Multi-branch lambda conversation time (again) :p

Here's a list of what each lambda type does with more than one branch

map: refocus the context variable each time
filter: keep where true under all branches
sort: tbd
reduce: tbd
normal lamda: nothing

No clue what sort and reduce should do.

I know there were some suggestions, but nothing concrete was decided upon

Didn't you already decide that it would also sort by those other branches?

That made sense to me

If two elements are equal when they go through the first branch, try the second branch. If they're equal there too, try the third, ...

it made sense at first, but then when I went to implement it, I realised that might be too complex

unless it's just "select the smallest each time, if the smallest is more than 1 item, filter until either 1 item or end of branches"

that could work

Wouldn't sort(branch1, branch2, ...) be the same as sort(x => [branch1(x), branch2(x), ...])?

Oh I see actually implementing is hard

How about if there's more than one branch, you group by the result of the first branch, then inside that, group by the result of the second, and so on until each group only has one element, and then sort?

@user good idea in theory, gets verbose real quick in practice

+ Maps are confusing to work with

@lyxal Verbose is fine, isn't it?

Because I also tried that

@user not when it starts becoming confusingly verbose

I ended up with things nested 5 levels in parenthesis

Because you have to wrap each branch in a lambda and then call that in a call to VFun.execute and compare those in a call to MiscHelpers.compareExact which is already in a call to group by and it's not just a flat list of VAny, it's a list of 2tuples because ctx-m exists so then there's indexing into each 2tuple

And then you mix recursion into all of that and it just becomes unpleasant

+ it's really easy to get lost in the types being thrown around

You've got a Seq in one place, some sort of iterable thing in another, VList required and a Map with VAny as keys

Perhaps we could use sortWith?

I feel like it'd be pretty straightforward to implement using sortWith (you provide your own comparator)

@user this was using sortwith

Or was it sorted?

I don't remember which method I tried using

list.sortWith { (a, b) =>
  val (aRes, bRes) = branches.map { f => (f(a), f(b)) }.dropWhile(_ == _).head
  aRes > bRes
}

Something like ^ maybe?

@lyxal I changed vyxal-bot to vyxal-bot-2 q:

Fair enough

@user just gotta figure out how to incorporate zipwithindex into this :p

Now that there's literate mode, I don't think I'll use normal form very much anymore :p

Except for golf

it automatically generates normal form

so I don't even have to use it then

!!/dQw4w9WgXcQ

[Dqw4W9Wgxcq]!? STEFFAN!? WHAT ARE YOU TALKING ABOUT!?

We gotta see the source code

@VyxalBot WRONG

!!/wrong

[Wrong]!? STEFFAN!? WHAT ARE YOU TALKING ABOUT!?

@VyxalBot WRONG

!!/run 2 2+

[Run]!? STEFFAN!? WHAT ARE YOU TALKING ABOUT!?

!!/status

[Status]!? STEFFAN!? WHAT ARE YOU TALKING ABOUT!?

DEE BOT NO WORK

!!/die

[Permissions] NO

IT'S TIME TO BE A [Big Shot]

!!/amiprivileged

[Amiprivileged]!? STEFFAN!? WHAT ARE YOU TALKING ABOUT!?

Uh why did the bot just restart

Wait did it shut down

!!/die

[Permissions] NO

SEE YOU KID!

IT'S TIME TO BE A [Big Shot]

It literally tells me no and then shuts down lol

!!/die

SEE YOU KID!

IT'S TIME TO BE A [Big Shot]

!!/die

[Permissions] NO

lol

!!/die

it's gone

ah

I think I crashed it

!!/mom

@Steffan :+1:

@Ginger I BROKE DEE BOT

*we

IT'S TIME TO BE A [Big Shot]

I think you'll find I played a non-trivial part in crashing it :p

oh it's back

!!/!!/

[!!/]!? LYXAL!? WHAT ARE YOU TALKING ABOUT!?

try not to break the bot challenge (impossible)

@Steffan er

oh no

eh, it's restarting fine

I'll fix it tomorrow :p

@Steffan here ya go

@lyxal Do the lambdas need indices? I guess the first line could be list.zipWithIndex.sortWith { case ((a, i), (b, j)) =>?

