Vyxal

user

2:56 AM

It shouldn't: scastie.scala-lang.org/WYc2YJq4QpaThtpB9MYV4A

Filter lambda has same problem

lyxal

3:23 AM

@user but the method on Seqs

does that force evaluation?

user

It shouldn't, this is where it comes from: github.com/scala/scala3/blob/3.5.0/scala2-library-bootstrapped/…

Views are lazy, so it should be lazy

lyxal

well the only thing which would cause mapping over an inf list to hang is to.zipWithIndex.map

which extends Seq

VList extends Seq[VAny],
      SeqOps[VAny, Seq, VList]

def zipWithIndex: Iterator[(A, Int)]^{this} = new AbstractIterator[(A, Int)] {
    var idx = 0
    override def knownSize = self.knownSize
    def hasNext = self.hasNext
    def next() = {
      val ret = (self.next(), idx)
      idx += 1
      ret
    }
  }

that's how it's implemented

user

Wait this is how we implemented it?

With the capability syntax?

lyxal

no that's how it's implemented in Scala

user

I thought that was experimental

Oh ok

lyxal

3:35 AM

deep down in Scala code

user

Okay well that should be lazy too, iterators are supposed to be lazy

lyxal

is map on a seq lazy?

user

This one's mapping an iterator, right? So it should be lazy

Could you make a unit test that creates an infinite VList and zipWithIndexes it and maps it?

Don't have the Vyxal repo on my machine at the moment

I can look at it tomorrow if you like (just ping me)

lyxal

I'm good to make the test now

one tick

user

poggers

lyxal

3:43 AM

test pushed

ÞPƛ}5Θ should print [2, 3, 5, 7, 11]

but online it hangs

the test will be whether the test passes for jvm

user

Aaaaaaaaa it's hanging

lyxal

okay well at least it's hanging on jvm tests too

user

Is it just prime numbers?

lyxal

yeah

ÞP5Θ doesn't hang

user

Maybe we're just not constructing the prime numbers properly?

lyxal

3:46 AM

the map shouldn't impact anything

e.g. ÞPƛ}4 shouldn't hang

user

Nope never mind

Yeah it hangs with naturals too: vyxal.github.io/…

lyxal

you can tell it's forcing evaluation because that should just print 4

it's funny, if this were any other golflang, this wouldn't be an issue (well, maybe apart from 05AB1E and Husk)

because there'd wouldn't be anything infinite to break :p

user

Yeahhh

Could you do something like this?

val x: LazyList[VNum] = VNum(1) #:: x
val v = VList(x)
val useless = x.zipWithIndex

That'd be the whole test. Just to see if zipWithIndex does something dumb

lyxal

where in?

user

Uhh literally anywhere

And if that doesn't hang, stick a .map after the .zipWithIndex

lyxal

3:55 AM

gotta wait for node to finish installing something first

user

:thumbsup:

If we absolutely can't figure this out, we can write our own .zipWithIndex override that that just calls lst.zipWithIndex

lyxal

it's using most my computer power Node

user

Love modern tech

lyxal

running jvm.runLocal

@user mfw this'll take updating all instances

user

Wait no it shouldn't require that

It'd be called .zipWithIndex too

lyxal

3:57 AM

I tried something like that but it had the wrong return type

@user you're gonna laugh

it'd errored because x isn't VAny

user

bruh

VList.from, I forgot

lyxal

recompiling

> x is a forward reference extending over the definition of x

user

aaaaaaaaaaaaaaaaaaaaaaaaaaaaa lazy val

Why can't the computer just read our minds and do the right thing

lyxal

ikr

dind't hang

user

Hmm

So it's not .zipWithIndex by itself

lyxal

4:03 AM

added map

compiling

 x.zipWithIndex.map { case (v, i) => v + i }

didn't hang there either

changing it to v.zipWithIndex.map

that errored because of type mismatch

so I'll try v.zipWithIndex because that's what I really wanted to test

user

Wait what's the type mismatch with v.zipWithIndex?

lyxal

@user v + i

user

Oh ok

lyxal

it was because I didn't update the map function

but anyhow, v.zipWithIndex hangs

user

Hmm

lyxal

4:06 AM

meaning it's forcing evaluation

user

@user Let's just do this tbh

Should be fairly easy, just something like

def zipWithIndex: Seq[(VAny, Int)] = lst.zipWithIndex

lyxal

compiling

had to make it override

user

Ah that makes sense

lyxal

also, Int?

that'd be better as a VNum

user

Can't override then

lyxal

4:09 AM

I know

I mean to say custom might be the solution

even if zipWithIndex works

user

lst.zipWithIndex itself will never give you anything beyond Int.MaxValue, so I don't think it's necessary to use VNum

Like a whole new implementation?

@lyxal I guess you could copy this then

lyxal

well it worked

didn't hang

user

Even if you make a custom zipWithIndexVNum, I'd prefer it if there was also a lazy override of zipWithIndex

\o/

lyxal

pushed

I suspect somewhere in the OOP structure something was casting to finite Seq

user

Maybe

lyxal

4:17 AM

Pushed the zipWithBigdex function

user

bigdex

lyxal

Vyxal 3.4.8

lyxal

4:55 AM

I've had a brilliant idea

two more context variables

first is ζ

inside a mapping/filter/reduce/whatever lambda, it'll be the thing that was popped as the argument

tbd what in normal lambda execution

second is γ

now this one is a toss up between two options

a) top of outer-most stack (definitely useful)

b) last calculated value

the idea would be something like this

<value> <operation on value> <structureLambda> <reference to original value before lambda>

for the pattern where you do something to a value before lambdaing

not sure if that part is 100% useful

(hence the outer stack value)

I've also added ⬱ to index from the outer stack

the idea behind these ideas is that lambdas and values outside the lambdas seem to be a bottleneck

so some way of remedying the situation seems like a good and new idea

lyxal

5:18 AM

Oh no I've had a horrible idea.

