development

[sharp geyser]

I guess in case some software intercepts it and "creates" it itself

it tells you what exactly is creating it

[frosty gale]

it gets stored on your phone locally

if your phone has a TEE it gets stored there (aka hardware attestation)

if your windows machine has a TPM it gets stored there also

[quartz kindle]

what if i lose the phone? arent they keys backed up to google?

[sharp geyser]

I think they are

Cause I have keys I saved to my android on my iphone because I have google logged in

[frosty gale]

not sure if its backed up to google but its possible

[quartz kindle]

so basically we are handing it all to google

[sharp geyser]

not sure

[frosty gale]

im sure if that happens you can toggle it off though

as bad as google is they cant force you to

[quartz kindle]

my main concern tho is how to access my accounts if i lose my phone and/or google account

[sharp geyser]

sucks to suck

don't lose your phone L

[frosty gale]

similar problem as to losing a password manager or password

but im sure you can backup the keys somehow

its probably OS dependent

you can view them here on windows https://support.microsoft.com/en-us/windows/manage-your-passkeys-in-windows-6a70599a-25e1-4461-86be-d67d1023c69f#:~:text=You can use the Settings,saved for apps or websites.&text=Note%3A some passkeys for login,the device and Microsoft services.

[sharp geyser]

I mean at a certain point there is no way to prevent something

they can allow you to backup your keys, but if you lose access to the places you backed them up there's nothing else you can do except maybe get new ones generated but idk how that works

[civic scroll]

tim

[quartz kindle]

ok i just tested it out, looks pretty cool

[civic scroll]

i think i will need to learn shaders if i want to make anyway remotely cool

[quartz kindle]

when logging in i can chose whether i want to use username+password or a passkey device

so i can still use user+pass if i dont have myphone

[quartz kindle]

i pray for your mental health

[frosty gale]

and the best part is the server only gets your public key, so if the company/server gets breached no user credentials get leaked

[sharp geyser]

damn no rust implementation

im not trying to impl webauthn myself

fuckkkk that

[frosty gale]

its really easy actually

i did it in rust

[civic scroll]

i made this and thought it would be infinitely more cool if it has some kind of shader animations to it

[frosty gale]

my backend is going to be in rust

it was easier to implement than in nodejs

[sharp geyser]

I dont have the paitence to use rust rn

[quartz kindle]

:O

awesome

[frosty gale]

but there are a few steps involved and you need to understand what to verify

so its best to use a library

there is one for rust

[sharp geyser]

where

they must of not updated teh website for it then

https://crates.io/crates/webauthn-rs

[frosty gale]

https://github.com/kanidm/webauthn-rs

[sharp geyser]

I guess there is that

[frosty gale]

you think the rust community wouldve missed a single opportunity to port something to their language?

[sharp geyser]

webauthn is cool ig, but I dont see its uses for me rn

[civic scroll]

thank

[sharp geyser]

Actually I can use it for vidcord

not any of my other ones

[sharp geyser]

not really

[quartz kindle]

icba seting up vaulwarden, jhust gonna use bitwarden itself

now i just need to manually move over all my 50+ accounts that i have written in plain text

:^)

[frosty gale]

bad thing with it though is the webauthn standard supports authenticator attestation

that means a website can force you to use a single authenticator to login

specific*

[quartz kindle]

that makes sense for a business perspective, a company might want employees to use the company's own internal authenticator

[frosty gale]

google could make it so if you want to use webauthn, you have to have an unrooted phone or another phone approved by google/the company

if its rooted the TEE will not sign the attestation

and google will know

but thats the only downside and i doubt theyd do that

[quartz kindle]

also, i know why my windows doesnt have the option to store locally

i dont use lock screen

:^)

[frosty gale]

most secure IT system in brazil

[quartz kindle]

i havent used a windows password in 20 years

[frosty gale]

longer than how old most people on this server are

[sharp geyser]

tim

you need a password.....

omw to tims computer rn

[quartz kindle]

good luck finding your way through the brazilian jungle

[frosty gale]

Windows encrypts your passwords and passkeys with your password so that makes sense

[quartz kindle]

i've always found windows passwords the most useless thing in the world

because 1, theres only me, and 2, i can just boot into some recovery iso and remove the password

[sharp geyser]

dw, im very resourceful

Just tell me a 50 ft radius at least

[quartz kindle]

:^)

[civic scroll]

perhaps your windows security is the brazil itself

[quartz kindle]

third world security: danger

[sharp geyser]

yea tbf

Id rather jump off a mountain than go through the brazilian jungle

[radiant kraken]

my country's government literally only use windows defender as a security measure in their database 😭

[quartz kindle]

when relatively famous but loved-by-community people make a kickstarter campaign

[civic scroll]

this is timnet

[quartz kindle]

i wish it was me

[civic scroll]

me too

[frosty gale]

secure boot + bitlocker + password protected bios:

[quartz kindle]

bootlicker

[frosty gale]

• rust backend to ensure prevention of memory corruption

[sharp geyser]

Right, so im not too sure on how swap memory works, but if a program requires at least 12gb of ram to function properly...could I substitute that with swap if I don't have the full 12gb?

[quartz kindle]

you dont control swap from inside the program

[sharp geyser]

Well no, but as the program is using ram, I can make the os prefer swap over ram no?

like on linux vm.swappiness controls how often the os swaps between ram and swap memory

[lament rock]

Yeah you can adjust the "swapiness"

[sharp geyser]

but I am wondering will this be sufficient

I dont want to upgrade to a higher ram server rn to run a test rust server

so I was wondering if I could just use swap in place of the missing ram needed

[lament rock]

I think your best bet is to optimize your program first

[quartz kindle]

you dont need to tho

[sharp geyser]

wdym

rust requires at least 8-10gb (12gb preferably)

I only got 6

[quartz kindle]

the program will ask for whatever memory it needs, the OS will simply do its best, which includes many things, including swap, memory compression, paging files, etc...

