The Nineteenth Byte

ATaco

01:02

CMC: Given a non-negative integer of at most 2^31-1 (2147483647), determine the number of rotations in no particular direction that results in the largest integer. Direction of rotation (left or right) is programmers choice, and may be constant or not.

caird coinheringaahin g

Digit rotations?

ATaco

Correct, Digit rotations.

Decimal digit rotations in particlar.

Input may be as an int, a string, an array of numbers, a stream, etc.

caird coinheringaahin g

Mind providing a quick example? I'm guessing 123 -> 1 (a single rotation to the right provides 312)

ATaco

123 -> 1 (Rotated RIGHT) 19298 -> 3 Rotated LEFT (Gives 98192)

Dannyu NDos

Leading zeros discarded?

lyxal

01:09

@ATaco Vyxal, 39 bits1, 4.875 bytes: ẏvǔṅÞM (vyncoded)

yes I can chat markdown why do you ask

ATaco

Oooh, Vyncode is new to me.

lyxal

well, tbf it did only get added to the base language last Saturday

caird coinheringaahin g

@lyxal of course

I'm not even surprised that Vyxal now has a way to have fractional byte encodings, and tbh, it looks more interesting that most other langs with it, but I still wasn't expecting it

lyxal

neither was I

but then one day I saw this:

in Vyxal, Apr 20 at 21:00, by AndrovT

I made a range coder for Vyxal to see how well it works and it looks like it decreases length by about 10% on average.
https://github.com/AndrovT/Vycoder Here is the code if anyone wants to play around with it.

(and yes, vyx-µ was a thing a while ago, but for some reason it never got incorporated)

(besides, range coding is apparently better than huffman encoding)

> looks more interesting

It's more like packed stax than nibbles

The CMC was just formatted like that for convenience

The actual bit string is 101110100011100010011010001110010001110

Apr 21 at 1:36, by lyxal

CMQ: Say I have a language where programs are bit strings of any length. And say the length of the bit string is important, meaning that padding it to a whole byte is out of the question. Could I use an encoding like a = 0, b = 1, c = 10, d = 11, e = 100, f = 101... up until any arbitrary length to shorten the bit string for code golf answers?

why else do y'all think I asked that :p

ATaco

Encoding theoretically never matters as long as it'll run :P

lyxal

01:23

y'all might be interested in the fact that the encoding utility is written in scala, and turned into js for the website via scala.js and bundled as a jar for the offline interpreter

cursed, yes, but the best solution we could think of

user

01:58

Not sure if this is the right place to talk about it but we could use it like a library instead

Either publish to Maven Central or use Git submodules

Anyone here used Git submodules to be able to use another one of their projects as a library without publishing to some package repository?

And if so, how painful was it to get set up?

Any hacks required to keep it working and updating the submodule?

Bbrk24

02:24

@user I've used submodules once but I never had to update it so I can't fully answer it. It was painful to move the submodule from one folder to another though

lyxal

@user wdym library?

How tf is python gonna use a maven library?

Seggan

maven isnt just java

gradle can famously work for c++ and swift

lyxal

@Seggan Mmm except in this case this is java/scala

Seggan

@lyxal most kotlin packages are on maven central as it seems to be the only repo that can handle kotlin multiplatform

Unrelated String

like the direction of rotation doesn't actually have to be consistent, your program just has to output either the correct number of right rotations or left rotations for any particular input? ooh
i'm having a hard time thinking of what kind of solution could make that useful but i'd love to see one

lyxal

02:31

@Seggan but getting the raw Java/scala isn't going to help. You'd need to upload just the jar to maven to make it of any use

Bbrk24

@Seggan It can WHAT

Seggan

@lyxal yeah

@Bbrk24 yep :P

Bbrk24

There's already CocoaPods and Carthage and SPM, we don't need gradle too

Seggan

apparently it can handle any lang, but builtin support is java, kotlin, scala, groovy, c(++), swift, and javascript

Bbrk24

Why Swift? /gen

Unrelated String

02:35

anyways DṙJ$M for 5 in jelly with consistent left rotations; remove the D$ if digit list input is allowed

Seggan

@Bbrk24 I’m thinking bc perhaps they needed to support kotlin/iOS and adding official swift support was trivial

Bbrk24

> kotlin/iOS

you what

Seggan

yeah bc kotlin/iOS may need to have some code written in swift

@Bbrk24 never heard of kotlin multiplatform?