Do we want scala-steward btw?

@Ginger WOW I didn't know that .gitignore was so powerful

@Steffan wdym?

That's amazing, you wrote a whole BOT in .gitignore

Better link for Scala Steward, actually. It's a bot that keeps sbt plugins and dependencies and stuff up-to-date

@Steffan you're looking at the wrong branch dumbnut

._.

In the code, it just sends [Permissions] NO if you send !!/die and aren't privileged, but continues to quit the bot

!!/die

[Permissions] NO

SEE YOU KID!

bruh

IT'S TIME TO BE A [Big Shot]

I may be stupid :b

!!/die

[Permissions] NO

that's stupid enough that I'm going to go fix it right now

IT'S TIME TO BE A [Big Shot]

s

!!/die

[Permissions] NO

lmao

!!die

stop

IT'S TIME TO BE A [Big Shot]

!!/die

[Permissions] NO

What other commands are there?

STOP KILLING THE BOT

I'M TRYING TO FIX IT

This isn't interfering with you fixing the code, right?

message sending? no

Let us have our fun. Pleeeeease?

but restarting it sure is

@user no

@Ginger Really? How?

Bot works in mysterious ways :b

i suppose the admins rn are just you and lyxal?

but has been shut down

Btw, I don't think it's necessary for it to announce itself every time it joins. We can see it in the users list

(quick interlude and a reminder for me for after lunch: I need to look into why a variable followed by execute-lambda at the end of a program seems to be moved to the front - also, look into whether or not variables are being combined into groups in the arity grouping phase. Also I need to specifically test to see how something like ... $abs execute-lambda is being parsed - I'll make a test for it)

Resume bot discussion

IT'S TIME TO BE A [Big Shot]

it'll probably work now

@user it's so I can easily tell if it crashed

@Steffan hyper, lyxal, and I

!!/die

[Permissions] NO

ok it work

!!/díe

[DíE]!? LYXAL!? WHAT ARE YOU TALKING ABOUT!?

nice

lol

go commit díe

→ 2 messages moved to ­Trash

You'll have to me-proof this:

E

true

@emanresuA eh? looks fine to me

Click on the link ;)

Steffan153 released v1000000[](github.com/Vyxal/Sandbox/releases/tag/v1000) in Vyxal/Sandbox

Steffan153 created tag v1000 in Vyxal/Sandbox

@Ginger whatever you do don't add a dice rolling command

why not oh no I have fallen for your trap

@Ginger pxeger happens otherwise

oh no

Mfer froze the bot

I mean it did roll 4 billion dice q:

@lyxal chat.stackexchange.com/transcript/message/62982325#62982325

Do we want a bot to automatically update our libraries and sbt plugins and stuff?

sure

Why not

@user yes lambdas need the index for context variable m

Ah ok

I've got something locally not pushed yet

@steffan an idea for the website - the ability to open up multiple output and code tabs

You could send a program's output to a different box each time or run multiple programs at once

Can't people just open another tab in their browser for that?

All on the same page though

Convenience

And then you could allow programs to read from other output boxes and do some weird sort of multithreading program communication thing

Uh-oh

Who's going to implement all this multithreading program communication stuff?

Because webworkers are pain

You can't have two at once?

You can

Well then you just make multiple webworkers

Communicating between them will require callbacks or something, right?

And they all include which box they're sending output to in their messages

@user no no the program literally scrapes the output box :p

o.O

The web workers just have a number assigned to them

Or something like that

They already have a session value if that helps

@user you say mutability bad, but it's all my brain could handle at 11pm :p

I mean, mutability isn't always bad, it's just that to gain admittance into the FP community, I must yell MUTABILITY BAD whenever I see the word var

@lyxal It did actually terminate the following day.

that's crazy

it took a whole day to roll a random number?!

all day

the backlog on the bot was one of the backlogs ever

I thought surely it would just crash or segfault or something

it just kept going

@Steffan Generating numbers takes O(n^n), didn't you know

Generating the nth natural number is one of the hardest problems in CS

am I an nth natural number?