[sharp geyser]

ok

but tim

thats not what im asking

[quartz kindle]

if the OS cant meet the requirements, the program will just run slow or crash

[sharp geyser]

I know the os will do its best

but will swap be sufficient to fill in the holes or is that not how swap works

[lament rock]

Swap usually starts by dumping inactive programs memory to disk

[quartz kindle]

if you have 6gb ram and 6gb swap, it will be sufficient to run the program, it will just run slowly

[lament rock]

The memory still has to be in physical memory at some point

[sharp geyser]

Fair enough thank you

[lament rock]

But if the same program has to use all physical memory and swap at the same time it probably wont work

[sharp geyser]

well

Its worth a try

I don't have the money to upgrade rn, and quite frankly I am bored and want to try my hand at making rust plugins again

[eternal osprey]

const enum variable6 {
test = 1, test2 = 3
}
console.log(typeof (3 as variable6))

why does this still show as a number?

like i can remember, in java we could take normal numbers, and cast them to enum types.
How can we do that in ts?
I assume that as is like cast in java/c++?

[neon leaf]

typescript?

[eternal osprey]

ye

[neon leaf]

as in typescript only changes the ts type

not the js type

in this case you would need to write custom logic to do this

there is no way to cast a number to an enum built in

[quartz kindle]

look at the compiled code and you'll see why

[sharp geyser]

poise?

I mean, you can definitely make use of macros here

It would be similar to what @radiant kraken made for our rocket routes

let me

get the code

one sec

#[macro_export]
macro_rules! gen_command_routes {
    ($($command:ident),*) => {
        $(mod $command;)*

        #[inline(always)]
        pub fn routes() -> Vec<Something> {
            vec![$($command::$command),*]
        }
    }
}

and you'd use it

gen_command_routes! {
  ping
}

providing that your ping cmd file is called ping and your command fn is also called ping

so

// ping.rs
pub async fn ping() {}

then you would just

commands::routes() where you tell poise what commands to load

Sorry if this is confusing (I myself hardly understand macros) I just vaguely know how to read it

@covert gale It helps to think macros like regex

* matches zero or more

tbh

macros are exactly just regex

but rust specific regex

btw

that macro expands to

mod ping;

pub fn routes() -> Vec<Something> {
  vec![ping::ping,]
}

now if you gave it multiple

gen_command_routes! {
  ping,
  avatar
}

it expands to

mod ping;
mod avatar;

pub fn routes() -> Vec<Something> {
  vec![ping::ping,avatar::avatar,]
}

Let me break it down for you

#[macro_export]
macro_rules! gen_command_routes {
    ($($command:ident),*) => {
        $(mod $command;)*

        #[inline(always)]
        pub fn routes() -> Vec<Something> {
            vec![$($command::$command),*]
        }
    }
}

Note: $ declares a macro variable containing the matched result
($($command:ident),*) - Match a identifiers that can either be zero or more (aka you can have zero, or you can have more than zero)
$(mod command) - basically puts mod ping into the file
No idea what #[inline(always)] does tbh, I know it inlines it but no idea why you specify always
vec![$($command::$command),*] - Specify that you are putting in, zero or more references to $command (the macro variable containing the identifier passed to the macro) becoming something like vec![ping::ping

@covert gale ^^^

This is the best I can do to explain it

@radiant kraken can likely explain it better if I missed anything

Pretty much

Its quite literally just rust specific regex

https://doc.rust-lang.org/book/ch19-06-macros.html

Here is a detailed explanation

I cba to read it tho

No problemo

Just make sure to credit null

realistically you'd be doing this in a mod file inside the folder commands

// mod.rs inside commands folder
gen_command_routes! {
  ping
}

then inside main.rs

commands::routes()

to get the vector of commands

[proven lantern]

here's a good comparison of rust and zig
https://www.youtube.com/watch?v=rKXsZno9ijw&t=176s

[sharp geyser]

[proven lantern]

Using zig in sandboxed environments like AWS lambda is even a use case to use zig over rust
https://youtu.be/rKXsZno9ijw?t=778

[radiant kraken]

inline always strongly suggest the rust compiler to inline the function, as just #[inline] is not a 100% guarantee but merely a suggestion

[radiant kraken]

you don't need to! hahahaha

you explained it pretty well :)

[frosty gale]

programmers insisting on explicitly specifying inline even though single line functions already get inlined automatically

if applicable*

depending on usage of functions you can make inlining impossible

[winged linden]

can someone help? i am having a weird ass issue

const { client } = require("../../ufo");

if i have this in a specific file it works, if in another file it doesnt

[lyric mountain]

is the "another file" in the same folder?

[winged linden]

in extra, its undefined in in BotStatus its working fine

[lyric mountain]

did u save the file?

[quartz kindle]

does ufo.js require extra.js?

[winged linden]

ye ye lol

[winged linden]

not really

[quartz kindle]

does ufo.js require a file that requires extra.js?

[lyric mountain]

that import chain restriction is ridiculous lmao

[winged linden]

no lol

[lyric mountain]

where is ufo in the file tree?

[quartz kindle]

you sure about that? because this smells like a circular dependency

[winged linden]

outside functions and then outside src

[lyric mountain]

try creating a new file and importing it

if it works, then it's circ depend

[quartz kindle]

easy way to test (and fix), in extra.js, remove the require from the top of the file, and put it inside the function that needs to access the client

[winged linden]

yeah it works like that

[lyric mountain]

then it's what tim said

[winged linden]

hmm

[lyric mountain]

I dont get why js has issues with that, only other lang I know that cannot deal with bidi imports is delphi

[quartz kindle]

because cjs is line-by-line sync

it loads a file and parses it, when it reads another require, it loads the next file and parses that, before finishing parsing the previous file

[lyric mountain]

that sounds inefficient

[quartz kindle]

welp, thats how they chose to do it

because the file being required is not just parsed, it is also executed

[lyric mountain]

does esm also have that issue?

