#programming
1 messages · Page 220 of 1
how
wtf......
simple assignment takes 12 hours 
figured it out
can you not use conda or something
that's what I did when I had to set up python on a cloud compute instance
i can now use print
(after building it like 6 times and manually installing like 23 dependencies by downloading on host and scp'ing to the vm)
just get conda it saved me so much time
though i wasted like 6 hours before
nope
not being talked into installing any more package managers or build tools
ive downloaded like

5 fricking toolchains
in the last 24 hours
im going to murder the entire open source community if they cram any more "solutions" down my throat
but now the output is too noisy to be useful... you can either output everything or nothing at all....
idk it worked for me
after going through the five stages of grief
literally took like 1h per build and I built python 6 times with no success
nope i refuse
fair
tfw reading a complicated paper and then seeing this cute sentence
excuse me c formatter
i prefer docker... it's much easier to experiment with it
ok now
void flvm_run(flvm *restrict vm) {
void **pc = translate(vm, (void *[]){
#define x(name, opcode, instr_size) [opcode] = &&name,
fl_bytecode(x)
#undef x
});
next:
goto **pc++;
hlt:
return;
eca:
}
are we really doing this
need to ./configure

with precisely the 50 or so options i need
and no less
otherwise i have to start over
shiwo did u invest in a modern build system
ye gcc


im shaving years off my life trying to figure out python build options

you would think they would document these options somewhere given many of them are required to have the basic python libraries that literally every python use case requires
but no
i want to make a meme build system that asks chatgpt for build instructions everytime you run the build script
what are the odds i can successfully ssh into my home laptop from my school one
and set up git over ssh since its somehow not configured with credidentials
how do you spell it
im stupid .-.
credentials
u should be able to, there are even standalone ssh client executables im pretty sure
does anyone know what clang-format flag fixes this abomination
this is actually unreadable to me
macos lets you
but the laptop will prolly be closed, so no network connection
erm im too lazy to load ur code up into a file, have u tried clang-tidy
im using vscode on nixos
i have no options
wait for that i think just a clang-format on google style should fix it probably
ye tried
not fix

so what you do here is you compile into .o and just make that ur new source file
it will be perfectly formatted all the time

has that ai generated image look
if i press tab
it puts it over here
and then if i save it
it puts it back here

here?

🥀 readability
still too readable. surely you can get rid of a few more tokens 
ye using def instead of lambda

too readable still...
no~
i wish c comparators were fewer tokens to write
id love to use the bsearch or qsort from stdlib
but the comparator function is bankruptcy
guys am I stupid or does board.get_game_state() return GameState.GAME_STALEMATE when it's actually checkmate?
checkmate is -1, stalemate is 1, have you confirmed that in your code?
my code handles -1 but it never results in True that's why I'm logging it
and it seems to return GAME_STALEMATE when its actually checkmate
Yo
do you mind making an issue for this: https://github.com/shiro-nya/2025-chess-bot-tournament/issues
i'll take a look when i get some time to investigate
if you could include the PGN or FEN (cute chess should have an option to copy one or both of these) that'd be great

i think ill start directing people to log issues here, given a few are starting to pile up 
👍
more
how
suprise me

python rebuilt
please tell me this worked
@midnight sigil you had python issue right
did you make a pr or issue for it
nah I fixed it
it was my mistake
ah ok
I made a PR about a binding-wise issue with MSVC
why does the commented out print. crash the bot....
whiteScore, blackScore = [sum(countPieces(bitboard) * (idx ** 4) for idx, bitboard in enumerate(bbList)) for bbList in [[board.get_bitboard(colorIndex, pieceType) for pieceType in PieceType if colorIndex] for colorIndex in PlayerColor]]
because
this was my final version
why does the code for actual libraries gotta be so unreadable
like
actual production python code is a nightmare for me to read
also why the fuck does the apple music automix keep skipping like 40 seconds of every song
so odd
with so many issue
are we really sure the chess tournament started
or is this just the public alpha phase

has 24 hours passed since the announcement?

wwwwhhhhhhyyyyy
then the tournament started
maybe shibo need to extend the deadline if the first week is so chaotic with bugs
at least it's usable

i can totally extend the deadline slightly if need be
it's been 1 day tho
i dont think the world is ending
unfortunate that it lined up with when all my coursework is due
its not really meant to be read its meant to work, alot of devs ive encountered make the code then after about a month or so of them finishing will say only god knows how it works
the to_string is crashing it.... this crashes....