Bbrk24

Like React Native but worse

Seggan

there’s kotlin/jvm, kotlin/android, kotlin/iOS, kotlin/native, kotlin/js, and (new) kotlin/wasm

@Bbrk24 no

better

js and anything to do with it is near universally bad

Bbrk24

03:03

I disagree

@Seggan ergo kotlin is categorically bad

more seriously, JS has improved a lot recently, especially with certain preprocessors such as what modern Angular does

Switching very suddenly from Angular 1.6 to 14.2 was an experience, but after the adjustment I can say I vastly prefer the newer one

user

03:31

@lyxal Oh you mean for v2, mb

@Bbrk24 Aww, okay

If moving stuff is gonna be painful, best to avoid them, I guess. I try to stay away from Git black magic :P

@Bbrk24 I mean, Gradle supports a TON of languages, doesn't mean it's actually used for them much. Some company out there may be a Java shop that only occasionally makes Swift stuff, so they might prefer to stick with Gradle even for Swift

I bet there's people out there using Make for languages that are nothing like C/C++

@lyxal ~~Jython~~

lyxal

@user oh hell no. Ain't no way

Man's really suggested the goofy version

user

You'd have to ditch PA, of course

lyxal

@user there exists make for js called jake

user

aaaaaaaaaaa

Jake more like Joke

lyxal

@user That was UnrelatedString's reaction too

ATaco

03:39

@UnrelatedString I'm of the school of thought that where possible, a challenge should remove as many restrictions as possible, to open as many strange solutions as possible.

user

based

lyxal

Nov 18, 2022 at 5:45, by Unrelated String

wait wait wait is jake fucking
make
in JAVASCRIPT

user

Quick, delete that before Unrelated sees it again or they'll have to take benadryl and go to sleep again

@UnrelatedString Please ignore the messages above this

Bbrk24

03:58

and the ones below it for good measure

lyxal

ignore this message if I can take all your money

lyxal

04:26

I think there's been a change to chat that's broken one of the userscripts I use - something about s.data.replace being undefined when I use chat with userscripts keeps coming up in the error console. And it makes firefox start using a whole lot of RAM

Unrelated String

05:18

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

lyxal

jake, me taking your money, or the userscript thing?

caird coinheringaahin g

I get this weird feeling, apropos of nothing, that Unrelating String is having trouble with something

Unrelated String

@ATaco yeah same

this is just a way i've never seen it done before

and it's so cool

Nov 18, 2022 at 5:45, by Unrelated String

this isn't funny anymore

caird coinheringaahin g

Apr 16 at 21:01, by caird coinheringaahin g

@RydwolfPrograms Disagree

Jan 28, 2022 at 21:07, by caird coinheringaahing

People can be very funny when they find a very bad idea :P

lyxal

in Sandbox, 7 secs ago, by lyxal

ooh are we doing that talk in replies thing again?

caird coinheringaahin g

05:23

Sep 22, 2021 at 21:35, by caird coinheringaahing

Ah, yeah, that's cheating

lyxal

Jan 20 at 0:12, by lyxal

don't care + didn't ask + L + ratio + nerd

caird coinheringaahin g

On the other hand, if I keep talking like this, then I'll eventually run out of messages to quote