[quartz kindle]

and to be executed, it needs to have everything it needs loaded

[quartz kindle]

it does have a few more ways of dealing with it, but yes, it has the same issues

because like, if you have a file1.js like this:

console.log(10)

and you require that file, it will log, because the code executes

buf if your file1.js is like this: ```js
require("saomething else")
// or import something else

console.log(10)

the console.log needs to execute regardless

so what if it needs to log the contents of the loaded files?

not possible to log that without the full file chain being resolved synchronously

[lyric mountain]

so if a single lib in the chain has bad code like an infini loop then it'll get stuck if a single file references it

[quartz kindle]

yes

unlike compiled langs where a compiler examines the code and puts it together beforehand, an interpreted lang will dynamically load and run the file during runtime

// a.js
const b = require("./b.js");
console.log(b);
module.exports = "A";

// b.js
const a = require("./a.js");
console.log(a);
module.exports = "B";

what would happen if you run either of these?

a.js needs to log the contents exported by b.js, but when b.js loads, it needs to first log the contents exported by a.js, but a.js is still waiting for b.js to deliver its exported content, but it wont do it before it logs what a.js is exporting

so it will loop infinitely until it exceeds maximum stack size (not sure about this, maybe node is smart and just stops it from looping), and just return undefined

[winged linden]

Basically i might have found the issue

ufo.js requires a path called api.js
api.js requires ufo.js (it works fine though)
api.js requires a functions file extra.js
extra.js requires client.js

[quartz kindle]

generally, one should never export anything from the main file

if ufo.js is your main file, the one you run node ufo.js on, it should never be exporting anything

if it has something that needs to be exported, put that something in a separate file and import it in the main file

[winged linden]

i am only exporting redisClient & the discord client

i think the issue is with the redis client

gonna do it in a seperate file

thanks 🤝

[quartz kindle]

also, ideally you should not be exporting created instances

for example if you have a class in a file (like the discord "Client" class), the file should export the class, and the file who imports it should instantiate (new Client) the class

the file where the class is should not export a "new Client" directly

its also better for the class or instance of a class to call/require the functions/files it needs, and not the other way around

[winged linden]

how do you properly pass the discord client to your api routes file then?

[quartz kindle]

a function in extras.js should not require client, the client should require extras.js

[quartz kindle]

if the bot is the main program, when you run the bot, you create the api server and pass the bot instance as a parameter to use in the routes
if the api server is the main program and the bot client is just an extra, you create the api server and inside it somethere you create the bot client to use in the routes

there are a million different ways to do things, but in general there is usually 1 thing that is the main thing the program is doing, and everything else is extras or addons to the main program

its much easier to design things with that in mind than trying to have multiple "main things"

[winged linden]

Yeah exactly

Main thing is obv the bot in my case

the api is an addition for my dashboard

const app = express();
app.use(cors());
app.use(bodyParser.json());
app.listen(PORT, () => {
  console.log(`API running on port ${PORT}`);
});
app.use("/", require("./routes/api"));```

thats how i currently do it, this is in the main ufo.js file

[lyric mountain]

why ufo btw

[winged linden]

bot name haha

[lyric mountain]

ah

[quartz kindle]

how does api.js look like?

[winged linden]

This is what it requires

const router = express.Router();
const { v4: uuidv4 } = require("uuid");
const { redisClient } = require("../ufo");
const { PublicKey } = require("@solana/web3.js");
const { Web3 } = require("web3");
const { BotStatus } = require("../src/functions/BotStatus");
const web3 = new Web3();
const { client } = require("../ufo");```

[quartz kindle]

what does it export?

the router?

[winged linden]

correct

[quartz kindle]

you can do something like this:

// api.js
module.exports = function createRouter(client) {
  router.get(...);
  router.post(...);
  return router;
}

// ufo.js
const createRouter = require("./routes/api");
app.use("/", createRouter(client));

in fact you can move the entire express part into the api.js file

dont need express at all in ufo.js

ufo.js can just do ```js
const api = require("./routes/api");
api.start(client, PORT);

and have a start function do all of the initialization and routing of express

[winged linden]

thats smart!! cool cool thanks alot

[quartz kindle]

if you want to store them to look at them later, then yes

but there are special dbs for that

one that supports append-only writes

such as time-series dbs

you can also let the OS handle that

linux itself for example stores all logs in various append-only log files

you can easily setup something that redirects all stdout into some log file and then just print/console.log stuff away

but a proper time-series database will give you a lot more tools to work with such data

[eternal osprey]

ah i see, thank you!

Shit's bullshit let me cast my thing pls

pls ts

mannn all my bot does is dm ppl who join the server..
shit's annoying

[neon leaf]

@quartz kindle do you like my horrible physic thing https://8000.webhook.pet/

[quartz kindle]

"casting" is for types, types are not values

in js types dont exist, only values

in ts, types exist, but they are removed when compiling

[eternal osprey]

yeah i know, it's sad cuz i always liked the feature

though it's not needed, i can just use a switch right

[lyric mountain]

ngl, this earns an instant block from me

[deft wolf]

[*]

[lyric mountain]

dyno and mee6 are blocked for this very reason

[eternal osprey]

😭

mannn

are they still blocked

cuz idk it;s not harmful right

[lyric mountain]

yes they are

I wont unblock anytime soon

it's not harmful but annoying af

[eternal osprey]

damn

[lyric mountain]

especially mee6 cuz it's everywhere

[eternal osprey]

sorry ig

[deft wolf]

It really depends. I often saw advertisements for "partner" servers when joining a server

[eternal osprey]

yeah that shit is annoying asf indeed

[quartz kindle]

lmfao looks pretty fun hahahah

[eternal osprey]

i am joining a gaming community and out of nowhere i get 18+ servers shoved into my screen

[deft wolf]

fr fr

[eternal osprey]

ion know gang but the physics seems off

[quartz kindle]

one got stuck, cant create more xd

[neon leaf]

ye its very hard to fix when idk shit about physics

[quartz kindle]

xD

[pearl trail]

lmfao me too

[neon leaf]

I may need to redo my amazing js

https://8000.webhook.pet/index.js

(I should just use a library for this)

[quartz kindle]

you probably wanna use requestAnimationFrame()

[neon leaf]

ok now it uses matterjs for physics which seems to work a lot better (who couldve guessed)

[quartz kindle]

is it live?

i just see a dark screen with some settings buttons, but i dont see a way to start/create the blocks

[neon leaf]

right click

[quartz kindle]

ah

they dont move tho?

