Vyxal

emanresu A

02:03

@lyxal What do you think of ^^?

(Could improve efficiency and compiled code readabikity)

lyxal

I don't know if it's worth it in the long run

emanresu A

¯\_(ツ)_/¯ Just an idea

lyxal

How do you determine when to stop grouping?

emanresu A

When you reach an instruction that involves some form of stack manipulation and/or when you reach a structure.

user

02:23

I think we should definitely consider this at some point but it would be hard, so not right now

lyxal

^

emanresu A

Ok :P

I might experiment with it once the rewrite's done.

lyxal

Good idea

Maybe we could add other optimisations too

Like maybe create an autogolfer based on common patterns

emanresu A

That's what I was thinking, & I have another idea:

What if we have a statically compiled elements.py and dynamically import what we need?

lyxal

That would just be printing the elements dictionary after it gets generated

lyxal

02:54

@emanresuA I finally implemented functions as variables btw

It works good

And I was thinking of using ° for complex number creation

So like 3°2 has real part 3 and imaginary part 2

3.5°2.5 -> real = 3.5 imag = 2.5

°<digits> -> real = 0, imag = digits

<digits>° -> real = digits, imag = 1 (shorthand for 0°1)

emanresu A

03:53

@lyxal Thanks!

@lyxal Stealing from Jelly, are we?

lyxal

@emanresuA actually no

if jelly uses that for complex stuff then that's completely coincidential

hyper-neutrino

it doesn't

it uses a character that doesn't exist in vyxal's codepage, ı

emanresu A

Just Jelly has the same sorta syntax for complex stuff

hyper-neutrino

sure, it functions the same way with a different character

but this is like, the only sane way to do it

emanresu A

Also, are we going to have complex rationals now or what?

Just edit it to ^^ :P

lyxal

03:56

@emanresuA well if sympy supports it, then yes

@emanresuA how 'bout i dont

hyper-neutrino

complex rationals are pretty easy

sympy should be able to handle them perfectly fine

lyxal

>>> sympy.S(1.3 + 4.5j, rational=True)
1.3 + 4.5*I

>>> sympy.Rational(1.3 + 4.5j)
Traceback (most recent call last):
...
TypeError: invalid input: 1.3 + 4.5*I

emanresu A

ಠ_ಠ

hyper-neutrino

well, yeah, it's not a rational :p

lyxal

>>> sympy.Rational(1, 4) * 1j
0.25*I

thank you sympy

because you couldn't have left it as a fraction

hyper-neutrino

04:09

>>> sympy.Rational(1, 4) * sympy.I
I/4

git gud /s

lyxal

Mmm I do love how python complex numbers and sympy complex numbers are different

hyper-neutrino

lol

lyxal

Also, fun fact: Sympy has a to_LaTeX function for expressions

Meaning you'll be able to create tex in just 2 bytes lol

emanresu A

That's pretty cool

lyxal

05:02

in The Nineteenth Byte, 2 days ago, by hyper-neutrino

man every time i try using vyxal i suffer from how every useful list builtin either doesn't exist or only works for strings because it's an overload of a math function

@hyper-neutrino would you be okay to create a list of such list built-ins?

Asking so we can implement them

hyper-neutrino

05:18

uh

I can't really come up with a list just off the top of my head but i can make one and just add to it as i think of things

Vyxal Built-in Wishlist

lyxal

05:32

@hyper-neutrino that's fine too

Vyxal Bot

06:30

chunkybanana closed issue #265 (declaring a function is the same thing as assigning a variable, Vyxal/Vyxal)

emanresu A

Since Lyxal implemented it

lyxal

yes

PyGamer0

09:42

@hyper-neutrino $ git gud

@VyxalBot so you can push functions on stack?

emanresu A

You could do that before, but now it's nicer and a byte's been freed up.

lyxal

10:15

You can still push them to the stack

They're just variables now

@PyGamer0 ^

So where you'd once write

!!/run @triple:1|3*; ⟨1|2|3⟩ °triple; M

Vyxal Bot

@lyxal ⟨3|6|9⟩

lyxal

You'd now write @triple:1|3*; ⟨1|2|3⟩ ←triple M

Which actually saves a byte lol

emanresu A

And the syntax is cleaner, and it makes more sense since most functional languages do the same.