Unrelated String

Jan 9, 2022 at 3:03, by caird coinheringaahing

Possibly

lyxal

Mar 22 at 22:10, by lyxal

@user "hey wait gosh dang it not again"

caird coinheringaahin g

I regret this

lyxal

05:26

in Snowpaw's Den, Apr 25 at 22:39, by lyxal

@Snowpaw puts you in microwave for 3 minutes 30 seconds

caird coinheringaahin g

Jun 9, 2017 at 15:57, by caird coinheringaahing

@programmer5000 i dont think so

lyxal

in Snowpaw's Den, Apr 20 at 6:16, by lyxal

Oh well worth a shot.

in Snowpaw's Den, Apr 20 at 6:16, by lyxal

@Snowpaw there'll be plenty more chances.

caird coinheringaahin g

05:48

It's like 7am here, my brain doesn't work yet :P

lyxal

Side topic: it really is a shame SE isn't paying attention to community ads very much anymore

caird coinheringaahin g

Honestly, I wouldn't care so much about them advertising to us, so long as we had competent community ads

mousetail

06:10

Time for a community adds userscript :)

user

06:36

Aug 1, 2021 at 18:15, by user

Missed a fun conversation :/

Unrelated String

07:28

Jul 4, 2017 at 21:12, by caird coinheringaahing

now I'm sad

lyxal

2 hours ago, by lyxal

in Sandbox, 7 secs ago, by lyxal

ooh are we doing that talk in replies thing again?

mousetail

On topic: We currently have 21 imgur links that are going to die soon: https://data.stackexchange.com/codegolf/query/1747560/imgur-links-that-are-about-to-go-down

Anyone want to help me edit them out?

This gist has fixed links: raw.githubusercontent.com/Glorfindel83/SE-Userscripts/master/…

lyxal

@mousetail there's a userscript for this

I'll find it real quick

github.com/Glorfindel83/SE-Userscripts/blob/master/archivist/…

It archives images and links and gives you a replacement to edit in

mousetail

07:44

Yea I already downloaded the spreadsheet with all the replacements

That's my second link

Fixed SEDA link: data.stackexchange.com/codegolf/query/1747560/…

Dannyu NDos

08:32

Does any known CPU architecture implement fused multiply-add for signed integers?

Dannyu NDos

08:45

Nvm, I planned to make an OISC doing that, but I don't want the instruction length to be too long.

In my current sketch, the pseudocode is: if a == b then goto a else [*a, *b] = *a - *b

2^32 registers with 64 bits each; that's 32GB in total.

mousetail

That's not really registers anymore, that's normal RAM

Dannyu NDos

Yeah, just really quickly accessible RAM.

Would require a lot of silicon tho.

(Help me, Samsung!!)

Neil

@mousetail (for what reason?)

mousetail

34

Q: Imgur TOS change on May 15th 2023 could cause the loss of thousands of Stack Exchange images!

On May 15th 2023, Imgur will implement its new TOS, in which We will be focused on removing old, unused, and inactive content that is not tied to a user account. I, and presumably many other users, have uploaded things to Imgur without logging in, and then posted those images to Stack Exchange,...

Dannyu NDos

09:03

Wait, if a byte is supposed to be the smallest addressable unit of memory, does that mean 1 byte = 64 bits in my OISC?

Such an overkill to encode a character.

mousetail

@DannyuNDos Yep

But only if there is no single instruction to access less values

lyxal

20

A: Is the change of policy of Imgur going to create problems for all posts with images on SO?

Nope. SE runs a corporate version of Imgur, so image persistence is guaranteed. I also see no valid reason for nudity on SO, so that seems extremely unlikely to have any (negative) effect.

mousetail

@lyxal That's why I'm specifically searching for images not posted with the corporate version of imgur

And replacing them with images that are hosted on the corporate version

Images on stack.imgur.com/* will work, but any other .imgur.com images will be deleted probably

Unrelated String

^

Ginger

10:50

May the 4th be with you all

Dannyu NDos

May the momentum prime be with you, too.