[neon leaf]

did you set it to static

[quartz kindle]

ah

xD

physics engines always have issues with "resting" position

throw a bunch of objects in a pit and they will keep moving forever and never settle

[neon leaf]

ye

[quartz kindle]

still fun

[neon leaf]

oh god

[quartz kindle]

xD

[neon leaf]

this doesnt seem to work well tbh

oh wait i broke importing oof

nvm

[pearl trail]

nice

[lyric mountain]

I can almost hear source engine's clank sound from the bottom part

[eternal osprey]

Am i a bitch that i secretely fall in love with wordpress?

I used to do all my website coding in html, css etc but damn that was crazy

Wordpress is so easy to manage

[lyric mountain]

that's why non-coders use it

[eternal osprey]

😭

[lyric mountain]

it's easy & cheap to get a website running

[eternal osprey]

yeah

So far ive been creating my sites in plain js/ts and html/css with bootstrap

Its been a pain cant lie

Is react worth it to speed up the process of developing a site?

[lyric mountain]

personally I hate react

first because it's very prone to dependency hell

[eternal osprey]

I see

[lyric mountain]

second because the source quickly becomes unreadable if you lower your guard

[eternal osprey]

I see yeah thats true

Could i basically rely solely upon a css library and plain js/html?

[lyric mountain]

on react? not really

you need to use react in react

[eternal osprey]

No i mean without react

[lyric mountain]

I like doing sites with pure html5 kit

[eternal osprey]

i hear everyone talking about react and thst its a must have etc, but idk i find it bs cuz ive been creating some good sites without

[lyric mountain]

dynamic pages are a bit complicated, but jquery is a thing

[lyric mountain]

but idk i find it bs cuz ive been creating some good sites without
react is the mainstream thing, it by no means mean you cant make good sites without it

it's like python, it's easily the most popular lang

[eternal osprey]

Yeah true

[lyric mountain]

but it doesn't mean anything else is inferior

I'd say use html or the closest-to-html library you can find

the output becomes clean and you get better grip on your changes

[eternal osprey]

Damn you gave me hype. Ill stick with the vanilla shit and bootstrap then

[scenic kelp]

i am a big svelte advocate i love you svelte

i used to be hardcore vanilla css/html

i am like 10x more productive with svelte and tailwind

[lyric mountain]

I'd even ditch bootstrap tbh

I never liked how "pre made" it felt

[eternal osprey]

That shit makes my site responsive tho

Without i am lost

Lost in the sauce

[lyric mountain]

nah, you can easily make responsive sites by using flexbox

bootstrap is more for standardized themes

you can easily make your own stylesheets and have 999x less classes on elements

[eternal osprey]

crazy

yeah that's true

i've been relying way too much on bootstrap

fuck it guys i go in raw

plain html and js

[lyric mountain]

but if u like bootstrap then it's fine too, my reason for avoiding is that it feels too much like theme packages I've used before

it works dandy, until you need to change something

then it's documentation hell to find out how to do it without cascade breaking

[lyric mountain]

btw, intellij has built-in browser preview if u use it

u dont need to keep refreshing to see changes

[eternal osprey]

yeah that's true

[eternal osprey]

overridding is litearlly a cancerous job to do

[quartz kindle]

if you're gonna write your own CSS, check out LESS or even SASS

being able to nest CSS rules makes everything so much easier and smoother

[lyric mountain]

also never forget that instead of ```css
#parent #this {

}

#parent #that {

}

#parent .those {

}
u can just docss
#parent {
#this {

}

#that {

}

.those {

}
}

[quartz kindle]

wait, is that built in now?

i dont remmeber that ever being possible without LESS/SASS

[lyric mountain]

idk, found out when I was making the roulette thing

it was made using pure html + css (+ js later)

no libs or anything

[quartz kindle]

oh dayum

i didnt know they actually made it built in, finally holy shit

[lyric mountain]

ie is dead, so no need to worry abt it anymore

[solemn latch]

Just waiting for jxl to get support 😦

[quartz kindle]

jxl?

[solemn latch]

jpeg-xl

[quartz kindle]

oh

[solemn latch]

chrome had a flag for it, then it was removed.
https://caniuse.com/jpegxl

[quartz kindle]

considering what happened to jpeg2000

i wouldnt count too much on it

[solemn latch]

Safari supports it, chromium just chose avif over it

https://caniuse.com/avif

[lyric mountain]

all I want is anything other than gif on discord

[solemn latch]

I just want sharp to add native support, but they refuse because chrome wont.
But chrome wont, because webdevs dont.

So its a circle of no one adding support because others wont.

[eternal osprey]

what yall think of this logo

imo it's too white

could also do this

[lyric mountain]

what about black?

white aint gonna work as the background is bright already

[quartz kindle]

yeah try black

[lyric mountain]

could even use transparent, then darken the logo if page background is too bright

[eternal osprey]

hmm black let me see

[neon leaf]

i made it more fun and deployed sir @quartz kindle https://physics.rjansen.dev/

[quartz kindle]

xD

[neon leaf]

horrible src is here https://github.com/0x7d8/physics

[neon leaf]

smh abusing the cubes

[quartz kindle]

:^)

[lyric mountain]

if it had sound

[quartz kindle]

xDDDD

[neon leaf]

add it

i dare you

[lyric mountain]

will do it later

oh wait, u mean to the app?

just get the source engine collision sound and add it to onCollide or whatever you use

[neon leaf]

yes

[crystal wigeon]

can i send svgs in embed image?

or just image

[quartz kindle]

just image afaik

[crystal wigeon]

mhm

[crystal wigeon]

ok so is it better to use canvas or svg?

ik canvas is more cpu intensive. im thinking of drawing stuff using svg then load it onto canvas and stream the jpg to embed image

idk if that will be faster because im like drawing shapes and texts and some icons on it

[quartz kindle]

thats gonna be more cpu intensive than using canvas directly, but it depends on what you're drawing

[crystal wigeon]

a couple of rectangles

and couple of images

like a banner

[quartz kindle]

if the end result is gonna be to output an image, going canvas directly is probably preferable

[crystal wigeon]

hmm

i feel like if i were to add some animations canvas will start causing issues

or lag

[quartz kindle]

where do you want to display those animations?