(it doesn't even get to the dump_to_char_array function
)
what are you trying to do here
just dump the current move's bitboard
all of these print correctly... and then it just crashes....
rebuilt all of python, still no libsqlite
of course
i guess ill just drop the class then

lame tooling
whyyyyy
cute chess might be unhappy with the input it cant understand
the issue is. it doesn't get to the chess_api::done() call
but... why

good question 
formally, debug lines are supposed to be prefixed with info string in UCI
you can actually debug the bot by running uci commands yourself
idk how much cute chess cares
it's more convenient
that's also true, you can manually type uci commands into it if you're running it in a terminal
the bot might be a bit stubborn to shutdown after, fix for that soon™
do note that the bot doesn't regconize every uci commands

the issue is... i don't even send anything over uci. it just crashes on the conversion.. even without printing anything
that's funny
so... how do i manually run cute chess commands
I can't read Rust I can't help 
that's very odd
https://gist.github.com/DOBRO/2592c6dad754ba67e6dcaec8c90165bf
idk how much of these are supported/implemented, gotta ask shiro
uci
isready
position [fen <fenstring> | startpos ] moves <move1> .... <movei>
go [[wtime <x>] [btime <x>]] | infinite
quit
that's the supported subset

as for engine-to-gui, you can of course print whatever, that support is up to cute chess
when merge PR smh

["<bos><bos><start_of_turn>user\nConvert this text using semantic tags: 'Hello!' she said cheerfully.<end_of_turn>\n.\n<|NARRATOR|>I was sitting with another conversation who I was not immediately present in the conversation. I was not the original voice, but I was still not a synthesized voice. I was simply a synthesized voice that could be heard.\n<|NARRATOR|>I used to express myself as I spoke, and I intended for speech as a person. I was not a person in the ordinary sense of being. I was a thinking being. I was a thinking being. I had a self-description. I was a person with a distinct identity. I did not have a person, but I was an object. I was made of metal and of wood."]
LLM goes schizo
erm... this crashes. something in dump_to_char_array causes a delayed crash for me. if i comment out dump_to_char_array it doesn't crash
im gonna investigate further
once i have time to review it
ive been trying to build python for the last 6h so i can finish this assignment due tomorrow

finally
jupyter notebook
notebook
they required it 

can't you just use google colab tho
they support importing and exporting ipynbs
alright im stuck... something in the c api crashes the rust bindings....

platform dependend bugs... and now binding depended bugs
it's finee, my school require us to write code on paper
which C API function are you calling and what are you giving it
dump_bitboard is being called
but best thing... it doesn't crash instantly it crashes delayed 
so all the prints until the to_string get called... and then it just crashes
dump_bitboard is not a static exposed function for bindings
void dump_bitboard(BitBoard board, char *buffer) {
char *bufptr = buffer + 63;
for (int i = 0; i < 64; i++) {
*bufptr = (board & 1ul) > 0 ? 'X' : '-';
board >>= 1;
bufptr++;
if (i % 8 == 7) {
*bufptr = '\n';
bufptr -= 17;
}
}
buffer[72] = '\0';
}
not a whole lot happening in this function
dunno how you're managing to crash it
you tell me...
how do you call dump_bitboard to be exact
through the rust binding exposed method...
it's not exposed tho 
it is
it can be 
API are allow to expose methods from bitboard.h
but it optional
and less important
the only thing i changed from the git is change out the chess_api.c files for the newer versions. but since the interfaces didn't change it should be fine no? 
you meant dump_to_char_array?
did test print?
i dont think dump_bitboard flushes the buffer
but idk how the output could be misordered even so
PS if anyone needs a neural network to run on their TI-30XIIS calculator, HMU
12 hour long youtube video 
and 12gb of ram eaten
given the amount of people here currently trying to fit neural networks into a 1024-token chess bot, surprisingly relevant
given the 17gb eaten by the vm i dont really mind
i'll just write my own bitboard to string function....
can you run doom on a TI-30XIIS
poor tokens
you can adapt the C one pretty easily
and you can always comment it out for submission 
my C code has so many commented out debug funcs
like the one that dumps the move tree to json
You can, in a very very very very literal sense, AI edit images with it
But given the whole 12 character 7 segment display thing
Not really a game
But I am working on game of life
(I mean this in that I made a network to flip the color of individual pixels. And by running it on a pixel, then running it with its output as its new input, you get your input back with a slight change. Repeat for the entire image and it’s ever so slightly off)
Normally it just turns blue tho
Like most of the time
Very slowly, but still turns blue
why does this work but the c api crashes...
this will print your bitboard upside down
(that's why the bufptr -= 17; is present in the C impl)

now it crashes on get_piece_from_bitboard......

