The Nineteenth Byte

New Posts

00:05

0

Q: Bash on *nix: bi-directional sorted mapping with adding single key

Write code in most recent Bash 4, may use POSIX utilities or coreutils. You can reserve up to 20 global variable names. Neither kind A nor kind B string values contain newlines or null characters. Given is a global indexed array of valid kind A string values, that may have elements disappear fro...

Seggan

00:48

@UnrelatedString what engine tho

Unrelated String

probably one that doesn't take too long to run :P

Seggan

also the evaluation could vary with time

@UnrelatedString i think stockfish is pretty much default

OldSandboxPosts

01:00

Sandbox posts last active a week ago: My prefix is food, my suffix is rude

lyxal

oml I found APL in a theory of computing textbook

> APL is not one of the major programming languages in use today

oof

Seggan

lyxal

you're either showing something cursed from one of your projects (or CGR) or you're pointing out that I have bad programming practices :p

Seggan

why you?

(this is GGR)

lyxal

@Seggan I swear I have something like that somewhere in vyxal 3

nvm it's only a list of String Int pairs

but at some stage I was probably doing something like that

Ginger

01:11

@lyxal link or it didn't happen

lyxal

Chapter 1 of Automata, Computability and Complexity. Theory and Applications by Elaine Rich

9th edition

Ginger

I WANT A [Hyperlink Blocked] NOT A [Peer-Reviewed] [Rip People Off]!

lyxal

Well I can't give you a link

Not legally at least

Rydwolf Programs

oOoooh Ginger tryna break the law

Ginger

~~YAR HAR FIDDLE DEE DEE BEING A PIRATE IS ALRIGHT TO BE~~

Rydwolf Programs

01:15

Worms have been sent to your location

Ginger

ooh yummy

Seggan

@lyxal why not

ah you paid?

lyxal

@Ginger for legal reasons I do not condone piracy or using websites like genesis library or whatever it's called for finding books like that

Ginger

@lyxal me neither, but I wanted to make the joke :p

lyxal

@Ginger I know but that's also what uni workshop leaders say when they talk about textbooks

Ginger

01:18

~~YAR HAR FIDDLE DEE DEE~~

Rydwolf Programs

IMO scientific articles and to a slightly lesser degree, textbooks are like, perfectly valid to pirate. And I'm generally against piracy.

Knowledge should be free

And the publishing industry deserves it

Ginger

~~BEING A PIRATE IS ALRIGHT TO BE~~

Bbrk24

Okay I think I've finished updating the Trilangle online interpreter links

lyxal

@RydwolfPrograms but how should the people who collate the knowledge be compensated?

I don't disagree with you, I just think it's an unfortunate dilemma

Rydwolf Programs

Scientists don't get compensated by journals anyway, and textbook authors can profit plenty from ones bought by schools and stuff

lyxal

01:24

But if knowledge is free, why would schools pay?

Rydwolf Programs

For scientific papers there's literally no moral dilemma I can find. For textbooks it still exists, but not to much of a degree

@lyxal Same reason an enterprise customer pays 10× what I do for SSDs

Bbrk24

One place I'm active has an interesting policy: they generally allow sharing papers etc since the vast majority (if not all) of the money goes to the publisher, and none to the authors. However, some books they don't allow because the author is making money from the sales

Rydwolf Programs

They have the money, they want to do things "the right way", and they care about having the most up-to-date option with the best customer support and stuff

Less the last two with ✨knowledge✨ but still

lyxal

@RydwolfPrograms That's a little different though. Hardware still costs money. But if textbooks as PDFs are free, why would schools pay for licenced copies?

Rydwolf Programs

Because governments, companies, and schools are goody two shoes

They could get free PDFs now if they wanted to

They still do things legit because that's just what they do

lyxal

01:27

But we're assuming that textbook PDFs are free legally

Rydwolf Programs

Well it kinda falls apart once it's legal

lyxal

That's what I was saying

Rydwolf Programs

IMO it should stay ""illegal"" because what we have now is a pretty nice system

Bbrk24

In one of my classes, the former professor wrote the textbook, so the current professor has a legal PDF, and he just uploaded it for all of us to just have

Seggan

so im writing an assembler rn for CGR

and i cant imagine writing the first assemblers in machine code

ATaco

01:38

No need, write it in a second language, then write it in the assembly.