[crystal wigeon]

whereas svg is just simpler to work with no? and it just loads faster

on the canvas

[quartz kindle]

i mean, where on the user's end

[crystal wigeon]

ah on embeds images

[quartz kindle]

on discord?

[crystal wigeon]

yea

the animations will be in the image itself

[quartz kindle]

what animation formats does discord support?

[crystal wigeon]

idk that

[quartz kindle]

gif, mp4, anything else?

[crystal wigeon]

gif

ig

[quartz kindle]

idk if they support apgn or animated webp

[crystal wigeon]

mp4 will be weird

[quartz kindle]

in any case, those are all done better with canvas

[crystal wigeon]

ive seen png images ebing animated and shown on discord

got it

[quartz kindle]

svg animations are done by css, you cant put that into an image

[crystal wigeon]

mm, ok so is there a way for me to add text into a rect on canvas and tell me to automatically take the width of the text?

or do i have to calculate the text width and adjust it

[quartz kindle]

you have to calculate stuff, but canvas does offer a measureText() method

[crystal wigeon]

ic, also canvas positioning is all absolute right?

is there any relative positioning i could use

[quartz kindle]

yes, its all based on XY 2d coordinates

[crystal wigeon]

alr

[quartz kindle]

you'll have to calculate those yourself

[crystal wigeon]

man wish there was an easier way

[quartz kindle]

welp, its a good learning experience

i enjoyed watching my shit slowly turn up good looking with canvas

[crystal wigeon]

yeah. gotta learn some optimal ways to write text and draw on canvas and make it load in less than 50ms

[lyric mountain]

when choosing one or the other, you just need to question yourself:

[quartz kindle]

and the funny bugs i got from miscalculating stuff

[lyric mountain]

do you NEED infinite resolution?

[crystal wigeon]

my only concern is canvas taking too long

[lyric mountain]

(also discord doesnt display svgs)

[quartz kindle]

canvas scales exponentially with resolution, how large do you need your images to be?

[crystal wigeon]

im trying to compress all my images considerably small. but if i allow users to like upload their own image it might be problematic

[quartz kindle]

well not exponentially but like cubically or something, idk

[crystal wigeon]

like showing a banner

[lyric mountain]

it's cubically yeah square

w * h

[quartz kindle]

you can resize user uploads if they are too large

[quartz kindle]

how big is the banner? like 600x200?

[crystal wigeon]

yeah smaller than that

[quartz kindle]

then its fine

you can easily generate those in < 100ms

[crystal wigeon]

yeah well there gonna be 2 images, lets see. also is there way i can like fit this image to have cover object fit into a rect?

like on canvas i have a rect of some 100x200, i want to fit the banner in this box

and have rounded corners

[quartz kindle]

yes, canvas's drawImage method has variables for source cropping/resizing and destination resizing

[quartz kindle]

for rounded corners you will need to use clip paths

[crystal wigeon]

can.t i use roundRect and then add some image as bg?

[quartz kindle]

sure

[crystal wigeon]

how

[quartz kindle]

how what exactly?

[crystal wigeon]

how do i do this

ig need to use clip path

[neon leaf]

TPS: 22 / 240 (9%)
Bodies: 2917, Removed 1513

very small slowdown

[quartz kindle]

canvas has something called "composite operation"

[crystal wigeon]

thanks

[quartz kindle]

https://developer.mozilla.org/en-US/docs/Web/API/CanvasRenderingContext2D/globalCompositeOperation

scroll down to example images

for example

you draw the image you want

then you change mode to destination

then you draw the rectangle on top of the image

otherwise you can also do clip paths, with beginPath(), moveTo(), lineTo(), arc(), closePath() and then clip()

[neon leaf]

@lyric mountain it has been completed

it sounds absolutely horrible

[lyric mountain]

but it SOUNDS

[quartz kindle]

scale the sound pitch with the size of the object

xDDDD

[neon leaf]

nah someone make a pr (ok=

[quartz kindle]

just got a dm from someone

"hi, im a developer and i have a client that wants to use your api, can you help me"
"i asked chatgpt to parse the response your api gives me, but the result is wrong"

developer in 2024

[eternal osprey]

crazy

[quartz kindle]

its not even hard, i literally gave him copy pastable code to parse the api response into the format he wants

and hes still asking chatgpt to do it for him

and getting wrong results

lmfao

"developer" my ass

[solemn latch]

We need a no-ai website like nohello

[real rose]

dontasktoask is also one of my personal favourites

i wish i could send it on Teams to my colleagues

[solemn latch]

lol

[eternal osprey]

what yall think of the little design on the bottom left corner

it's a webpage that explains how we check for downtime.

tryna make a very minimalistic site

(site colors are orange, white and black btw)

maybe black actually loops beter!

[quartz kindle]

in this instance i like the white better

but black doesnt look bad either

[lyric mountain]

I prefer black

feels lighter on the eye

[neon leaf]

same

[proven lantern]

have you ever seen this? should all these people just be banned?

