The Nineteenth Byte

Tsumiki Miniwa

00:00

Are any of them more efficient than console.log()?

I'm ... not sure if I need to run a console.log() at the very end

(Maybe I don't need to run a console.log())

Seggan

not if u use a lambda

Adám

@TsumikiMiniwa V8 has print()

Bbrk24

Only debug builds, I thought. Node 14 certainly doesn't expose it

Besides, you're allowed to use alert()

Tsumiki Miniwa

@Seggan Is that legal if I'm using multiple functions? I'm doing the "Golf a number bigger than Loader's number" challenge and I highly doubt I can do that in one function

Bbrk24

you're allowed to say f=()=>"foo";g=()=>f and then the function you need is called g

f=()=>"foo" is treated the same as function f(){return"foo"} for the purposes of code golf

Tsumiki Miniwa

00:06

Guess I'll do that, then.

I assume g=_=>...code... would be better.

Bbrk24

Yes, _ is better than () if it works either way

Tsumiki Miniwa

Great.

So no difference between the various types of JS for this, then.

Adám

Why doesn't JS allow ==> instead of =()=>‽

Bbrk24

That could be parsed as == > which is obviously illegal

Seggan

@Adám your interrobang makes me wonder if theres a character whose sole purpose is to combine the chars beside it

like a binary combining diacritic

Adám

00:09

@Bbrk24 And hence there's no ambiguity.

Bbrk24

something something maximal munch

Tsumiki Miniwa

@Seggan start extracting random bits out of some database by entering text input with that diacritic a million times

Seggan

lol

my username is a combining diacritic

@TsumikiMiniwa i wanna try that now :P

Bbrk24

@Seggan Like a ZWJ? Or like U+0361?

Seggan

@TsumikiMiniwa i meant binary as having 2 inputs

@Bbrk24 like the second but merging the character after it into the character before it

Tsumiki Miniwa

00:12

oh

if it combines the character before and after it

Seggan

yeah

Bbrk24

ZWJ does that for some pairs of characters in some fonts, like uh

Unrelated String

@Seggan i know there are those cjk combiners but i don't think there's anything that just serves as a negative space

Tsumiki Miniwa

put two of them together, upload it on social media, crash all poorly-designed browsers and apps

Bbrk24

ᛖ‍ᛞ

In Babelstone Runic Beowulf

Unrelated String

00:13

you have a font that gives you bind runes???

Adám

@Seggan ‽ has a distinct purpose, though; indication a rhetorical question.

Seggan

@TsumikiMiniwa your browser be like zalgo

@Adám i know

Bbrk24

@UnrelatedString Yeah, the Babelstone Runic families do

Unrelated String

nice

Adám

F=ㅤ=>"Hello World"; print(F()) Try it online!

Bbrk24

00:15

ᛞ‍ᛖ becomes:

Seggan

@Adám whaaaaaaaaaaaaaaaaaat

Bbrk24

Seggan

how many chars is that

Bbrk24

@Adám Did you use an invisible character that JS allows as an identifier

Adám

@Seggan You don't know JS, it seems. ^

Tsumiki Miniwa

00:15

@Adám oh that's interesting I put that chara into my "view non-printable unicode characters" thing and it's invisible in that too

\u3164

compart.com/en/unicode/U+3164

Adám

@TsumikiMiniwa It is printable. A normal (but zero-width) letter.

Seggan

@Adám nope

Tsumiki Miniwa

zero-width?!

it's not zero width on my screen

Adám

It is supposed to be.

Bbrk24

Oh yeah, the Mongolian Vowel Separator tends to break things. If you include one in a Discord message, iOS fails to render the entire message

Seggan

00:17

@Adám zero width space?

Adám

No, zero-width letter.

Seggan

mm

Bbrk24

᠎

that thing

Adám

And since various length sequences of them form distinct identifiers, you can do things like:

Jul 5, 2022 at 14:00, by Adám

avg=ㅤ=>ㅤ.reduce((ㅤㅤ,ㅤㅤㅤ)=>ㅤㅤ+ㅤㅤㅤ)/ㅤ.length;avg([3,1,4,1,5])
2.8

I find this much more readable than with visible identifiers. Very similar to APL's +/÷≢ where +/ is reduction using + and ≢ is the length function.

