Vyxal

emanresu A

00:16

I have a 7 that only works on a new version of Vyxal

Due to removing debug prints

So 3 7s

Would you mind asking them to update?

lyxal

I already have

I asked when I released 2.15.2

emanresu A

Can you ask again?

lyxal

Done

Steffan

Ok mucho problemo

opens weekgolf

lyxal

@emanresuA later today is when it'll be done

Steffan

00:22

Got 7. It was the easiest thing on earth

lyxal

Ikr

Steffan

I just opened up the tab and did it instantly. I don't know why I didn't do it earlier

Lynn and Lyndon/Lydxn are super good at golfing tho

In literally any language

Is Lyndon/Lydxn even on CGCC?

I actually have two 7-byters. One would also be 7-bytes if the input was given normally, the other wouldn't.

emanresu A

@Steffan I suspect I've got a quite different approach to y'all

Mine starts with ¶

And it took fricking half an hour to find.

lyxal

Wait what

Tf you doing with that ¶?

emanresu A

You'll see at the end of the week

lyxal

00:35

Well not if you don't submit it to the site

emanresu A

I will... when Vyxal gets updated

lyxal

Imagine needing 2.15.2

Made by works in 2.14.2 gang

emanresu A

20 mins ago, by emanresu A

Due to removing debug prints

If you look at the difference between v2.14.2 and v2.15.2 in terms of print, you'll quite easily find it

Steffan

@lyxal One of mine starts with ¶

The other doesn't

emanresu A

@Steffan I think we found the same thing

And I also found the one y'all use

@Steffan Are the last four bytes of your two the same?

Steffan

00:44

Yes

But mine works in 2.14.2

emanresu A

Huh. I had ẇ÷, what did you have?

Steffan

wait no, it doesn't.

emanresu A

What were your first three bytes for the ¶ one?

Steffan

the debug message you're talking about was never removed, though. It's still there, and updating to 2.15.2 won't help it.

My other doesn't work. I thoguht it would but it didn't.

emanresu A

@Steffan Huh. Ctrl-Fing for print didn't yield anything

Wait nvm lol

The irony is, it took me two minutes to find the one y'all were using

lyxal

00:51

Oh how I wish ȯ was 1 indexed right now

If it was, a 6er would be possible

emanresu A

Ooh, what?

lyxal

@emanresuA ₄ȯ

But alas, 26 includes the first newline because 0 indexing

emanresu A

That ... still should work tho

lyxal

You'd need 27, but that's 2 bytes taking it back to 7 bytes defeating the point

@emanresuA not if you truly have the 7 I have

I tried it

Leading newline gets included

emanresu A

Wait nvm

lyxal

00:59

Does any of y'all's 7s start with kA?

emanresu A

Nope

lyxal

Then you don't have my 7 byter

emanresu A

Found it

lyxal

Amazed at how stupid it is?

Like why bother struggling with the input format when you can just take advantage of it

emanresu A

Yep lol

Steffan

01:10

Mine doesn't start with kA lol

But it should be super easy to find

emanresu A

Neither did my two

Steffan

Just different stack oper

emanresu A

I seriously doubt there's a 6

Steffan

Yeah

lyxal

@Steffan wow so I had something completely different to both of you

emanresu A

01:10

Random idea: Use a sort to get kA

Steffan

Ooh

lyxal

problem: newlines

and duplicate letters

source: I tried that

emanresu A

lol

Steffan

shall I reveal what I have lol

emanresu A

It's the same as what I have

Steffan

01:14

lyxal doesn't know though

I have another 7 with sort

lyxal

@Steffan something to do with <newline const><w updot>?

Steffan

yeah

no

lyxal

well we all have 7 anyway, so it should be fine

emanresu A

@Steffan Huh nice

What?

Steffan

↵ḣ⁋Ȯs„Ŀ

lyxal

01:18

nice

Steffan

The other is kA instead of Ȯs

@lyxal ¶ẇ doesn't work due to debugging messages

lyxal

imagine not starting with kA$Ŀ

emanresu A

Works online due to print not being redirected

Steffan

yes

I was 100% sure you had the same thing as me lol

I didn't even think of using kA$ (or kA? or kA⁰ or kA^ or kA‟ or kA„)

lyxal

2 hours ago, by lyxal

Well actually it's the input format that got me from 8 to 7

I thought "well it already has newlines, just use those lol"

and cheesed it

last three bytes can include ↵Ḣ⁋ and 27ȯ

Steffan

01:23

Wait what

Without even using transliterate?

lyxal

3 mins ago, by lyxal

imagine not starting with kA$Ŀ

43 secs ago, by lyxal

last three bytes can include ↵Ḣ⁋ and 27ȯ

Steffan

OH WHAT ON EARTH

ASDASDASD

lyxal

is that a reaction to how stupid the solution is?