[pearl trail]

well those people are people who flagged by discord/cloudflare because spamming scam links

idk but that's my guess

[quartz kindle]

i have a few of those on mine, i didnt do anything to them

they didnt do anything either (yet)

[spark flint]

It just means they are marked as a spammer

You already see that marker in chat, up to you if you want to ban

They are more than likely already disabled / restricted from chatting in or joining servers

[proven lantern]

i should just kick them. they might spam other members

[sharp geyser]

Uhm I highly doubt that

As I was looking in a server I own, and null had that next to her

[spark flint]

They disable only some of the time

Generally if it’s a new account or if the account shows indicators of compromise

[pearl trail]

i joined 121 servers and i'm surprised there's none that is a spamming server

[lyric mountain]

I mean, u need to kinda look for to be in one

[solemn latch]

👀

[sharp geyser]

Love it

[pearl trail]

yeah fair enough

[feral rose]

Why does on_dbl_test but not on_dbl_vote literally makes no sense

[sharp geyser]

Anyone here ever use grafana with nginx and cloudflare origin certs? I am running into an issue where visiting the page it shows a 400.

grafana.ini

[server]
# Protocol (http, https, h2, socket)
protocol = https

# This is the minimum TLS version allowed. By default, this value is empty. Accepted values are: TLS1.2, TLS1.3. If nothing is set TLS1.2 would be taken
;min_tls_version = ""

# The ip address to bind to, empty will bind to all interfaces
;http_addr =

# The http port  to use
http_port = 3000

# The public facing domain name used to access grafana from a browser
domain = grafana.greenzone101.com

# Redirect to correct domain if host header does not match domain
# Prevents DNS rebinding attacks
enforce_domain = false

# The full public facing url you use in browser, used for redirects and emails
# If you use reverse proxy and sub path specify full url (with sub path)
root_url = https://grafana.greenzone101.com/

# Serve Grafana from subpath specified in `root_url` setting. By default it is set to `false` for compatibility reasons.
serve_from_sub_path = false

# Log web requests
;router_logging = false

# the path relative working path
;static_root_path = public

# enable gzip
;enable_gzip = false

# https certs & key file
cert_file = /etc/ssl/certs/somecert.pem
cert_key = /etc/ssl/certs/somecert.key

nginx

server {
    listen 443 ssl http2;
    listen [::]:443 ssl http2;

    server_name grafana.greenzone101.com;

    ssl_certificate /etc/ssl/certs/somecert.pem;
    ssl_certificate_key /etc/ssl/certs/somecert.key;

    ssl_verify_client on;
    ssl_client_certificate /etc/ssl/certs/cloudflare.crt;

    location / {
      proxy_pass    http://127.0.0.1:3000;
      proxy_http_version 1.1;
      proxy_set_header Upgrade $http_upgrade;
      proxy_set_header Connection 'upgrade';
      proxy_set_header Host $host;
      proxy_cache_bypass $http_upgrade;
    }
}

nginx looks setup properly

it seems to be the grafana.ini file causing problems

[quartz kindle]

remove the certs from grafana.ini and set it to http only

you're not using them at all, you're connecting via http

[tribal bridge]

Since I got my bot added to Top.gg, I've been getting errors with my bot, which forces it to go offline. Been getting an error with webhooks, and also this one here.

      throw new DiscordAPIError(data, "code" in data ? data.code : data.error, status, method, url, requestData);
            ^

DiscordAPIError[50013]: Missing Permissions
    at handleErrors (F:\YannaBot\Bot\node_modules\@discordjs\rest\dist\index.js:730:13)
    at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
    at async SequentialHandler.runRequest (F:\YannaBot\Bot\node_modules\@discordjs\rest\dist\index.js:1133:23)
    at async SequentialHandler.queueRequest (F:\YannaBot\Bot\node_modules\@discordjs\rest\dist\index.js:963:14)
    at async _REST.request (F:\YannaBot\Bot\node_modules\@discordjs\rest\dist\index.js:1278:22)
    at async TextChannel.send (F:\YannaBot\Bot\node_modules\discord.js\src\structures\interfaces\TextBasedChannel.js:177:15) {
  requestBody: {
    files: [],
    json: {
      content: undefined,
      tts: false,
      nonce: undefined,
      enforce_nonce: false,
      embeds: [
        {
          color: 3447003,
          description: '![YannaLevelUp](https://cdn.discordapp.com/emojis/1267046635217879091.webp?size=128 "YannaLevelUp") @coarse belfry, you have reached Level 1!'
        }
      ],
      components: undefined,
      username: undefined,
      avatar_url: undefined,
      allowed_mentions: undefined,
      flags: undefined,
      message_reference: undefined,
      attachments: undefined,
      sticker_ids: undefined,
      thread_name: undefined,
      applied_tags: undefined,
      poll: undefined
    }
  },
  rawError: { message: 'Missing Permissions', code: 50013 },
  code: 50013,
  status: 403,
  method: 'POST',
  url: 'https://discord.com/api/v10/channels/707673693160996895/messages'
}

Node.js v20.14.0```

[quartz kindle]

your bot is trying to send messages to a channel where it has no permissions

the channel ID is 707673693160996895

you can look up this channel in your bot's channel cache, and check which permissions are causing the issue

but ultimately you will need to check permissions before sending the message, and you also need to catch the error if sending fails to prevent the crash

[tribal bridge]

Only thing I can think of is that its within a server thats added my bot

[sharp geyser]

Well I fixed it partially

now its sending this tho

[quartz kindle]

whats your config now?

[sharp geyser]

server {
    listen 443 ssl http2;
    listen [::]:443 ssl http2;

    server_name grafana.greenzone101.com;

    ssl_certificate /etc/ssl/certs/greenzone101.com.pem;
    ssl_certificate_key /etc/ssl/certs/greenzone101.com.key;

    ssl_verify_client on;
    ssl_client_certificate /etc/ssl/certs/cloudflare.crt;

    location / {
      proxy_pass    http://127.0.0.1:3000;
      proxy_http_version 1.1;
      proxy_set_header Upgrade $http_upgrade;
      proxy_set_header Connection 'upgrade';
      proxy_set_header Host $host;
      proxy_cache_bypass $http_upgrade;
    }
}

and in grafana I set http_addr to be 127.0.0.1

since it was assigning to * aka all interfaces

[quartz kindle]

did you remove https from grafana?

[sharp geyser]

no

because I need it to be https

[quartz kindle]

you dont need grafana to be https

[sharp geyser]

oh?