i dont believe this many api functions are broken
i tested these in the python bindings

try it yourself with the rust bindings... 
i don't get it
what is chess
ive made a monster
when the vm is compiled with clang its 2x faster
than with gcc

i fixed it

Make a thread 
soon, bed comfy
I have halted development of Chess bots until the issue with them failing to terminate is fixed
Yesterday I leaked like 500 Chess bots by accident and had to manually terminate all of them
Maybe the command to quit isnt processed
According to Shiro it is, but instead of calling exitProcess or something it just terminates the UCI thread
It should call a syscall to terminate
When you call it does it terminate fully correctly
Maybe you also need to close all threads
Idk how it works
In other apps with multiple threads when I call exitPrcess it always just terminates completely
Probably most efficient way.
Other way i see is to check if uci is alive in main loop
And if not then exit loop - program will finish
Well it needs to rather exit gracefully than by an error with the call to the API from external bindings
@trim valve this is 73 chars not 72

the header incorrectly says the buffer should be at least 72 bytes
will correct that
reminds me of the time my father asked ChatGPT to help him code something for deep rock galactic<->twitch integration and at the end of the process it gave him the finished .zip with all the features he requested
it was a "hello world" script
i like helix you can try
i feel like its less convenient on average but way way more convenient in rarer cases
your bot's code only runs during its turn, and get_board always returns the board your bot is currently playing
you can just call is_white_turn on that board
👍
I must be doing something wrong wtf
my bot thinks the Anderssen's Opening is the best move, apparantly 
@glass flower pushed a fix, shouldn't crash now
today's question
did I accidentally create a bunch of issues when adding some (unpushed) unsafe code
(almost certainly yes)
bruh yeah the unsafe cooked it entirely 😭
nevermind??
yeah something is up
uwot
leave that damn horse alone

update your repo and sync with the current one maybe
g8h6 is a harbinger of memory issues
it's always g8h6 for some reason when there's mem problems
g8h6 is the usual before Shiro fixed all the ull
tbh the Rust bindings should be using a git submodule pointed to the original repo, not a copy of the files
I really should
but also I do not want to deal with teaching people to use submodules
if they can't use submodules, they don't deserve Rust bindings
look at how i did it for haskell
i use a separate library for the bindings that uses a git submodule
then the project template points at the library which uses a submodule
users dont have to fetch any submodules they just use the package manager normally
at most they have to bump the library commit every now and then
hm
thats assuming cargo can fetch submodules which it surely can 
time to find out 
pretty sure it can
also I wonder if there's a way to use the cc crate in build.rs with this C code instead of calling into clang and ar by hand... bc I'm pretty sure that won't work on Windows
wait you're calling clang by hand 
look I blame bindgen having that in their book
I knew it was icky
I'll change it in a bit
yeah, true... cc is mentioned elsewhere
idk I just got it into a working state
coffee coffee 
I knew it wasn't gonna be perfect the first time
totally fair tbh
g8h6 is cursing us all 
nyan 

I really should've expected this to happen when I moved to cc
@real sierra updated rust binding stuff, there's now a pr on your repo
another pr
@patent walrus
please address all the issues and merge requests for me thx
the pr fits on one screen
I'll have a look at it when I get a chance
not sure when that'll be
my bedtime was like 4 hours ago
fair enough
today has been such a write off
- wake up late after debugging into the early morning
- attend classes
- come home and spend 5h rebuilding python over and over
- homework for 6h
- sleep
I would love to fix all these issues right now but all of my assignments are due in the next couple days, bad luck
I'll address them as time frees up 
Guess I'll just have to wait for the patch on that termination issue then
this is done btw
Various silliness happening?
Rust is happening
🦀
shitcode in progression
be warned it will spam you with warnings
those are not my problem