Ginger

I have a feeling that this Imgur TOS change is at least partially for freeing space on their servers :p

Unrelated String

11:13

crazy

mousetail

I can't manage to get this gif under 2 MB, anyone else want to try? codegolf.stackexchange.com/a/111121/91213

emanresu A

Crop it?

mousetail

Want to try?

Bbrk24

11:31

> while cd

what

emanresu A

I had to crop it, maximally compress it, delete every 4th frame and slow it down 4x to counter that.

Unrelated String

@Bbrk24 probably the shortest way to get a successful return code :P

mousetail

@emanresuA TY

emanresu A

You may find ezgif.com helpful, I'm going to bed now

Unrelated String

just make sure that whatever you need it for does in fact mandate the gif format :P

o/

oh for porting to stack imgur

hmm

Unrelated String

11:47

in any case this is what just naively cranking the bitrate down in ffmpeg does

Bbrk24

@UnrelatedString What about :?

Unrelated String

ooh

true

maybe fish doesn't have it

mousetail

@UnrelatedString Getting the size down if converted to webp is very easy but imgur doesn't support it

Unrelated String

f

and it also seems that proper video formats are right out

Bbrk24

APNG? /hj

mousetail

11:56

Could work

Ginger

JPEG2000

mousetail

But it's fixed now so I don't really have any need to put more effort into it

4

Q: Could we support WebP image format when uploading images?

SE is a leading example of top multimedia streamlining for ease of use. WebP is currently unsupported when attempting to upload an image. It's fast becoming a very common image format because it is superior to other formats in various ways. If .webp becomes useable on SE it could be a good upgrad...

feature-request images uploader

Bbrk24

12:31

Can you link to badges like you can tags?

Unrelated String

probably not but good question

lyxal

[badge:code-golf]

no

Bbrk24

my only post on workplace is at 2.5k views and voted +12/-9

Ginger

a bit of a controversial subject, that :b

Bbrk24

you get a silver badge for 2.5k views so within 36 hours of joining I got a silver badge

it’s also my first post to make it to HNQ

lyxal

12:34

@Bbrk24 second

Bbrk24

First one SE decided to send me a notification that it reached HNQ

lyxal

must be a workplace thing

Bbrk24

(was the other one my sort by duplication?)

lyxal

it was

I've never received a notification for HNQ status before

Rydwolf Programs

13:41

Yeah that must be new or sth

mousetail

Do mods get automatically flagged if you do a lot of edits?

New Posts

14:10

0

Q: Prefix notation in JS

This challenge requires me to create an anonymous function that given a string in prefix notation, must return the result. The string can only contain the following operations +, -, *, /, ^ and integers (negative, zero, positive, etc). I've come to a point where I can't reduce my function any mor...

code-golf

Rydwolf Programs

15:11

@mousetail I don't think so, but it does clutter the homepage either way, which is very visible

holy shit you were not lying when you said "a lot of edits"

usually you should ask on meta before doing more than like 4-5 edits at once

Ginger

jesus mousetail

Rydwolf Programs

my favorite religious deity

we all remember when he died on a mousetrap so we could all have cheese

Ginger

15:30

ಠ_ಠ

mousetail

@RydwolfPrograms Still about 200 to go

Planning to do about 50 a day

Rydwolf Programs

4-5 a day is typically the recommended rate

What exactly do you need to edit so much

mousetail

We only have 2 weeks

It's not going to be done in time if I do it that slow

8 hours ago, by mousetail

On topic: We currently have 21 imgur links that are going to die soon: https://data.stackexchange.com/codegolf/query/1747560/imgur-links-that-are-about-to-go-down

Anyone want to help me edit them out?