Steffan

yes

lyxal

hehe

Steffan

01:25

I would've never thought of doing that

It's so weird

emanresu A

I count 40 different 7-byters

Steffan

lol

Because of the stupid amount of stack operationsd

lyxal

@Steffan that's the beauty of transliterate not acting on characters not in the translation alphabet :p

emanresu A

@Steffan Yep. kA, 6 stack ops, and one of ↵Ḣ⁋, 27ȯ, ₄›ȯ,↵ḣ⁋, ¶ẇt, ¶ẇṫ for 36, plus (↵ḣ⁋ or ¶ẇ÷), (kA or Ȯs) and „Ŀ

Or as a regex, kA[$?⁰^‟„](↵[ḣḢ]⁋|(₄›|27)ȯ|¶ẇ[tṫ])|(↵ḣ⁋|¶ẇ÷)(kA|Ȯs)„Ŀ

I thought I might have a cheaty 6 but no

My idea was transliterate without popping, then remove the key. but nah

There's possibly some hackery you can do with øA

lyxal

01:52

anyhow, completely unrelated, did y'all see the codepage remodelling I did?

and what do you think of it?

emanresu A

Pretty good

PyGamer0

04:29

@lyxal no wher?

lyxal

14 hours ago, by lyxal

  0123456789ABCDEF
 ┌────────────────
0│ᵃᵇᶜᵈᵉᶠᶢᴴᶤᶨᵏ¶ᶪᵐⁿᵒ
1│ᵖᴿᶳᵗᵘᵛᵂᵡᵞᶻᶴ′″‴⁗ᴥ
2│ !"#$%&'()*+,-./
3│0123456789:;<=>?
4│@ABCDEFGHIJKLMNO
5│PQRSTUVWXYZ[\]^_
6│`abcdefghijklmno
7│pqrstuvwxyz{|}~¦
8│ȦḂĊḊĖḞĠḢİĿṀṄȮṖṘṠ
9│ṪẆẊẎŻȧḃċḋėḟġḣŀṁṅ
0│ȯṗṙṡṫẇẋẏżẠḄḌẸỊḲḶ
A│ṆỌṚṢṬạḅḍẹịḳḷṇọṛṣ
B│ṭ≤≥≠₌⁺⁻⁾√∑«»⌐∴∵⊻
C│₀₁₂₃₄₅₆₇₈₉λƛΩΛµ∆
D│øÞ½ʑʒʓ×÷£¥
E│
F│⁺∥∦

PyGamer0

@lyxal not a lot of greek letters

and are
the ones at the top are for modifiers?

lyxal

They are.

I decided to move away from too much greek

PyGamer0

ok lol

lyxal

You know, JoKing could be on to something

things like $?⁰^‟„ could all do different things based on a) how many inputs there are and b) how many items are on the stack

so ^ with an empty stack could push all inputs reversed

and ^ with 1/2 items could do something different too

because ^ with 1 stack item is a NOP, and ^ with 2 stack items is $

same with ‟„

$ and ? wouldn't have changed behaviour though

they'd remain the same regardless of situational context

(JoKing's message was in regards to the 40 different ways to solve this week's week.golf problem)

but it's only stack manipulation stuff that could be changed

all the other variations are because there's several ways to accomplish the same thing but only because of the nature of the challenge and input format

emanresu A

05:40

If we ever run out of digraph space, we can get more by making Þ∞ one byte again and passing it into stuff that can't support infinite lists