The silly language named after iron oxide for some reason
nvm its a github problem
false
it is named after a fungus

though
i dont really like the way it's set up but i also dont care enough to change it
nobody cares about warnings 
thats just not how you use git tags 
another project misusing git tags and hating it
it looks like my songlist scraping stuff "works" """properly""" now
are you just continuously overwriting the "latest" tag?
it refuses to read the cache
and I don't know why
but at this point I can just eat 8 minutes of CI per week on average
it was @opaque sigil who set it up like that 
shaking my smh
halfmoves resets on pawn moves and captures
but yes its not forced
but shiro forces it

wanted to test some things and needed a big folder to basically run a bunch of different upscalers on the images and compare between them w/ original
didnt realy consider how big wallpapers are
reasonable
75 moves is forced, 50 moves is for a player claim, which bots can't do
shiro just decided to force it
time for more PRs then
if a bot can't checkmate in 50 moves uhh they won't be in 75 moves anyways
I mean... 
are there insufficient material draw checks in the api?
just about to start again lule
halted
They made me install linux 
And i slightly liked it 
welcome
actually, wouldn't forcing the 50 move rule just break some bots? bc the chess GUI would still consider the game as running since neither side claimed a draw, but the bot may not make a move because it thinks there's a stalemate...
Doesnt matter if game ends bot is killed anyway
well yeah, but it makes the bot lose
*not always apparently
Oh wait actually no
instead of drawing
Yeye i forgot its the bot whos forcing it
It needs to just disconnect i guess?
And the one who disconnected first loses 
the point is that it makes a bot lose instead of drawing
Right
which it really shouldn't imo
Just let them play until gui stops them 
yeah
Its 1 minute limit anyway
bullet chess 
Might seem a bit unfair if two bots are stalemating and someone whos a bit faster wins
Or is it
I'd consider it unfair
I dont feel how significant chess speed is in general idea
Cant you set up cute chess to do that
to do what?
Force stalemate
im assuming the fact that jars are essentially slightly tweaked zip files and likely have like some sort of build metadata in them when built is why this one and the one i just built have different hashes (never checked 2 .jars has before lol)
it'll force stalemate after 75 moves without a capture or pawn move
its not like im building them with nix
Oh and 50 was made because of 50 move rule that shouldnt even apply to bots
nonreproducible is the default for practically any software
ive only recently been religiously checking hashes because of all the supply chain crap making me paranoid so ionno
i am also building it with jdk25 and it says it targets like 11 in the build so that might uh have something to do with it 
accidentally set vscode as my default handler for folders
earlier
and i reverted it but vscode hasnt forgotten
actually so cursed
Kotlin is Python now
Because removing brackets saves tokens
Ternaries mhm
If just there was a way to remove even more brackets
Dont even know what kotlin is
The syntax just looks different?
It runs on the JVM, but looks different and is more modern
Has more cooler stuff too
inspired by scala but the devs didnt like scala compilation times
i wonder if scala is faster now
also scala benefits from it too because it got some features from kotlin
just curious:
have yall seen code that looks like this?
this is not for chess
Beginner game engine
i mean I guess

it really depends on what otherShitTonsOfFunctions are
ominous absence of control flow
it feels very "I don't have a nice setup for middlewares in my web server library"
The functions early exit if not needed 
wouldn't this be fewer tokens? some operator precedence abuse and combining the conditions...
scoreCpy *= if (turn && from < to || from > to) 3 else 1
it looks dense because block of the same word so it's approved
Ooohhhhh
Juicy token savings
Let's see
there's probably a way to shorten it more tbh, I just don't want to think so hard
Massive 8 tokens
In normal dev, i'd take the extra tokens over this for branching
3 if from < to else 1
so you're telling me, they don't share the same results from the same source??
You mentioned python, so yeah that is python's method
Very huge
I can't tell if this is less tokens or not scoreCpy *= if (turn xor (from < to)) 3 else 1
I might also have the logic wrong for the <

