wowie that's a lot of people

Woah

@user yes

@user some random code i forgot to remove

Opened a PR for that

@lyxal noice

@user why

@Seggan oh ok I’ll just remove it then

i removed it already, its in the main branch

@Seggan something something micro optimization

@user mondas or dyads?

@Seggan oh didn’t know that

@Seggan dyads

hmm idk much about annotation processors

Also monads ig

monads stay the way they are rn

I’ve always wanted to try them out

btw i gtg bed, we can talk more tomorrow

!!/'night

@Seggan o/

0/

@lyxal :|

cmnowastaken forked Vyxal/Vyxal into cmnowastaken/Vyxal

cmnowastaken opened PR #1050 (cmnowastaken/Vyxal → Vyxal/Vyxal) (cmnowastaken:main → Vyxal:main): added ඞ unicode

chunkybanana approved on PR #1050 (cmnowastaken/Vyxal → Vyxal/Vyxal)

We absolutely needed ^^, it's hilarious

Lyxal approved on PR #1050 (cmnowastaken/Vyxal → Vyxal/Vyxal): "Welp guess this is part of vyxal now"

Lyxal merged PR #1050 (cmnowastaken/Vyxal → Vyxal/Vyxal) (cmnowastaken:main → Vyxal:main): added ඞ unicode

I may have helped them with that and tbh I think we should have better guides

Y'all ever just wonder if SBCS will be able to viably compete with fractional byte languages once they start evolving towards competitiveness?

Because I sure hope so

As with all the latest stuff happening, it makes me wonder if v3 will even be relevant when it's released

i feel like fractional byte languages don't have that much of an edge and sbcs will always be superior aesthetically

Aesthetically yes, but someone's gonna win the fractional arms race eventually

And that winner is obviously going to be competitive

@lyxal what is the latest stuff?

@PyGamer0 radvylf's, pxeger's, EmanresuA's, dlosc's and the creator of golfscript's ventures into non-integer byte count territory

i see

The current fractional byte ideas are a bit unusable and don't end up much better than SBCS in practice, but if there's anything minecraft speedruning and speed cubing have shown, the meta/tools used evolve over time where what was previously cool and good becomes warm and bad

So a better question is how do you keep SBCS competitive in a fractional age?

lmao minecraft speedrunning and speedcubing lol

what

Essentially, how do we take vyxal from the 3rd generation to the 4th?

That's the important question

What if bi-directional modifiers?

That is, if there's nothing to the left take from the right

@UnrelatedString what are these unparseable nilads you speak of

Any program ending in a nilad is redundant

@lyxal Now I want to write another blog post about that

@emanresuA then make trailing nilads into monads or dyads

That could be an interesting experiment

Hmm but then that runs into the issue of functionality being locked away from the rest of the program

What if nilads at the end were yeeted to the front of the program

+ 3 4 would become 3 4+ before execution

That's two redundancies solved in just one hour

More. I need to know more!

More redundancies must be solved!

@UnrelatedString minecraft speedruns used to be 20 minutes, now they're sub-10. The strats used originally are no longer relevant. Speed cubes have had many mechanism changes and lubing metas meaning that a cube from 2010 is very cringe compared to a 2022 cube.

I sometimes wonder if vyxal can keep up with new innovations while still being its own language

Maybe we delete 2-5 characters from the code page and use EmanresuA's idea

@lyxal ono

That's not to say it's being abandoned

V3 is full steam ahead

It's just that some clever innovations will be needed to keep it up to par with fractional

Gotta keep vyxal future proof :p

And this isn't some knee-jerk reaction to pxeger's language announcement - this is something I've been considering for a while

@lyxal the only concern I have with this though is that the ease of just typing a character and it equalling a single command is lost with fractional languages

^

I suspect most future golflangs are gonna be syntax sugared though

So how does one keep it simple to type and use but still fractional?

Pretty much a compiler

Say, for catstruct you might end up typing a program like this:

(+ (sqrt input) 2)

And that'd get compiled to a bytestring

What if the normal SBCS version gets compiled to a bytestring?

now that's an idea

that's a genius idea

I'm not sure I understand what you mean

Ohhh waiit

so instead of (+ (sqrt input) 2) you'd have √?+2 and that'd get compiled into the byte-string

Do you mean that there's a sbcs-like sublanguage which compiles to a complicated bytestring?

exactly

@lyxal Ahh yeah I see

it retains the simplicity and readability of SBCS with the benefits of fractional byte counts

hell yeah

4th generation here we come!