Number was wrong, there are actually like 300

Rydwolf Programs

ohh

I saw that on another site's meta, didn't realize it applied to us too

Opinion: Restricting who can edit what in order to maintain the front page is dumb

mousetail

It applies to the entire internet lol

Reddit is going to be a ghost town too

Rydwolf Programs

15:39

"noooo don't [necropost/do ...], you're ruining the front page" is what makes other forums absolutely suck

We shouldn't be sacrificing moderation for...what benefit even is the active page?

mousetail

There really should be a way for high-rep users to edit without bumping

Maybe then going to some sort of queue for review

Rydwolf Programs

^

Maybe not a queue, just like, a page in that useless 10k tools thing

mousetail

That would be fine too

Honestly, I'd appreciate if there was some system where someone checked my work. It's entirely possible I made some mistakes

Rydwolf Programs

Yeah but the review queue could be the bottleneck then

For mass edits the 20 reviews/person/queue/day would be annoying

mousetail

You would need 2 people working in tandem

Reviewing each others posts in the queue when it fills up

Ginger

16:01

someone's probably working to archive as much of imgur as possible before the cutoff

mousetail

I don't think there is actually any kind of index though so you could list the contents

So you'd need to crawl the internet for links

Bbrk24

16:35

@RydwolfPrograms recent meta said 10/day, I thought?

Sandbox Posts

0

A: Sandbox for Proposed Challenges

2-Cyclic Quine In this challenge, you have to make a programme that satisfies these properties: The programme must output another valid programme. This second programme must output the original programme. It must obey all the rules of quines (see exception below). It must not print itself. This...

mousetail

@Bbrk24 it's a different situation if you are editing your own posts or doing a moderation task

The Thonnu

16:56

@TheThonnu then why is the tag still here? — Hippopotomonstrosesquipedalian 3 mins ago

Anyone know how to answer that better than I did?

mousetail

Popularity contents are still allowed

But only in very exceptional circomstances

Seggan

17:57

As far as I can tell it’s for the slightly subjective contests

New Posts

0

Q: I'm in the process of replacing all ananymous imgur images with stack.imgur images, and I need your help

As stated on mother meta all imgur images not associated with a account or the "stack.imgur.com" official corporate instance will be deleted on May 15th. If we don't replace those links before that date, they will be permanently lost. Some questions and answers may lose unessential fluff but many...

discussion

Bbrk24

someone made another TC set of Swift types

DLosc

20:02

The flag of Bangladesh is hard on colorblind people...

Jacob

20:19

LDQ: When, if ever, is it okay to sacrifice consistency for readability?

Rydwolf Programs

Can you give an example?

Seggan

^

Jacob

I realize that was a bit too general lol

Gimme a sec

Rydwolf Programs

I think a lot of the time consistency is in service of readability, so if they're conflicting, readability is probably a semi-safe bet.

Seggan

whoah

TIL Voyager 1 is 22 light hours away

thats far

Rydwolf Programs

20:23

How many hours has it been since it departed tho? :p

Seggan

a lot

but still, 22 light hours is... considering the moon is a light second away

user

@RydwolfPrograms At least 22

Seggan

still a while till a light year

Rydwolf Programs

@user Not necessarily, it may have travelled through a wormhole

Seggan

just a bit till light day

user

20:24

@Jacob Like, adding sugar for some specific case?

Rydwolf Programs

Added sugar is unhealthy. Minimize how much you use.

DLosc

@RydwolfPrograms But a spoonful of sugar helps the medicine go down, so in that case it may be a net positive for your health

Jacob

hmm okay

Rydwolf Programs

@RydwolfPrograms (this was a joke btw. syntax sugar is often good)

Jacob

the context is i’m making a language that’s somewhat golfy but less general-purpose

okay thank you for the feedback 👍🙏🪆🛎️💰

Rydwolf Programs

20:31

thumbs up clapping...russian nesting dolls...bell money?

You used this language to get hired as the well-paid bellringer in a russian church? :p

Seggan

20:47

*orthodox

Bbrk24

21:06

@RydwolfPrograms 🙏 isn't clapping, clapping is 👏

🪆 is just a box on my end

DLosc

@Jacob Still a bit vague even with the provided context, so my answer is "sometimes" ;)

Jacob

22:01

i can give a more specific example, i was on my phone on the train earlier so i couldn't really type

the idea i had was a construct for doing lookup-table-ish mappings

Dannyu NDos

How do OISCs deal with syscalls?