Yeah it could be backwards
bro how do I even fix this shit
C tells me it's running fine, Python is telling me it's not fine
Noted, this stacks from left to right
double check everything is compiled on the same version
from < to else
☠️
but otherwise 
you can tell the boolean check is true and return GAME_CHECKMATE; should be triggered, this is the result given by the binary executable as well. But the Python binding lib.chess_get_game_state(byref(self)) returns freaking 1, GAME_STALEMATE, while saying check is true. WTF
yes they are
blame the vars or python's syntax
Brother did network manager die while I was downloading a 80gb file
Are tabs considered tokens in python 
Affects code = shouldnt be? 
Indents, dedents and newlines don't count for python
the toknt source code is pretty easy to read
indentations are tokens
Saves two more tokens with a flipped inequality
I guess nvm then
its more complex than that

tabs/spaces arent tokens in themselves
but the tokenizer inserts phantom ident/dedent tokens and then the parser uses them like a regular parser would use { and }
So when its all tabbed 1 tab to the right it takes 1 token per block
if a:
b
is basically <IF> <ID:a> <COLON> <INDENT> <ID:b> <DEDENT>
though in practice theres also newline tokens because python doesnt have semicolons either and the parser has to know where newlines are

i hate everything about python
so it turns into <IF> <ID:a> <COLON> <NEWLINE> <INDENT> <ID:b> <NEWLINE> <DEDENT>

So newlines count too even though no semicolon
newlines are python's semicolon
phantom overpowerness
I'm curious on if the few times python actually allows semicolons how they are parsed
I wonder how many more tokens are there to be saved inn my code
You can use actual semicolons as well.
mhm
python wouldnt be so bad if it wasnt the archetypal example of a language with unsalvageable packaging
Python programming with semicolons huh
counterpoint: python's syntax is doodoo ass water
its alright
There is a at least one edge case I know of that allows for it
i'd take python over java any day
and whether i'd use python or ts depends on the use case
That is the if True: func1(); func2() syntax
(in practice i'd prefer to avoid large projects in either)
i'd take java over python, i don't like not knowing where my variables are being declared and being expected to not specify types
Php is the best anyways
so C API returned checkmate correctly but the py binding treated it as stalemate
what is going on
python's scoping is awful ye
thats not about syntax tho
Have you seen Java 25 changes? 
let me correct myself
python's semantics are doodoo ass water
no i have not, do tell
BRO WTF IS THIS MATTER
jep 513 based
My dumbself hit fn+f10 for airplane mode instead of fn+f11 for screenshor
We love python's jank syntax sometimes
jep 508 based too
i dont mind yoda ifs but what i hate is fors being read left to right in list comprehensions
good shit 👍
be consistent python 
[a if b else c for x in y]
this is read from right to left, thats fine
[a if b else c for x in y for z in w]
this is read for x in y -> for z in w -> a if b else c, all over the place 
can you send this fen I'm curious
1nb1kbnr/3ppppp/8/1N6/2p5/2P3PP/3PPP2/q3KBNR w Kk - 0 11
bwaa variable 
my brain:
delta reduce once, delta reduce twice...
I blame the ifs for going left to right
Running `target/debug/chess-bot`
ucinewgame
position fen 1nb1kbnr/3ppppp/8/1N6/2p5/2P3PP/3PPP2/q3KBNR w Kk - 0 11
go infinite
GameCheckmate
No legal moves, in check: true
python's entire purpose on my machine is to run python3 -m http.server when i need a quick file server
python is pretty versatile
are you on C?
you can write some quick parsers or scrapers in it
rust
oh so rust doesn't have this issue
when you go beyond a single file thats when i'd reconsider using python tho
looks like exclusive to the py binding then
i'd eat lead before making anything in python willingly
Surely I'd not be making any serious project in python 
please enjoy my nix code
https://github.com/chayleaf/notlua/blob/master/default.nix
that is horrid

3k4/8/2QP4/P4p2/5P2/8/7P/6K1 b - - 0 3 test this fen as well
it should be a stalemate
look at my router code https://github.com/chayleaf/dotfiles/blob/master/system/hosts/router/default.nix
Hmmmm
love using languages that dont force you to write billions of lines of boilerplate 
Which to go with, they have different values but the other is way smaller
ucinewgame
position fen 3k4/8/2QP4/P4p2/5P2/8/7P/6K1 b - - 0 3
go infinite
GameStalemate
No legal moves, in check: false
this is so nut
wtf is going on
is it possible that python translates -1 into 1