Any more feedback on my latest Vyxal MVP? I think I only heard from mathcat and ginger

They complained it was too high contrast

lol it's not done

it's an MVP for a reason

that's why it's feedback :p

I'm going to implement it but I'm going to do some other things first lol, like adding inputs and moving buttons where they belong

And getting Pinia set up

What's pinia again?

Store

pinia.vuejs.org

@Steffan Love it! I'm not sure the option to open a tab in a new window is needed, though

^

it's really cool that you can change the layout

I only ever went with the TIO look because it was a safe layout style

but this new way of doing this is epic

{ctx-n 0 < if ctx-n negate else ctx-n endif} :=abs
[12|4|7|3|9|3N] sort-lam $abs execute-lambda | negate end

> [ 3 | -3 | 4 | 7 | 9 | 12 ]

Pushed a new commit to use Pinia

I disabled the maximize on tabs, I'd hate it to accidentally click it lol

{ctx-n 0 < if ctx-n negate else ctx-n endif} :=abs
[12|4|7|3|9|3N] sort-lam $abs end

Are we... implementing vyxal in vyxal?

no

it's because there's no abs element yet :p

and I needed one for testing

the point was more that you don't need execute-lambda after $abs anymore

I think the multi-branch stuff of the PR is mostly done (except for reduce lambdas with multiple branches)

time for the parameter parsing part

I'm thinking to just reuse the same : syntax used in v-2

so you might write lambda x:y:z body $x $y add $z times end

and that'll make a lambda with arguments x, y and z

or lambda x:1:y ...

which puts an argument into x, pops 1 value and pushes it to the stack and then an argument into y

and * for varargs

lambda ~ body ... end could be for lambdas that operate on the stack

I set up the sidebar and moved the run button there. Going to bed now lol

o/

ooh looks nice

@Steffan Looks really nice. Minor bug, you can see the options to maximize and stuff if you move a tab ^

Gonna switch : to , for lambda parameters

Because it looks nicer in literate mode and consistency counts

!!/test

[Test]!? GINGER!? WHAT ARE YOU TALKING ABOUT!?

wow, the bot survived the night

that's a nice change :p

!!/die

SEE YOU KID!

IT'S TIME TO BE A [Big Shot]

bruh

I had to fix that

please don't, I want to see how long it'll run stably for

and now you've reset my progress

The fact it survived 7 hours indicates that it should run stably unless the server restarts

no, it really doesn't

not when it's sechat we're talking about :b

I clearly have more confidence in the library than you :p

in other news, I wrote a function for the parser and completely forgot to use it

and spent the last 5 minutes wondering why things weren't working

5 6 λx,y|#$x #$y +} Ė

> 11

and the tests fail

wonderful

   java.lang.NullPointerException: Cannot invoke "String.isEmpty()" because "$this" is null
[info]     at scala.collection.StringOps$.nonEmpty$extension(StringOps.scala:1168)
[info]     at vyxal.Context.pop(Context.scala:57)

huh

you know you've done something wrong when you get a java error in your scala

@lyxal lemme just say I have never seen a library that loved failing silently after working fine for a week more than sechat v1

github.com/Vyxal/Vyxal/security/dependabot/1?

1 2 3 3 lambda * do fold- add end execute-lambda

> 6

aight, time for more bot work

today we will be adding commands

yknow I just realized the bot is offline :|

I didn't do it

IT'S TIME TO BE A [Big Shot]

> ValueError: 34 is not a valid EventType

I'm sorry what

wtf is event 34

34: "EventUserNameOrAvatarChanged",

you caused it

you killed the bot yourself :p

fuq

dignus :p

wait, where is that list?

github.com/Vyxal/VyxalBotSE/blob/master/chatbot.py

weird, I'm pretty sure I copied that list verbatim when I made sechat v1

bruh

I did, but I mistyped the last 3 event numbers when I copied them over back in *checks GitHub* June 2022

that's been a bug for like half a year wow

SEE YOU KID!