[quartz kindle]

you already have nginx

proxy_pass http://127.0.0.1:3000

nginx connects to grafana via http

[sharp geyser]

Thanks tim

you are a godsend

[quartz kindle]

https is only for outside networks, you dont need https for internal networks

[sharp geyser]

I was under the assumption grafana needed to be https as well

Cause thats what the docs told me

[quartz kindle]

only if you expose it to the outside

and access it without nginx

[sharp geyser]

well thanks a lot tim

<3

[quartz kindle]

np

[quartz kindle]

yes but you still need to do those checks in your code, otherwise your bot will keep crashing

[feral rose]

is someone able to help me in #topgg-api

[frosty gale]

what is grafana

[sharp geyser]

        private async Task<long> GetBalance(string playerId)
        {
            long res = 0;
            _sqlite.Query(new Sql("SELECT balance FROM player WHERE playerId = ?", new object[] { playerId }), _conn, async result =>
            {
                result[0].TryGetValue("balance", out object balance);
                if (balance == null)
                {
                    Debug.LogError($"Economics: Error reading balance: {balance}");
                }
                res = await Task.FromResult(Convert.ToInt64(balance));
            });

            return await Task.FromResult(res);
        }

For some reason this looks wrong

It doesn't seem to error but like

        private async Task<long> GetBalance(string playerId)
        {
            // Set an initial res value to be 0
            long res = 0;

            // Perform an sql query to get the balance based off the playerId
            _sqlite.Query(new Sql("SELECT balance FROM player WHERE playerId = ?", new object[] { playerId }), _conn, result =>
            {
                // Check if result is empty, if so report it. Later make this so it actually reports at some point.
                if (result.IsEmpty())
                {
                    Debug.LogError("Economics: Nothing found for that user.");
                }

                // If result is not empty, balance must exist.
                var balance = result[0]["balance"];

                // Convert balance to be an Int64 aka long
                res = Convert.ToInt64(balance);
            });

            // Return res as a task result
            return await Task.FromResult(res);
        }

[digital swan]

Cool graphs

[ionic schooner]

grafana is beautiful

we love grafana

[eternal osprey]

prepare-statements in sql lets go 🗣️

i see so many sql prijects on github not preparing their statements.

is there a way to preview html files on discord?
Not the actual code inside the html file, but a preview of the actual site the file described.

[lament rock]

Me getting prepared for my statements against sql and how you should totally just use a json DB :)

Oops I forgot my /s

[scenic kelp]

what's the use of Task#FromResult here? not critiquing just literally never seen it

[quartz kindle]

not really no

[winged linden]

Bro, js is weird as fuck, now extra.js and BotStatus have only 1 function easy, exactly the same, one of them works fine once i require a file and the other it doesnt

[quartz kindle]

still having circular dependency issues?

[proven lantern]

i can't believe intel stock just crashed. it only too 2 generations of faulty processors to make it happen

[wheat mesa]

Tbf many chipmakers aren’t doing well rn

[quartz kindle]

its a pretty major deal though, its a huge blow to the company's percieved trust

its like the boeing situation

[proven lantern]

yeah, it's been rough for the silicon industry

[wheat mesa]

My poor AMD shares are down pretty bad rn

Still up from what I bought them at but still

Same with my NVIDIA stock

Never would’ve bought Intel stock though. They’ve been downhill for years

[quartz kindle]

lel

[proven lantern]

they might be a buy in a a year or so

[wheat mesa]

Doubt it

[proven lantern]

they'll be bailed out

[wheat mesa]

People have been saying that for years

Their stock peaked in literally 2000

[proven lantern]

dont they have a new foundry being built now

[wheat mesa]

They haven’t been able to capitalize off of the modern era of chip making and to me that’s a very bad sign that something isn’t right there

[proven lantern]

something isn't right

[wheat mesa]

AMD is doing much better than Intel ever has, I trust them a lot more

[winged linden]

yeah, just fixed it seperated everything lol found out it was a stupid ass mistake from my side

[quartz kindle]

xD

[sharp geyser]

So i can return a value while using async

Ive not used async without using Task so idk how else to do it

[scenic kelp]

ah that makes sense

you absolutely should be using task for async

[frosty gale]

this isnt the norm yet?

immediate critical vulnerability if you dont use them

even sanitizing values before inserting into an sql query is risky

best to tell the db engine whats sql and what isnt

[eternal osprey]

no look at github

like 40% of the projects don't prepare their statements

or sanitize anything

[pearl trail]

https://tenor.com/view/face-mask-guns-redneck-hazmat-gif-16608474

[eternal osprey]

cron.schedule('0 5 * * *', () => {
  scrape();
}, {
  scheduled: true,
  timezone: "America/New_York" 
});

hey guys i am a bit restarted right now.
i want to run the script at 12 am, each day in est time. Would this be configured right/

idk about the whole am pm type thing, i use 24h like any other country

[pearl trail]

https://crontab.cronhub.io/

that means at 5am

[neon leaf]

0 0 * * * should be correct

[eternal osprey]

owh damn

[eternal osprey]

so that's 12 in the morning right?

[neon leaf]

its 0:00 so 12am

[eternal osprey]

wait no i remember 0 0 being at night

ah so 0 12 would be in the morning?

[neon leaf]

i think you are talking about 12pm

yes

[eternal osprey]

ah yeah

thankss!

[deft wolf]

r word

[eternal osprey]

it is tho ngl

shit got me confused asf

[neon leaf]

i dont get why 12pm is 12:00

and 1pm is 13:00 ????

that together is so weird

couldve just done 1-12 = am
13-24 = pm

[pearl trail]

that is called 24h-time format

[neon leaf]

Ye ik

That's normal

[neon flicker]

It throws an error to try to edit the components an interaction reply that already has a few

How may I avoid it?