I just realised how easy it is to DoS the online interpreter

Oh well :P

lol

@lyxal imagine this: fractional byte counts suck and someone proves that SBCS is better than fractional byte stuff :P

there are now 258 items in the core functions list for version 3. Come forth and leave feedback

!!/hyperping

@lyxal @lyxal @Steffan @mathcat @emanresuA @user @UnrelatedString @hyper-neutrino @Seggan @PyGamer0 ^

@lyxal but isn't the max 256?

some of the things there will be overloads

meaning that the 258 probably won't be enough

meaning more stuff needs to be added

plus some of those things are numerous things in one

it's meant to be a list of roughly 250-300 items that absolutely 100% need to be implemented

woah

my original plans for v3 called for a total of 532 things

but those may not be the most useful

hence the new list

what's dyadic maximum/minimum?

@lyxal how would they ever get an edge? they might have smaller byte sizes but that translates to less commands which translates to longer code for non-trivial tasks

@mathcat ∴ and ∵

@Seggan no doubt someone will find a way one day to create a reasonable balance

Nibble by the creator of golfscript comes closeish

@lyxal theyre only sub-10 because piglins

:61249908 got it

@Seggan well who knows what the piglins of golfing languages are

perhaps it just hasn't been discovered yet

but in other news:

Curo will beat everyone

they somehow captured the entirety of vyxal in 15 seconds

@mathcat after I've started/finished coding it

@lyxal aww

> atop

> over

how does that fit into vyxal?

173 is yoinked from jelly lol

@user yeah so rn every element has a boolean whether it vectorises. monads are automatically vectorised by the compiler, dyads are still manually vectorised. im thinking of somehow vectorising those dyads using a vectorise method and MethodHandles

annotations are too manual imo

@lyxal looks good

@PyGamer0 takes two arguments

is there a swap2 element, kinda like jvms DUP_2, POP?

where DUP_2 inserts a copy 2 below

@Seggan yep, that’s what my pr is for. Could you look it over and see if it’ll work?

@Seggan there is a 2dup, but no 2swap

maybe 2dup, pop will be enough

wait is swapd ( x y z -- y x z ) what you mean?

no more like x y z -> z x y

@lyxal shallow flat vs. deep flat?

@lyxal ooh vyxal+forth combo lang when?

@Seggan so basically ∇ right now?

wait we have that?

we do

a, b, c -> c, a, b

argh

i didnt notice it

@mathcat deep = fully flat, shallow = flatten by one layer

ah

Would be nixe to have a digraph to flatten by n layers

Or like that apl thing

you mean Þf?

How many weapons do you actually have?

yes

@user i cant review rn, but you should make the signatures Any b, Any a instead of Any a, Any b

as stuff is popped backwards

nice, type-checking elements!

not exactly type checking

also user your dyadvectorise wont work on infinite lists

Ah right

I’ll make the changes later today, thanks

@Seggan wdym?

oh, and one more obstacle to using kotlin multiplatform: MethodHandles dont exist in native

@mathcat were you replying to my message about signatures?

how are the elements not type checking?

oh i get it

no, I was mentioning the "is list?", "is string?", etc. elements for v3

ah

@Seggan That's not a problem, you just have two different bits of code for JVM and native

yea but idk how yad do that

Any idea why it'd say runtime.list isn't generated @Seggan?

probably because i moved it someplace else

yeah the private const val runtimeClasses = "/build/runtime-classes" line needs to be private const val runtimeClasses = "../runtime/build/runtimeLibs"

sadly i cant change it rn

Ah ok

only gonna be able to do it in the evening

I can change it

We need to make save in register peek in v3

I use duplicate + save in register wayyyyy more than just save in register

Seggan requested changes on PR #4 (Vyxal/Myxal)

Seggan approved on PR #3 (Vyxal/Myxal)

Seggan merged PR #3 (Vyxal/Myxal) (Vyxal:test-workflow → Vyxal:main): Add workflow to run tests

cgccuser deleted branch Myxal/test-workflow

Seggan requested changes on PR #2 (Vyxal/Myxal)

@user i made an actual dedicated test harness instead of a flag. we can actually check the stack after execution now

Ooh nice

chunkybanana opened PR #1051 (Vyxal/Vyxal) (Vyxal:main → Vyxal:production): Update production (31/5/22)

IMO we shouldn't need approvals on prod updates

cgccuser approved on PR #1051 (Vyxal/Vyxal)

cgccuser merged PR #1051 (Vyxal/Vyxal) (Vyxal:main → Vyxal:production): Update production (31/5/22)

It's still 30/5/22 for me lol

CMQ: Do we need approvals on prod PRs?

!!/hyperping

@emanresuA @lyxal @Steffan @mathcat @emanresuA @user @UnrelatedString @hyper-neutrino @Seggan @PyGamer0 ^

I think we do

1 approval is pretty easyto get

More would , however, be unreasonable

@emanresuA yes, to make sure other branches don't get merged into production

Ah frick

I guess yeah

That's why admins can override it on production

Because it's a sanity check/safety measure