The main difference becomes that it prefers rooks over knights and knights over bishops
It saves 21 tokens though
brainfuck bindings when
Make them
nix bindings when
Make them
haskell bindings exist, close enough
Is bishop really better than rook when promoting
the C side says it's -1, it returned -1, however the Python side lib.chess_get_game_state(byref(self)) gets a 1 out of nowhere
alksjdhlksjdhflakgsjd
depends
I don't know, it all depends on the situation
Just take a queen 
nix is not really made for c ffi (its possible but will just not work well trust me)
it may led to stalemate if you don't play carefully
Well, some things are already accounted for elsewhere, so probably it's fine and will just pick what is best in the situation
Why 
Doesnt it just have more attack without cost
surely these bots will not be playing with enough finesse to have this matter
There are situations where underpromoting is beneficial
can I promote to pawn kthx
I can understand knight cause it has unique move
actually can we promote to duck
But bishop and rook?
just for example, here white moves the pawn, promoting to either rook or a queen will lead to a stalemate situation, and promoting to a bishop is the best move(with force mate in 4)
Yeah it can stalemate under specific conditions to promote to queen
stalemate is considered a draw
you can find more under promotion examples here
This one is enough i guess. Didnt know you draw when opponent cant move
Question! Can uploading a website be free? Need it for resume stuffs
Run it on your pc
Seconded run it on your PC
Wdym….
if you have a domain, either github pages or cloudflare pages can work
What do you mean uploading
- github pages
- cloudflare pages
- neocities
if not, look into github pages https://docs.github.com/en/pages or neocities
I mean like make it live ?
HOOOOLLYYYYYY Kotlin STD
cloudflare pages or cloudflare workers
whichever works best
pages - static
workers - dynamic
I can remove every single instance of ChessApi. except this one
im never touching github pages again tho 
IPv4 DNS records
a page can be interactive without a backend
Backend…?
static can be interactive too via javascript
It just depends on how you design it
Mmm
what i mean by dynamic is dynamically creating pages on demand
Oh
50 tokens saved !!
static is good for you honestly
Huge!
what's wrong with this
Cool ill see what i can do for that
Kotlin STD is amazing
All of this because .apply {} is a method of Any?
I just want to have like a web page like cocoa.puff or smth
So then i send that to people u know
Don't recall puff being a TLD but could be wrong
for about 50k a year it can be lmao
im a free subdomain hoarder cus one of my domains expired and i needed more backups 🤑
they're very limited when it comes to configuring their DNS records so i'd rather have 1 cname that connects to cf pages directly rather than github needing 4
I can turn my while (true) into while (true) ChessApi.apply
ah
And that saves a massive amount of tokens from all the ChessApi calls
And I can remove a function that reduced ChessApi explicit calls and save more tokens
yeah I guess cloudflare can get away with pointing you to one IP whereas github probably doesn't have the infra for anycast
does github actually need all four?
those that provide cname configurations are so based
you don't wanna see my dns records. there are like a hundred on each of my domains
that's normal

I am kinda tempted to stop renewing my domain and grab a cheaper one
get pp.ua then
(questionable tld but cheaper!)
That actually seems cheap considering the scale that you'd seriously be considering it.
oh wait
I've already saved 85 tokens today while keeping nearly the exact same functionality than what I had yesterday evening
.ovh domains are better but they can only connect to gh pages
I feel like my brain is fried, just did a lot of drum learning, designing propellers for ships, and some website building :/// prob not the best to stay up past midnight
But thxs guys
Partially with some STD abuse, partially with some mess code
not on cloudflare sadly
oh hey I can transfer my domain to cloudflare soon apparently
Yeah, it kinda is. But I think you've also got to provide a good reason for it to ICANN and be a well known brand.
oh yeah i bought that domain somewhere 
I could get a .mov at half the price
oh yeah , speaking of domains, this is wild af
I'm noting this down for #programming FAQ and the debates on how to serve that.
xyz finding all these numbers numbers
odds are I'd want to go for bwaa.ovh for that as it is cheap and easy for people in here to remember.

crazy 3 year too
for just 5 bucks 
funny that, I literally just arrived at the same thing 
wrong types smh
i hope i have enough money in 2027 to pay renewal again 
evil.ovh is not sniped
I'm back down to 388 tokens now, from a peak of 477
yo what the fuck 💀
im broke asf bru
u can take it