(Java ≤ 7.0, I'm looking at you)

lyxal

@emanresuA corr

ect

emanresu A

'night

lyxal

11:12

125 failed, 189 passed in 58.25s

I'll take that for the first run of pytest I've done for a long time

I was expecting much worse

full output (warning: big)

@emanresuA no u

mfer calling ord on lhs not char

lyxal

12:45

hey there @Razetime

I've just been bust fixing little bugs shown by the auto-generated test case suite

because that's a thing we have

and I don't think I'll do any more tonight

9 fixed test cases is enough for one hour's work

!!/'night

Vyxal Bot

@lyxal o/

user

@lyxal It’s painful, isn’t it?

I have time today, so I think I can help with that

Razetime

13:10

@lyxal strange

user

13:22

Strange because it's auto-generated or because we have tests at all? :P

Btw thanks @lyxal for getting me back to the 10k+ club

Once I get enough rep, I'll give 50 rep (only once, though) to anyone who fixes 5 elements

2

RO abuse ftw

emanresu A

21:18

@lyxal You can blame Copilot for that one.

user

21:34

Okay, who's been misusing copilot again?

(this is inside right_bit_shift/<<)

emanresu A

ಠ_ಠ

user

There is so much wrong with that that I can't even

Actually, it's not totally terrible, should be a trivial fix

emanresu A

That was not me

user

nvm, it's in fb anyway

user

22:06

I just realized, every time you want to vectorise a function, vectorise(...) generates a new vectorised version of it :o

This is much inefficient

How about a @vectorised annotation? (I think I added it at one point but didn't do it well or sth)

user

22:29

...Why are there two vy_zips, one in elements.py and one in helpers.py?

user

22:49

The current documentation for summate in elements.yaml says that if the argument is a number, then it's sum(digits of a). But the implementation is basically sum(iterable(lhs)), which makes more sense to me. Update docs or update code?

~~110 failed tests, 215 more to go!~~ 215 passed tests, 110 more to go!

user

23:16

@user nvm that's not how it works

lyxal

@user that's not right?

emanresu A

It's getting a single type, of course it's not right

lyxal

Oh dang

@user it looks like the one in helpers can be deleted

user

Aight, on it

@lyxal There was also an extra wrap in helpers. I renamed it to wrap_with_width for now, but should I just delete it?

lyxal

@user I'll have to look at that in a bit, but I think that one might be needed

user

23:47

We should add a Pytest workflow for the digraph-implementation branch so that it we can check how much progress we've made

We had one for master before, I think

lyxal

Isn't it that workflows only occur on the default branch?

emanresu A

When copilot tries to inject PHP into an html file

lyxal

@user Because I've tried to set up workflows for fb when it wasn't default but it kept going to main

user

Oh, weird

I thought workflows could also be made on other branches

@emanresuA Uh-oh, we now know theyir real agenda: make PHP take over the world :P

lyxal

@user I mean, if you can find a way to get it to work on other branches, I'm all for testing on di

user

23:49

Aight

emanresu A

20

Q: How to run a GitHub Action from a branch other than master?

I have a repo in GitHub and I want to create an Action to build a Docker imagen and push it to the DockerHub. I know how to do it but: if I create the action in a branch other than master GitHub does not run it. This is a know problem (Workflow files only picked up from master?). Any ideas to fix...

lyxal

It's probably just that I'm trash at github workflow creation

user

lol it crashed with "no module named numpy"

lyxal

lol

user

Welp I'm going to push fixing that onto y'all because I need to go do actual work now

Wasted the whole day on my crappy esolang

Transcript for