Seggan

my assembler is 101 lines of a modern programming language. if that, then how much machine code did it take for their assemblers?

@ATaco except they didnt have second languages?

ATaco

You can write an assembler in any language, ideally something close to the target

Then, if you wish, rewrite it in the assembly.

Rydwolf Programs

...the point is there weren't any languages yet

Seggan

^

ATaco

Oooooh, Forgive me, Misunderstood.

Seggan

01:40

they wrote everything in machine code, until someone decided that memorizing numbers was too hard

Rydwolf Programs

Worth noting they'd write them in assembly then hand-compile them tho

ATaco

Writing the first assemblers would've sucked. But they were also doing it with Punchcards.

Rydwolf Programs

Not just sit down and vomit some binary

Seggan

@RydwolfPrograms oh yeah true

still, hand compilation mustve been tedious

logic error: did you mis-hand-compile it or did you actually mess up the program?

Aiden Chow

bruuuh guys i just realized that piet doesnt have a < operator, so i cant do (a>b)-(a<b) like the other answers... theres no absolute value either. what are some alternatives???

checking each manually seems like a pain too

Bbrk24

01:49

Trilangle's only branch instruction is "branch if negative", so I just did if isNegative(x) { print(-1) } elif isNegative(x - 1) { print(0) } else { print(1) }

I don't know what Piet's instruction set looks like

Aiden Chow

@Bbrk24 that looks like sign, not compare. i assume u did x=a-b?

Bbrk24

We might be thinking of different questions

Bubbler

@AidenChow Piet does have >

so you just do (a>b) - (b>a)

Aiden Chow

@Bubbler oh yea

that makes things a whole lot simpler lmfao

Bubbler

and I have this

for three-way comparison, I think chainging the leading I to II- in my "sign" is the golfiest at 21 bytes

(if colors work out, didn't check yet)

those rolls are damn heavy

Aiden Chow

01:57

@Bubbler wait what, u mean just adding an inN at the beginning then subtract or what?

dont u have to pad the rows below with black codels if u do that... or am i missing smth here?

Bubbler

@AidenChow yes

it's like pushing a folded rubber sheet, so it just adds two more colored pixels on the new rightmost column

Aiden Chow

oh nice, lemme see what a naive vertical structure would give compared with this

i suspect it will come out quite a bit longer

ok well i got 22 bytes

Bubbler

02:15

cool

that's super close

Aiden Chow

the two rolls is costing a lot tho

and also i just realized that i have to input in the order b a instead of a b lmfao, is that even allowed

Bubbler

well, ask the author lol

Aiden Chow

well im not gonna post the 22 bytes answer anyways lmfao

it would be longer than 22 bytes if order a b is required tho

cuz rn im doing (b>a)-(a>b) instead of (a>b)-(b>a)

Bubbler

I was about to say "it's just one swap away" and then realized one swap is like 4 bytes in Piet lol

Aiden Chow

my code for reference in case there are some obvious golfs: piet.bubbler.one/…

ATaco

02:19

Isn't it just one negation away?

Which I have no idea how easy is in Piet

Aiden Chow

no negation in piet

ATaco

0-n..?

Bubbler

negation is kinda worse, as you need something like 1 2 - * which is cost 5

Aiden Chow

^

ATaco

Ouchy

Aiden Chow

02:20

0 is a bit costly to push in piet as well

Bubbler

or 1 ! 2 1 roll - which is 7

Aiden Chow

u can only push positive integers, so u would have to do some form of arithmetic to get 0

becuz rn what im doing is inN dup inN dup then i do 3 2 roll, which puts a b b a on the stack... then greater than on the b a, roll it to the bottom, and do > again, which results in b>a a>b on the stack, minus then gives (b>a)-(a>b)

oh wait maybe i can just roll the top to the bottom at the beginning

then it would be b a a b

same byte count im pretty sure tho

Bubbler

yeah you can use 4 1 roll I think?

instead of 3 2 roll

Aiden Chow

@Bubbler yep, exactly what im thinking

ye that works

doesnt really matter anyways cuz its still 22 bytes

4 1 and 3 2 both take up the same amount of bytes lol

Deadcode

04:33

@ATaco As the developer of ReRegex, and still actively writing new answers in it, I'm curious... have you run up against the problem of Java having a very buggy regex engine? Of all the engines I've worked with, I've found Java's to be by far the buggiest, and with the hardest to explain bugs. And I tried to write an answer in ReRegex once, but had to scrap my plans, because the bug I ran up against was insurmountable.