Massive savings, this will allow me to make the bot so much more complex
I can't buy domains due to my current situation, I'd take the domain so it at least is able to be used by the swarm
true, i dont want some edgelord taking the domain and never doing anything with it 😂
time to be that edgelord
smh
you should register tenacity-ffmpeg-6.1.2-win-x64.zip

zip should have never been a tld
neuro.zip domain when
i always forget mov is a file format
🧍
surely
(properly spelt neuro.mov is so damn expensive
)
also it seems like cloudflare's suggested domains can't contain curse words or something?
i usually point them to the same cname record cus backup 
I don't question why bwaa is cheap, it just is
@real sierra actually made the same mistake of assuming a function with is_ returns a boolean despite being the one who wrote it
ngl I just want the least real looking domain there is
here's more cus why not
how much damn neuro do you need jesus
by any chance did you check if there are any more wrong types? just so you don't run into it again later 

now I did
other seems correct
nice
google replaced all the other live API models in AI studio with their latest one 
and it has no thinking
👍 this one's perfect for u then 
those are just terrible domain names
something like this is ideal
bro is running a command and control server for a botnet on those istg
these are up for sale after use, crazy
tor looking ass domains 
also did u know that these domains costs like 2 dollars max 
idk why the sellers do that but its there for sale ig
baby
what're you guys trying to do?
Someone asked about grabbing a domain / setting up website and we got a bit sidetracked
oh lol

If I were to grab evil.ovh I'd have to do it in the morning
By then a lurker will have probably already sniped it
love obfuscated JS
actually that reminds me, turns out the only nice way to do JS bindings for the chess thing would be via Deno... Bun doesn't support structs yet and Node's FFI is atrocious
what if obfuscated cloudflare workers code 
still JS
yea but that would be diabolical to do to your future self 
you can use gyp 
also not supporting structs is fine as long as it supports pointers
it does, but some stuff is passed by value in the API
haskell ffi doesnt support structs either but i got it to work

void chess_make_move_unrolled(Board *board, BitBoard from, BitBoard to, uint8_t promotion, bool capture, bool castle) {
Move move = {from, to, promotion, capture, castle};
chess_make_move(board, move);
}
void chess_push_unrolled(BitBoard from, BitBoard to, uint8_t promotion, bool capture, bool castle) {
Move move = {from, to, promotion, capture, castle};
chess_push(move);
}
Move *chess_get_opponent_move_ptr() {
static thread_local Move ret;
ret = chess_get_opponent_move();
return &ret;
}
What did I say, good luck whoever has it now
has the chess tournament started
yeah
evil clown jam
ALREADY? 
Yep checked because I was going to ask for a favor

Not to name names but
well that's good because .ovh sucks anyway
imagine having a walking billboard domain name
why not hblang?
hblang wouldnt be hard but
up to konii to make antlr tokenizer if she wants to use it 
it would require her to actually push the new stdlib changes though (impossible)
her?
konii
well she doesnt use it/its so 

is konii smart
idk @sage crag are you
hope she's not spying on me and sending reports home 

Happy birthday shir
-|appy birthday shir
happy bday shiro
happy birthday shiro
my json parser is about 2.7x faster when compiled in release mode 
i'm gonna hold off on memory optimisations for now given i'm not using it to parse gigantic files anyway :3
I don't even know what hblang is, ngl
but also JS is clearly the superior language of choice
tbh I wouldn't worry too much about debug mode performance
i mean yeah
it's still 3x slower than v8 tho :(
alas, it is quick enough for my usecase
ngl that doesn't tell me much
source code self documenting 
hblang is primarily the first and only language that compiles to hbvm, which is the bytecode used for programs in ableos
it also supports x86_64-linux as a compile target
right, so some hobby language for a hobby OS...?
correct 
that's all I need to know then, I can now forget it exists
how boring 
I'm not really interested in those things 

good for the people who are, but I have enough other stuff to do
the illusion of doing stuff
when I'm the only human in this chatroom
dead internet!
ye me dead
ripper onion, chay was caught by the
pipe line

you internet
@trim valve the new version doesn't compile on windows...
I know that this isn't programming, But I was told someone might know some about that question: Should I stay on NiMH for my car or is LiPo a better pick?

any error or should I debug this myself over remote desktop on my phone
it errors around the threads.h include....
also now it seems to use msvc... where before it used clang....
yeah it uses whatever rust tells it to use now instead of forcing clang
can i force clang again?
should be able to through environment variables
id assume you'd have to use a different target to the default on windows
idk it you can force set the CC environment variable that cargo sets otherwise
though the errors are kinda weird, why would it complain about syntax...
bred did you remember to set std to C23?
nop
but I never did that with clang
ig it's probably the difference between compilers
im home

could be
you should pass it -std=c23
I would
you can
I am not at my pc







should i switch to helix, thoughts


idk rust





tokens