Of course, if just a one-off, there's no need to name the function, so this will do:

(ㅤ=>ㅤ.reduce((ㅤㅤ,ㅤㅤㅤ)=>ㅤㅤ+ㅤㅤㅤ)/ㅤ.length)([3,1,4,1,5])

Bubbler

We need Whitespace 2 that uses these chars

Seggan

00:24

when u return to ur other computer that had tnb open

Bbrk24

@Seggan would you like some more?

KevinLPG

yeah @Seggan have some more

@Seggan let your cup of pings runneth over

Seggan

@Bbrk24 no, but i would like a pr review :P

Bbrk24

Did you see my latest comment?

Seggan

yeah, fixed

Bbrk24

00:25

oh you just did it. I didn't have Github open. Yeah I'll approve it, one sec

Seggan

ty

now we can finally work within the new module system

KevinLPG

@Bbrk24 you mentioned previously you were going to set a new github status. Did you settle on one? Because this is what I see

Bbrk24

Once we actually design the bytecode I'm excited to implement it in Trilangle. It shouldn't be too difficult actually but it might be a decent amount of code duplication

@KevinLPG . . .

and no, I didn't

@KevinLPG You just lost The Game

KevinLPG

I stopped playing a while ago

@Bbrk24 well hey, you're manually breathing and blinking now

Bbrk24

Joke's on you I already was

Seggan

00:28

@Bbrk24 the bytecode is gonna be quite high level and a bit different than usual

i.e. im planning for an operator/structure that implements the type overloading found in golflangs

Bbrk24

I think I already mentioned this, but I designed the --disassemble feature with the purpose of being able to translate it into bytecode. If I just add a different version of the function that puts out bytecode rather than pseudo-assembly I have a compiler with most of the work already done

Adám

Array.prototype.ㅤ=function(ㅤ){return ㅤ(this.valueOf())}
[3,1,4,1,5].ㅤ(ㅤ=>ㅤ.reduce((ㅤㅤ,ㅤㅤㅤ)=>ㅤㅤ+ㅤㅤㅤ)/ㅤ.length)
2.8

Yesss!

Bbrk24

Why?

Adám

Pure awesomeness.

Seggan

@Bbrk24 hes an APL ~~fan~~ thats an understatement

and that code is much like apl written out

Bbrk24

00:34

Okay but why is ' ' the identifier of choice

Seggan

at least i think thats why

@Bbrk24 because it looks like written out apl

Adám

@Bbrk24 Because then it looks like it isn't there. Subordination of insignificant detail.

Bbrk24

In Swift you get, without invisible identifiers,

extension Array
where Element == Int {
  func average() -> Int {
    return self.reduce(0, +) / self.count
  }
}

Adám

Hold on, are you adding an average method to the Array object?

Bbrk24

Count? Length? Size? I don't know anymore. C# has three different ways to spell it

Unrelated String

00:36

why is the keyword extension if you use it like that

Tsumiki Miniwa

I'm as surprised as you! I didn't think it was possible.

xkcd 391 since it auto-embedded

Unrelated String

it sounds like you're declaring an extension called Array instead of defining an extension for Array

Bbrk24

@Adám Yeah, Swift natively supports extension methods with the extension keyword. It's the same thing you did with Array.prototype in the JS before

Unrelated String

@TsumikiMiniwa the embed also links :P

Bbrk24

@UnrelatedString Rust impl T for S is spelled extension S: T in Swift

Adám

00:37

@Bbrk24 No, I didn't add an average method, I added a method to call any arbitrary function on an array using .() syntax.

Bbrk24

Yeah, but you still added a method

Unrelated String

@Bbrk24 yeah i assumed that's what it does, it's just confusing in a natural-language sense :P

Adám

@Bbrk24 So how would it look in Swift to add a method that allows calling a given function?

Bbrk24

I actually had to do this in real code once, so I have an existing implementation, let me grab it

extension View {
  func apply<T: View>(@ViewBuilder transform: (Self) throws -> T) rethrows -> T {
    return try transform(self)
  }
}

The entire reason this exists is that the @ViewBuilder annotation may be put there

Seggan

~~new pr dropped~~

user

00:48

Public Relations?

KevinLPG

@Seggan holy hell

Seggan

@user yeah

the more we drop, the better public relation we get as the closer it is to a finished product

user

But really, what were you referring to when you said new pr dropped?