Have you ever had to change your method of attack on a problem due to a Java regex bug? Or even had to give up on answering a particular question?

ATaco

@Deadcode Nothing Java Regex Specific, but I don't frequently find myself using the more complex features of it.

Deadcode

Answers where a Java regex bug has been exposed: Tralindromes, Fibonacci reversed!, Converge to a number, Do you make me up?

In every case, I've found the bug to persist in the latest version of Java.

ATaco

What's the Bug? Something Conditional related?

I'm pretty sure I also wrote ReRegex in Java 6..? So if it's a new bug it won't be affected.

Deadcode

Java doesn't have conditionals. And some of its bugs are pretty basic... like, it just doesn't try all the possibilities, i.e. gives up backtracking far too early.

ATaco

One could also Port ReRegex to a different flavour really easily, considering it's a very simple language.

Deadcode

04:50

@ATaco Yep. But then what would be the best way to identify the language in CGCC answers? "# ReRegex (PCRE)" perhaps? Or should it be more explicit, like "# ReRegex (PCRE flavor)" or "# ReRegex (PCRE engine)"?

ATaco

Probably exactly like that, yeah.

Deadcode

05:10

Oh yeah, and one more showing a Java bug (except in this case, Boost and Ruby are bugged as well): Determine if all decimal digits are unique

New Posts

06:40

0

Q: Find the linear transformation

Given a set of bit vectors \$A\$ and a binary matrix \$M\$, we can define the set \$MA = \{ Mx : x \in A \}\$, where \$ M x \$ is the result of the matrix multiplication of \$ M \$ by \$ x \$ over \$\mathbb{F}_2\$, the field with two elements. Given the set \$ A \$ and the set \$ M A \$, your tas...

binary fastest-code linear-algebra

Sandbox Posts

08:24

0

A: Sandbox for Proposed Challenges

It a problem I solved but now I am curious whit what techniques it can be improved. The original language is JAVA but here it doesn't matter. You start with a list of n strings with a length of 7 characters. What is the most efficient algorythm (fastest) to determen the langest substring all the ...

mousetail

09:29

11

Q: SE API multiple-escapes the returned data (at least display names, userlinks, and post content)

The Stack Exchange API will, unless you explicitly disable it, apply HTML escapes to certain strings. However, it will apply this transformation multiple times prior responding with the data. This is most easily visible when querying users. For any username containing the letter Ö, such escaping ...

bug status-review api

Lol this is dumb

lyxal

09:49

Obviously all html escaping should be implemented as escape until invariant

mousetail

What seems to be happening is:

load data from cache -> escape -> store escaped data in cache

lyxal

Ooh that's a fun little loop that won't lead to any problems

On one hand, I can see where that goes wrong. On the other hand, I kinda feel like I'd have done something stupid like that myself in that situation

mousetail

You could probably create a denial of service attack from this bug, by forcing it to escape and store a exponentially lengthening string

lyxal

It's already crashed smokey as indicated in the answer on that post. It probably could frick up the cache

mousetail

If smokey is on fire, then we have real problems. I guess that would create meta-smoke

6

lyxal

09:54

Congratulations you've won the pun of the day award

mousetail

Who won it yesterday?

lyxal

No one

Because I decided that there weren't any winners

Ginger

10:58

@mousetail lmao

CMQ: Is it okay to provide "bad grammar/formatting" as a reason for a downvote? I know I could edit the question/answer to fix it, but I feel like that should be the author's job so that the meaning doesn't get changed; however, not everyone speaks good English

mousetail

In the sandbox, no. On the main site yes

lyxal

@Ginger depends. If it's just a few typos, then no. And if it's still mostly understandable with the errors, then also no. But if it's a serious issue, then maybe. And even then, only if you're feeling particularly nitpicky and willing to waste one of your votes on grammar/formatting.