Jacob

where it's basically like indexing but somewhat more readable and concise

ATaco

@NewPosts Damn that's kinda cringe ngl.

Jacob

like "b"→"abcd"→[5, 23, 98, 2] would be 23, something like [5,23,98,2]["abcd".indexOf("b")]

ATaco

That doesn't seem particularly readable to me, as the two → symbols perform different actions.

Seggan

22:17

@DannyuNDos maybe a certain memory value or jump location?

Dannyu NDos

But register and volatile cannot mix?

Neil

@Seggan even the sun is only about 8 light minutes away, so 22 hours...

Bbrk24

Swift generally favors legibility over brevity. As a counterexample, fuckingclosuresyntax.com

lyxal

22:39

@Bbrk24 that just looks like scala syntax with extra steps

Bbrk24

There’s a lot of orthogonal things on display there

lyxal

> var closureName: (ParameterTypes) -> ReturnType = { ... }

Imagine not having inferred types

Bbrk24

- you can omit the return type if the compiler can infer it
- you can omit the argument types if the compiler can infer them
- you can omit the parens if you omit the argument types
- you can use `$0` and `$1` instead of explicitly listing the arguments
- you can omit the word `return` in a non-Void closure iff the body is a single expression

lyxal

@Bbrk24 as I said, Scala with extra steps

Bbrk24

- you can use the same syntax for unapplied method references as for unapplied operators

hence .sorted(by: >)

lyxal

22:43

Still scala with extra steps

Bbrk24

@lyxal You can have multiple trailing closures

Rydwolf Programs

wait I just thought of the most cursed thing

calling swift users swifties

Bbrk24

Button { doSomething() } label: { Text("Click me!") }

@RydwolfPrograms yes that is a thing

Rydwolf Programs

lol really

Bbrk24

@lyxal Does Scala have $n? I would’ve thought it uses it like Kotlin

lyxal

22:45

@Bbrk24 _

Dannyu NDos

I wanted to be a Rustacean.

And we have Go → Gopher.

lyxal

@Bbrk24 scastie.scala-lang.org/D0T6yeWjQ6ShlTpzn0pQPA

Bbrk24

@lyxal Oh yeah that too. var ignoresFirstTwo = { (_: Int, _: String, x: Double) in x }

lyxal

in x looks goofy

@Bbrk24 also, imagine not having last expression = return value

Bbrk24

@lyxal I meant like { $0 > 3 } as opposed to { x in x > 3}

lyxal

22:48

_ > 3

Dannyu NDos

(>3) as a Haskeller

Bbrk24

@lyxal Okay then, what about let f: (Bool, Int, Int) -> Int = { $0 ? $1 : $2 }

@lyxal That is a subject which has been vehemently debated on the Swift forums. The current consensus is that it’s not something that should be done, and regardless in some cases it would be breaking to add it

lyxal

@Bbrk24 you'd have to ask user

Bbrk24

@user ^

lyxal

@Bbrk24 scastie.scala-lang.org/DOU3eKFUSYKGQ1tqaTYlEw

Underscores all the way

val x : Function3[Boolean, Int, Int, Int] = if _ then _ else _
x(true, 3, 4)
x(false, 3, 4)

Bbrk24

22:59

you can what

okay then uh

let f: (Int, inout Int) -> Void = { $1 += $0 }

inout is equivalent to C# ref

lyxal

That's called bad code

That's not very functional programming of you

inout params are cringe tbh

Bbrk24

@lyxal Tell that to reduce(into:_:)