Seggan

i opened the pull request to add lazy lists

side effect abuse ftw

user

Made a couple comments

Bbrk24

00:58

I left my comments as well

Seggan

and i addressed 3

Neil

@mousetail SpiderMonkey also uses one-byte strings for ASCII

@Bbrk24 document.title acts as if it forwards to the <title>'s textContent

Bbrk24

Makes sense tbh

Neil

@Seggan did ChatGPT tell you?

Seggan

no

wikipedia

Bubbler

01:19

Ginger

nice

user

It is a message from the gods, you must not spend or pay any more

Bubbler

lol

Bbrk24

@Seggan Which operators? I didn't need it for map or filter

Rydwolf Programs

CMQ: Is there an idiomatic way to append an Option's contents to a Vec if it is Some?

I could use if let, but it feels like there'd be a better way

Seggan

01:31

@Bbrk24 when we are actually implementing the golfing operations

Bubbler

There's iter.filter_map if you're collecting at once, but not quite for pushing one by one

Bbrk24

I love if let

Seggan

@RydwolfPrograms java has flatMap

Bbrk24

Swift has Optional.map but it's rarely used because it can be conflated with Array.map

Unrelated String

yeah you should be able to just use map

@Bbrk24 that's a problem how? :P

Rydwolf Programs

01:33

Wait I think Option has a .iter() right?

Unrelated String

<$> gang

Bbrk24

I'm just saying I've never seen it used in production code because of the confusability of x.map vs x?.map

It's great for golfing

Unrelated String

@RydwolfPrograms yeah that's probably the cleanest solution tbh

Rydwolf Programs

So I could just do [vec, option.iter().collect()].concat() or something

Bubbler

Hmm yeah, Option<T> has IntoIterator

and Vec<T> has Extend

so... vec.extend(option) should work?

Unrelated String

01:36

niiiiice

Rydwolf Programs

Second question: I have a function which recurses through some parsed HTML looking for things to put in a list (I'm making a list of image srcs in the DOM). Would it be better to do it the FP way and have the function return a vec, or take a mut vec as an arg and push to it?

The first feels slightly cleaner from a purity standpoint, but from a simplicity/performance/etc. standpoint the latter seems better

Bbrk24

why not both

have the returning one just be let vec = []; operation(&vec); return vec

Bubbler

@Bubbler yes it works

Rydwolf Programs

@Bbrk24 Because I only need one?

Then you get the worst of both worlds

It's an impure function but also not simple or performant

Unrelated String

yeah would there not be some way to make it, like, emit the stuff

Seggan

01:38

@RydwolfPrograms fp way

Unrelated String

i guess pushing to a given mut vec is how you would do that :P

i don't see why not to do it that way; you can also provide a simple wrapper method that provides an empty mut vec to build on and returns that :P

Bubbler

@RydwolfPrograms The latter is called a visitor and is used in libs like syn

Rydwolf Programs

I'll do the latter, mostly just since this looks kinda clunky:

let mut srcs = Vec::new();

for child in &element.children {
    srcs.extend(search_node(&child));
}

srcs

Unrelated String

xd

Rydwolf Programs

Like if I've got a mut vec anyway I might as well pass it around

@RydwolfPrograms (and yeah I could definitely do this with a one-liner but whatever)

Bubbler

01:42

right

Rydwolf Programs

@RydwolfPrograms Actually I'm pretty sure this is just flatMap lol

Seggan

@RydwolfPrograms return element.children.mapNotNull(::search_node) Kotlin FTW

@RydwolfPrograms yep

11 mins ago, by Seggan

@RydwolfPrograms java has flatMap

Bubbler

@RydwolfPrograms Btw can't you just use crates like scraper

Rydwolf Programs

I'm doing some extra processing and stuff, simplest option is just a regular HTML parser and a 10-line recursive function

Bubbler

"a regular HTML parser" is not exactly simple but ok

Rydwolf Programs

01:47

Well I mean one from a library

html_parser

Bbrk24

@Bubbler Reminds me of stackoverflow.com/a/1732454

Seggan

@Bbrk24 who doesnt know of that post

why havent i upvoted it...

Bbrk24

It's iconic

@Seggan Because it's locked

Seggan

aww its locked

imagine the rep...

Bbrk24

I would have upvoted it by now if I could

Tsumiki Miniwa