lyxal

13:57

dddd
₿₿₿₿

huh

that's almost aligned right on mobile

  0123456789ABCDEF
 ┌────────────────
0│ᵃᵇᶜᵈᵉᶠᶢᴴᶤᶨᵏ¶ᶪᵐⁿᵒ
1│ᵖᴿᶳᵗᵘᵛᵂᵡᵞᶻᶴ′″‴⁗ᵜ
2│ !"#$%&'()*+,-./
3│0123456789:;<=>?
4│@ABCDEFGHIJKLMNO
5│PQRSTUVWXYZ[\]^_
6│`abcdefghijklmno
7│pqrstuvwxyz{|}~¦
8│ȦḂĊḊĖḞĠḢİĿṀṄȮṖṘṠ
9│ṪẆẊẎŻȧḃċḋėḟġḣŀṁṅ
0│ȯṗṙṡṫẇẋẏżẠḄḌẸḤỊḲ
A│ḶṂṆỌṚṢṬẒạḅḍẹḥịḳḷ
B│ṃṇọṛṣṭẓ≤≥≠₌⁺⁻⁾√∑
C│«»⌐∴∵⊻₀₁₂₃₄₅₆₇₈₉
D│λƛΩ₳µ∆øÞ½ʑʒʓ×÷£¥
E│←↑→↓±‡†Π¬∧∨⁰¹²³¤
F│⁺∥∦ıΩ₵ðß§¶ᶿᶲ•≈∫ꜝ

More codepage remodelling

lyxal

14:20

opinions?

user

Me no like the ′″‴⁗ section

Also, could ᶴ be replaced? It is distinguishable from ∫ but still too close for my liking

lyxal

it's supposed to be mnemonic for last n elements as a function

@user give me another single-width unicode character to replace the big version, and I will

user

Ah makes sense

The last two are just hard to distinguish

lyxal

@user only when they're next to each other

user

@lyxal What do the two stand for now?

@lyxal No, it's easier to distinguish when they're together, but when they're apart, you need to squint and count to make sure

lyxal

14:27

@user the small one is Reduce each overlap of size 2 by an element

the big one is unassigned

only the first two rows have anything assigned

Unrelated String

ʅ

user

lol

No a huge fan of ʒʓ looking alike but I guess they're distinguishable?

Unrelated String

i have no idea if it monospaces correctly but i can't let the ten minutes i spent hunting down sinological notation someone butchered go to waste

lyxal

@UnrelatedString seems to

user

New element: ʅ (Autodiff)

Unrelated String

14:30

big

user

Do the full size versions of ᶿᶲ not monospace correctly?

lyxal

idk if they do

I wanted the superscript versions

user

Because those would be easier to read

Oh ok

What are they for?

lyxal

modifiers

superscript = modifier

user

Ah that's a good idea

Unrelated String

14:30

in The Nineteenth Byte, Feb 8 at 5:36, by Unrelated String

it has both a graphical analogy and a verbal mnemonic since both of the ranges that contain 0 are zs

lyxal

except for a few cases

user

@UnrelatedString Ah ok

@lyxal What are those?

lyxal

like ], ), ⁺∥∦, /, ` \ `

user

But all superscripts are modifiers, right?

lyxal

yes

user

14:32

Great

It's fine if not all modifiers are superscripts ig if there's only a few of them and they're recognizable

lyxal

superscript → modifier

that is, superscript implies modifier

user

This codepage lgtm for the most part but I'm not an expert

@lyxal Damn we getting fancy with logic notation here

lyxal

you can thank the discrete mathematics course I'm taking for that

Unrelated String

looks good to me too

user

I've wanted to ask, why does it have to be so clandestine? Is the math in that course so forbidden that you can't do it out in the open? :P

Wait ¶

lyxal

14:34

@user it's too powerful for normal humans :p

@user 0A is literal newline

user

There's two of them

lyxal

yes

0A is supposed to be a literal newline

user

Oh the bottom one isn't a newline

lyxal

correct

it's a literal ¶

just in case y'all are wondering why the codepage re-work, it's because I felt the first draft was a little disjoint in some places

Steffan

16:34

Our current codepage has complex separator and first input looking the same

°⁰

In the Vyxal editor, they look the same.

In SE chat, they look just a tiny bit different

One is superscript O, the other is 0

Technically, we should remove O from the codepage because it looks too similar to 0 :P

!!/run °

Vyxal Bot

@Steffan I

Vyxal Bot

17:33

Steffan153 opened issue #1379 in Vyxal/Vyxal: Complex number with only multiplying and no adding errors

emanresu A

21:25

@lyxal Looks good! We may want to swap out some stuff...

Also, can we keep (useful) overloads the same where possible?

Deadcode

22:13

ṄL to calculate partitions in Vyxal is exponentially slower than my regex ^(\1x*|^x+)*$...

Maybe it can use a faster algorithm when just taking the length?

But even if it's storing the partitions in memory, it seems odd that it'd be that much slower than a regex

@Steffan Would you mind if I edited your posts to add more powerful test harnesses?

For example Try it Online!

(keeping the plain harness there too, of course.)

emanresu A

22:43

@Deadcode Pythonanywhere is slow

Steffan

23:58

@Deadcode Sure

Transcript for