was just interesting

@ngn If you find any clues, put it on the google doc (docs.google.com/document/d/…)

MilkyWay90

8:17 PM

youtube.com/watch?v=MtO589eu-Rc may seem like a rick-roll, but it isn't!

look at the title and description

ITS A CLUE

ngn

@MilkyWay90 oh, you're ahead? nice :) i'm afraid i must put this off till tomorrow

MilkyWay90

@ngn Okay.

MilkyWay90

8:49 PM

@ngn Yes! I did it! I found a cipher!

It's on the doc, if you guys want it for the puzzle

lirtosiast

in 05AB1E - Oasis, May 9 '18 at 14:41, by Magic Octopus Urn

I'd say the biggest thing that 05AB1E does well is everything, but the stuff it doesn't do well; it really... really doesn't do them well.

Is there a name for that property of a language, of being able to do things unsuited to the language without wasting too many bytes?

generality? worst-case golfiness? golfing robustness?

Anonymous

9:05 PM

@lirtosiast bf-ity

MilkyWay90

@Mego lol.

quartata

@lirtosiast "adversarial robustness" :p

MilkyWay90

@lirtosiast A practical programming language

dzaima

@MilkyWay90 APL and Mathematica disagree. (for completely opposite reasons though)

actually no what brain, mathematica completely agrees :|

Anonymous

9:21 PM

@dzaima "caprine animal"

lirtosiast

@MilkyWay90 Maybe. I wonder how Jelly would do in front-end dev though

I suppose it could be measured by, say, ratio of longest 5 outliers to mean code size on a collection of tasks meant to represent the universe of problems. Maybe drawn from PPCG problems, real-life code examples in a variety of areas, and reimplementing its own builtins.

The other question is, what causes a language to have or not have that?

Anonymous

@lirtosiast I think that some basic stuff like number theory (basic arithmetic, primality, fibonacci, factoring), string manipulation (reversing, contains, substring, replacement, Levenshtein distance), and fundamental data structures (array, linked list, dictionary, stack, queue) would be a pretty good corpus

Anonymous

Also GoL for good measure

lirtosiast

That seems better. With builtins banned of course.

If the corpus is small enough we can actually look at the histogram of normalized code size

colorized by the category of task

lirtosiast

9:45 PM

Come to think of it we could probably do that anyway

Pavel

11:27 PM

I'm working on a Python project with like 5 other people and magically everyone had Python 3.7.2 so I get to use Dataclasses and I'm pretty happy about that

MilkyWay90

11:38 PM

@Pavel Yay

El'endia Starman

@Pavel Huzzah! My work is currently on 3.6 though our code works with 3.7, so we'll get there eventually. :P

MilkyWay90

I work in 3.6.4

But everybody around me does 2.7, for some reason

Anonymous

@MilkyWay90 You should find better people to be around :P

El'endia Starman

If my company can transition to Python 3 last year, almost ten years after it was released, then people around you can too! :P