Sneak list operations into string overloads by making the string overload eval first

Given that string overloads get harder to make generally useful the more you add

Like once you've gotten the basic stuff, string overloads become somewhat arbitrary

Dunno if the idea has any merit, just something as a thought

RubenVerg

5:38 AM

@lyxal I added a glyph before you even did!

lyxal

5:59 AM

The two I removed balances out :p

emanresu A

5

A: Smallest prime q such that concatenation (p+q)"q is a prime

Vyxal g, 50 bitsv2, 6.25 bytes Þp'⁰+nJæ Þp'⁰+nJæ⁡‎‎⁪⁡⁪⁠⁪⁡⁪‏⁠‎⁪⁡⁪⁠⁪⁢⁪‏‏⁡⁠⁡‌⁢‎⁪⁪⁠⁪⁪⁠‎⁪⁡⁪⁠⁪⁣⁪‏‏⁡⁠⁡‌⁣‎‎⁪⁡⁪⁠⁪⁤⁪‏⁠‎⁪⁡⁪⁠⁪⁢⁡⁪‏‏⁡⁠⁡‌⁤‎‎⁪⁡⁪⁠⁪⁢⁢⁪‏⁠‎⁪⁡⁪⁠⁪⁢⁣⁪‏‏⁡⁠⁡‌⁢⁡‎‎⁪⁡⁪⁠⁪⁢⁤⁪‏‏⁡⁠⁡‌ Þp # ‎⁡a list of infinite primes ' # ‎⁢filtered by ⁰+ # ‎⁣(p+q) nJ # ‎⁤(p+q)"q ...

TIL that g is smart enough to know that the first item is the minimum

RubenVerg

6:48 AM

oh, one less constant too?

good for me, ⑩ looked not great in the font anyways

lyxal

7:17 AM

@emanresuA I don't know if it's smart or just lucky :p

Going through the implementation of monadic minimum shows no checking of the is infinite attribute of LazyLists

Ohhhh

That's sneaky

@emanresuA it isn't smart at all

Go to the link and look in the flag box

emanresu A

Ah

lyxal

Yeah :p

The template generator doesn't put the limit flag in the answer header because, well, I never actually thought anyone would use it for golf

Turns out it's more useful than just debugging :p

Clever usage of the flag though

Ginger

12:43 PM

@lyxal what was the thing you wanted me to do

lyxal

1:10 PM

I kinda forgot

@Ginger probably make it so that clicking on a codepage item inserts the character

Ginger

1:25 PM

@lyxal I already did that :p

oh right it was turning off the eof marker

lyxal

@Ginger not on the codepage tab you didn't

Ginger

oh right

Ginger

1:43 PM

@lyxal done

lyxal

nice

Ginger

3:23 PM

@lyxal bracket highlighting and the EOT marker are now settings

Ginger

4:07 PM

also, snow can now be turned on year-round by typing snow while the settings modal is open :3

Ginger

4:38 PM

mfw there's a race condition in the utility worker

and I predicted that it would happen a week ago but assumed it was okay

Ginger

4:52 PM

@lyxal the syntax info and grammar file seem to disagree? specifically about how the ngraphs starting with #: are spelled

the syntax info has #:R, #:@, #::, etc. but the grammar spells those with an extra colon

and is missing several of them

also I made λ a structure opener in the grammar because it wasn't for some reason

Ginger

5:43 PM

first Vyxal answer I've ever posted :3

Ginger

5:56 PM

wait no I goofed

Ginger

6:19 PM

it bothers me immensely that the only way to get a generator of every positive integer is ÞṬȦu

would also be cool if there was a "nth element satisfying a predicate" thing

also I think “ and N do the same thing when given a function (first positive integer where predicate is true)?

Ginger

6:47 PM

...is there really no "first item of list where predicate truthy"? there's like three ways to get the index of the item, but not the item itself

RubenVerg

Fh?

Ginger

that's two bytes though

RubenVerg

or /h if you prefer

@Ginger and there's three 1-bytes for the index?

Ginger

*two, ᶤḋ

one's a modifier and one isn't

RubenVerg

"find-by:" is a bad name for ᶤ, it implies it returns the item and not the index

though i agree it should at least be a function and maybe a modifier too

Ginger

6:52 PM

my use-case is Ω<stuff>}h, which could be ᶤϩ<stuff> and save me a byte if ᶤ returned the item and not the index

and there are probably cases where arity grouping would eliminate the ϩ too, although that doesn't work here

also I think drop should take its arguments in num, list order, not the other way around

RubenVerg

no because "drop 3" is λ3İ}

presumably there's a shorter way to write it but anyway you'd need to like dip

or swap

Ginger

oh yeah

ALSO ¬ should be a modifier

RubenVerg

i mean we just need a composition modifier

Ginger

a whar

RubenVerg

...that's just ϩ isn't it

Ginger

7:06 PM

yeah :p

something like Ω"frob"c¬}h could be /¬"frob"ch if ¬ were a modifier

makes arity grouping more effective, is my point

RubenVerg

7:20 PM

yeah but then not becomes ¬id or something?

i guess just have both

Ginger

bothisgood.gif

RubenVerg

8:51 PM

digraph for reduce from the right?

lyxal

10:58 PM

@Ginger is there not an all natural numbers built in?

@Ginger I look into this

Ginger

@lyxal I may be stupid

possibly

:p

lyxal

@Ginger it's hard to tell given that I know I missed a few important things in v3 :p

Ginger

(there is)

Transcript for