> Here's my favorite footgun in Swift:
```swift
func flatten1<T>(_ arr: [[T]]) -> [T] {
// O(n)
return arr.reduce(into: [], +=)
}

func flatten2<T>(_ arr: [[T]]) -> [T] {
// O(n^2)
return arr.reduce([], +)
}
```

emanresu A

That's just bad language design

Also, how

Bbrk24

As to "why": += does the exponential (amortized O(1)) growth and can use realloc, while + has to allocate a whole new array to store the results

lyxal

@Bbrk24 okay then

Bbrk24

23:04

The compiler is allowed to optimize out your accidentally-quadratic performance, but not required to

lyxal

@reduce(into:_:) inout parameters are cringe, as are you

Not having a reduce in O(n) by default is a skill issue

Bbrk24

Perhaps writing it out in full instead of using the shorthand syntax will make the performance difference more obvious

reduce itself is O(n), it's the allocations that aren't.

lyxal

Skill issue :p

Bbrk24

reduce(into: []) {
  $0 += $1
  // appends into the existing reserved capacity,
  // only allocating more as needed
}

reduce([]) {
  let ans = $0 + $1
  // because all three coexist, you have to allocate new memory for ans
  // that is where the quadratic performance lies
  return ans
}

it's the same idea as using push vs concat

lyxal

Why wouldn't the language create a variable ans = $0 and assign to that

Bbrk24

23:08

neither reduce is inherently O(n^2), it only works this way with dynamically-sized collections

besides, there's already .flatMap { $0 } which is shorter and more performant

@lyxal Again, it's allowed to, but not guaranteed to. Especially if you use a custom collection type instead of Swift.Array

lyxal

@Bbrk24 that sounds like a design flaw then

And really, custom collection types shouldn't change how it works

That doesn't sound like very good oop design

Bbrk24

If you're calling into a third-party library that only provides + and not +=, it simply cannot make that optimization

lyxal

Well that sounds like it breaks once you use operators not already defined

Like what if I want to use something that doesn't have an augmented assignment method

Bbrk24

Also, there's a reason you don't see reduce too often in the first place

Like I've definitely used it before, it has its place, but it's no worse to just manually write the for loop

lyxal

That's not very functional programming of you :p

Bbrk24

23:16

@lyxal I'm genuinely curious: do you even have a way to do this that isn't O(n^2)? I'm not convinced it's possible

lyxal

That's what I'm saying

The functionality of reduce into becomes wasted in that case

Bbrk24

reduce is equally powerful as for. It's literally just a matter of imperative vs functional style

You can write an O(n^2) for loop to flatten an array. You can write an O(n^2) reduce call to flatten an array.

This shouldn't be surprising. I only call it a footgun because of how easy the latter is to write

@lyxal Okay, non-inout example. let f: (Double, Double) -> Double = { $1 / $0 }

The point was less about the inout and more about changing the order

user

23:37

@lyxal Function3[Boolean, Int, Int, Int] can be (Boolean, Int, Int) => Int, I think

@Bbrk24 Wait what's the question

Bbrk24

how do you reference arbitrary parameters in a Scala closure without naming them

user

Oh you can't, there's only underscores

It doesn't have the dollar sign thing

Bbrk24

just _ only works if you don’t reorder them and or use them multiple times

user

Yeah

Bbrk24

I’ve used the dollar sign thing in a lot of code golf

user

23:39

To be fair, in real life, I try to avoid underscores if it gets over like 10 characters

But yeah, _0 or something would be nice to have for code golf

Bbrk24

Like I reference $1 twice in codegolf.stackexchange.com/questions/260075/…

I’ve done it in practical code too, I just have no examples onhand

Ah here’s one: sections.map { SectionContent(id: $0.key.position, heading: $0.key.label, items: $0.value) }

I don’t think referencing $0 thrice there is unreasonable

user

Scala wouldn't even be able to reference the parameter once because it's inside parens

Bbrk24

lame

user

Bad for code golf

But good for keeping people from doing shenanigans irl

@Bbrk24 I would probably use a named parameter here personally

Actually maybe not

Bbrk24

The other thing is: if you want to discard the first parameter you can say { $1 }. If you want to discard the second parameter, you have to write { x, _ in x }

user

23:48

Guess that makes sense

Transcript for