IT'S TIME TO BE A [Big Shot]

ok, that should be fixed now

anyway, time for commands

SEE YOU KID!

bot will be in the sandbox for a while

is it safe to do things that would otherwise send webhooks?

yes lol

if the bot isn't running nginx will just send a 502 back to GitHub

@Steffan @user #1611 - better lambdas is now ready for review

it needs tests before merging though

now, what's the best way to implement help

but it's 12:41am and I wanted to click button, so I'll add them later

o/

\o

Should we send a delegate to APL Seeds '23? d:

@mathcat this isn't anything to worry about, is it?

@mathcat are we using py directly?

lemme check

nope, we can just bump our pytest version to fix this

oh good

github.com/Vyxal/Vyxal/pull/1612

hmm ig you need to update poetry.lock too

wait I'll pull

just a second, we've got errors

lmao it needs to be locked

one second...

ok done

can we get some approvals

we can merge tho, right?

yes

boom done

back to golfing

I need to maintain my #1 week spot

@user Oops. That's because the options to show maximize or not is set in my default layout, so when the user moves them around, it resets

I don't think I can fix that

interesting design

maybe bigger font size for accessibility™?

I'm pretty sure wcag recommends 16px or larger

the tab names are 12px but they're not something long and important

yeah ig

just made my first vyxal answer :)

\o/ congrats

congrats

argh ninja'd by like 10 ms :P

already outgolfed thethonnu lol

bye 4 bytes -.-

Actually, I guess I can just select the maximize button in CSS and force it to be hidden

@Jacob oh wow

IT'S TIME TO BE A [Big Shot]

ok, the bot has some new commands

I'm going to be afk for a while so I'm starting it back up

!!/status

I am doing ballet at the Bolshoi Theatre in Moscow

!!/status

I am doing benevolent things like giving back to the community by volunteering at the local soup kitchen

also, !!/status has some new functionality

I'll let the bot explain:

!!/help status

!!/status: I will tell you what I'm doing (maybe). If you don't have a sense of humor and/or are a bot developer, run !!/status boring for actual status information.

!!/do you do ne do ne do ne do ne do ne are is do ne do ne do ne do ne do ne is are are are are is is is is is is is is are do ne do ne do ne is are are is are is are is do ne are is do ?

!!bug!!

the only new commands I added are !!/help, !!/status, and !!/info

so nothing else works

Does it use the gpt3 generated status messages tho?

it also doesn't have the various regexes from the old bot implemented yet, so there's only one way to run commands

@mathcat I copied the status messages verbatim, so if the old bot had them the new one does too

noice

we still haven't seen many of those

wait, were they in data.yml?

those were the only ones I copied

lemme check

yup that's right

the flowey quotes are almost nostalgic :p

!!/status

I am doing an environmental scan of the local area. You should really stop leaving those banana peels everywhere, it's attracting bugs

lmao

lol

Random question: Who actually started the cookie meme?

I'm not sure if it was user, lyxal or pygamer0

lol

Header and footer is now added to the website

I proposed the question a while ago and only got one answer: CM6 or Monaco?

monaco just because ik what that is :P

is the changes published?

@Seggan Ofc

CM6 is just CodeMirror 6

Atm I just have a textarea, so I ened to know CM6 or Monaco so I can use it

?

Welp people be gone

lol

Like I said, I'm kind of leaning towards using Monaco rn after actually trying it out

Yeah you have all VScode's nice features

CM6 was quite an improvement over v5 though, so it's pretty good

._.

I guess it's Monaco

We got 3 votes for monaco and 0 for cm6

Even though 8 billion people haven't voted

We can now flex that Vyxal is equivalent to vs code :d

!!/status boring

lmao yes