Ginger

alright

thabks

caird coinheringaahin g

11:40

@Ginger Also, you can vote for whatever reasons you want

Ginger

12:15

time to go back to screaming at Rabbit

I worry that this project is slowly driving me insane :p

mousetail

I kinda assumed you where always insane

lyxal

^

Ginger

Mad I was... until it worked

holy hell

I fixed the recursive type issue! :D

@Seggan

and it was so simple too: all I did was change the signature of instantiate a bit

abstract fun <T: ResolvedType> instantiate(parentContext: Context): RabbitObject<T>

and that can be overridden and have its types inferred

waiiit

shit

goddammit why can't you just work

fixed it (again) by completely removing the generics and making the signature into abstract fun instantiate(parentContext: Context): RabbitObject<*>

it can just be overridden

Bbrk24

12:43

What was the old/recursive signature?

Ginger

didn't have one, because I couldn't make it work at all within Kotlin's syntax

Ginger

12:58

okay, we've gone from 31 errors to 10

this is progress

oop it jumped to 23

:|

Seggan

13:36

@Ginger noice

Seggan

14:08

hrm so ive been messing with a certain error in gradle

and i finally fixed it...

by going to the link in the docs offered to me by the error smh

i should learn to use the docs more

@Bbrk24 you said mingw bash doesnt work well with vscode, have you ever considered using WSL?

Bbrk24

MinGW bash works fine with VS Code, it's gradle that it interacts poorly with

Gradle tries to overwrite text in-place, but I guess MinGW's ioctls are different than gradle expects, because that utterly fails

Seggan

@Bbrk24 ah

Ginger

14:46

me when 30 errors in my file

ok, we've gone from 9+ to 9

@Seggan what exactly is the type property used for in TNodes?

I see it being used for return types, variable types, lots of things

Rydwolf Programs

15:32

Well I'm dumb

I spent ages waiting for my genetic algorithm to start giving me better results then remembered I hadn't actually implemented the natural selection part yet

Ginger

lmao

@Seggan really what I'm trying to figure out is if my TypedNode's type parameter should be constrained to just ResolvedStructTypes

I think so, but I'm not 100% sure

Seggan

16:06

@Ginger ... for knowing the type of the node?

@RydwolfPrograms ooh genetic algorithms

what for?

Rydwolf Programs

Color names

Seggan

@Ginger whats the hierarchy of type classes?

@RydwolfPrograms why

how

Ginger

16:36

@Seggan oh boy you're gonna love this :p

@Seggan Type has children UnresolvedType and FunctionType; UnresolvedType has children UnresolvedStructType and UnresolvedFunctionType; ResolvedType has children ResolvedStructType, ResolvedTraitType, and ResolvedFunctionType; ResolvedStructType has children NothingType and NeverType (both singletons) and PrimitiveType; ResolvedFunctionType has child WrappedResolvedFunctionType

I think that's correct

Seggan

@Ginger yeah in that case make it Type as it can take a FunctionType

Ginger

hmm, right

will do once VSCode stops dying

mmm delicious 100% CPU usage on both cores

Seggan

17:05

ok so i finished the assembler for CGR

loadc "Hello, World!"
print

produces

0000: de ad c0 de 03 a6 53 54 52 49 4e 47 a5 4c 4f 41   ......STRING.LOA
0010: 44 43 a5 50 52 49 4e 54 01 00 ad 48 65 6c 6c 6f   DC.PRINT...Hello
0020: 2c 20 57 6f 72 6c 64 21 02 01 00 02               , World!....

scream very loudly is

label 0
loadc "A"
rprint
jump 0

and i can confirm the generated file works

Ginger

@Seggan you're just encoding the opcodes as plaintext???

Seggan

17:21

yep

due to the extensibility of the vm

Ginger

why?????????

Seggan

dont want id clashes

yesterday, by Seggan

we cant have conflicting byte ids

Ginger

okay fine I guess

still feels jank to me

Seggan

yesterday, by Seggan

ive had to suffer though byte id hell when writing certain minecraft plugins

yesterday, by Seggan

until the parent plugin switched to string ids, much easier

Ginger

FINALLY

ZERO ERRORS IN Type.kt

now on to the next file! :)

Seggan

17:25

@Ginger the strings are only stored once, and referred to by their index in the table, so no unnecessary dupes

the actual code is the last 3 bytes

Ginger

time to fix the 12 errors in Object.kt

also I guess I'm changing the type property of RunnableNode to ResolvedType

mmm I don't know if this is a good idea

@Seggan there are a lot of differences between function objects and struct objects

Seggan

40 mins ago, by Ginger

hmm, right

@Ginger theyre still types

Ginger

function objects don't have fields

although actually I guess that's not a giant issue

because fields are only accessed in one or two places

Seggan

2 mins ago, by Seggan

@Ginger theyre still types

Ginger

17:46

zero errors in Object.kt!

Seggan

noice

Ginger

also, nearly 2k sloc

Rabbit is now officially the biggest project I've ever made

although to be honest the experience of constructing it has been akin to being thrown off a cliff with several crates of parts and building an airplane on the way down

Ginger

18:15

@Seggan advanced Kotlin annotations: @RabbitFunction<PrimitiveNumberType>("+", Visibility.PUBLIC, PrimitiveNumberType::class)

I hate the redundancy here but it's needed in order for everything to work correctly

@Seggan can I have some advice?

I have an issue that I'm not sure how to solve

Rydwolf Programs

18:42

How should this answer be dealt with?

I guess for one it's invalid since it takes input with a variable

For two it's a copy of several existing answers

For three the source they took it from is MIT licensed

I've pointed out #1 and #2 to them, didn't suggest any action tho

Seggan

@Ginger what

@RydwolfPrograms which requires a license copy

i suggest we delete it

Rydwolf Programs

Yeah, but the code is so unoriginal and trivial that I don't think it can be feasibly claimed to be copyrighted

Seggan

true

Ginger

19:08

@Seggan so basically all subtypes of ResolvedType have an instantiate method which creates a new instance of them; I want to make ResolvedStructType's instantiate method take a list of args but I can't do that because it wouldn't be an override anymore

Seggan

sounds like you should make instantiate an external function

as in, not in the class

fun instanticantspellthisword(type: ResolvedType, ...)

or an extension function

fun ResolvedType.inwhatever(...)

then you can make pseudo-polymorphism by using when

Ginger

alternately I could make the args a property of the abstract method and just not have all implementations use them

nah

actually, I could use an extension function to pull this off

probably

Bbrk24

20:43

I keep expecting Kotlin's to operator to produce a range rather than a pair (like what's actually the .. operator).

Also, why can arrays not be const?

Seggan

21:06

because they cannot be JVM constants

@Bbrk24 you are looking for upTo

Bbrk24

...right. Of course to and upTo both exist

Seggan

theres also until and i think rangeTo

and downTo

Seggan

21:26

one thing that annoys me about kotlin is that it doesnt define a default companion object

leading to this behavior and not allowing to make nice List[1, 2, 3, ...] extensions

lyxal

21:58

@RydwolfPrograms duplicate answers are allowed

Rydwolf Programs

Ik

But invalid duplicate answers that violate copyright laws typically aren't :p

lyxal

It wasn't 100% copied from the linked repository :p

Also, you only need the notice for copies of the full software or substantial portions with MIT

> The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

Rydwolf Programs

Yeah the copyright isn't really any issue

I just needed three points for it to sound nice

lyxal

@RydwolfPrograms this is probably the first time in a while I've actually properly read the mit licence :p

Rydwolf Programs

Everyone knows "MIT license" actually just means "you can have my code I'm too broke to sue you"

ATaco

22:07

You can have my code because I'm too flattered to complain.

lyxal

To me it means "use it and if it blows stuff up that's on you idiot"

That's why I choose MIT :p

"do whatever the hell you want, but don't blame me for anything"

user

I just pick it because it seems to be the default and having a license makes my project more professional-looking

No one's ever going to use any of my code so it doesn't matter anyway :P

Rydwolf Programs

I'm just going to start using a license that mandates everyone to use my code in their projects

lyxal

@user all the cool kids with golfing languages use MIT :p

user

"By reading this license, you agree to pay me $50"

lyxal