[eternal osprey]

https://tenor.com/view/misinformation-spreading-spreading-misinformation-online-gif-4543841945782036495

[eternal osprey]

you need to set the hook to the other but then reverse it so that and then you need to reload it.
It's not that hard tbf

[deft wolf]

Real

[neon flicker]

Aw, I just realised that I forgot to use .setCustomId() for the new components

Nevermind, it still throws the same error

rawError: {
    message: 'Invalid Form Body',
    code: 50035,
    errors: {
      components: { '0': { components: { '0': [Object], '1': [Object] } } }
    }
  },

[eternal osprey]

i hjate it when i explain stuff to others but they end up ignoring me

lemme retry again, set the hook from one to the other and then reverse the entity to be able to reload it.

no jk, please show me your code @neon flicker

can't help without any form of code.

[eternal osprey]

i just created a ticket saving bot

what yall think 🗣️

it saves ticket transcripts to maria db, and i then host a private weblink for it for the users to see

only problem is that i need some way to make sure that people can't randomly enter someone else's ticket.

[frosty gale]

sorry you have used the "R" word which can be used to describe mentally challenged individuals in a derogatory way

back in my day it wasnt such a big deal

[frosty gale]

it would be a bit better if the date and time was aligned properly

[eternal osprey]

srry chloe

my sincere fucking apologies 🗣️

https://tenor.com/view/cat-kitten-cat-crying-kitten-crying-05starrynight-gif-10141647709992578610

[wheat mesa]

Because the day starts at 12:00am

Aka 00:00

Then by noon (12pm) it’s 12:00

An hour later it’s 1:00pm/13:00

[scenic kelp]

12pm and 12am are also completely arbitrary

even the US government used 12 am for noon as recently as 2000 smth

[neon flicker]

The issue was that I also forgot to set the styles of the buttons as well 🤦‍♀️

It fixed it eventually to set the styles of the buttons

[sharp geyser]

Hey guys, I recently decided to take a look at valve's RCON protocol and I can't really make heads or tails of it

https://developer.valvesoftware.com/wiki/Source_RCON_Protocol#Basic_Packet_Structure

I am not quite sure what exactly little-endian integers are, and reading the link to it is not exactly very descriptive for me.

I am also not exactly sure how to form a TCP packet to send over

[proven lantern]

AI knows

[sharp geyser]

I see

Makes sense

[proven lantern]

here's a zig example

const std = @import("std");

pub fn main() !void {
    // Create a little-endian integer
    const value: u32 = 0x12345678;

    // Convert host integer to little-endian bytes
    var bytes: [4]u8 = undefined;
    std.mem.writeIntLittle(u32, &bytes, value);

    // Print the bytes
    std.debug.print("Little-endian bytes: ", .{});
    for (bytes) |byte| {
        std.debug.print("{x:0>2} ", .{byte});
    }
    std.debug.print("\n", .{});

    // Read little-endian bytes back to an integer
    const read_value = std.mem.readIntLittle(u32, &bytes);
    std.debug.print("Read value: 0x{x:0>8}\n", .{read_value});

    // Example for RCON-like packet
    var packet: [12]u8 = undefined;
    std.mem.writeIntLittle(u32, packet[0..4], 8); // Size
    std.mem.writeIntLittle(u32, packet[4..8], 1); // ID
    std.mem.writeIntLittle(u32, packet[8..12], 3); // Type (SERVERDATA_AUTH)

    std.debug.print("RCON packet: ", .{});
    for (packet) |byte| {
        std.debug.print("{x:0>2} ", .{byte});
    }
    std.debug.print("\n", .{});
}

[sharp geyser]

Not sure about that for loop at the bottom.

That is zig specific syntax

[proven lantern]

zig doesn't do anything tricky. it's all simple stuff

[sharp geyser]

No idea what .{] or .{byte} is

x:0>2 is also a bit weird

I assume thats like pythons x[0..2]

[proven lantern]

no, it's different

it's like sprintf

[sharp geyser]

Either way x is undefined

[sharp geyser]

never heard of it

[proven lantern]

sprintf (buffer, "%d plus %d is %d", a, b, a+b)

[sharp geyser]

interesting

such a weird and annoying way to do that

[frosty gale]

zig users trying to be unique for absolutely no reason

[proven lantern]

btw writeIntLittle and readIntLittle dont actually exist if you try to run this

[sharp geyser]

100%

[proven lantern]

it is copying C there

that's not unique

[proven lantern]

these are c docs

[radiant kraken]

so real

the whole "hidden control flow" thing doesnt even get me interested

that's literally just C

[proven lantern]

it is literally just figuratively C

[radiant kraken]

Bun uses zig just to look different

[proven lantern]

isn't zig kinda a bad idea for a JS runtime? don't you actually need saftey for that?

i know zig is planning to have memory safety on par with rust someday, but that's not today

[radiant kraken]

well the v8 runtime is in C++ so

[proven lantern]

well it's a step in the right direction

zig having 3 release modes is a good idea though. safe/fast/small

[scenic kelp]

i also don't really think runtimes are very safe in the first place when you're writting JITters

you're literally just injecting assembly code to be run

[frosty gale]

thats not quite how it works

the bytecode doesnt get converted to assembly/machine code, it just gets interpreted by some low level program written in c/c++/rust/whatever that does something depending on that bytecode instruciton

kinda something like:

if (line.instruction.op === "ADD_NUMBER_TO_VARIABLE") {
  variable += line.parameter;
}

unless you have a memory corruption bug its impossible to run arbitrary binary code in a JIT engine from a flaw in the bytecode

nevermind, apparently there is actual compilation involved in most JIT runtimes

im not sure how it works but id imagine it would be a sequence of predefined instructions for the most part so it would still be very difficult to execute any arbitrary binary code using that

[sage bobcat]

One message removed from a suspended account.

[radiant kraken]

because no hidden control flow!!!! 💪

zig fans when operator overloading 🤮

[sage bobcat]

One message removed from a suspended account.

[radiant kraken]

so real

[proven lantern]

it's less complicated and has a faster runtime than rust

[scenic kelp]

yeah that's just a normal interpreter

[scenic kelp]

but this is still hard you have to ensure that no undefined behaviour arises

[quartz kindle]

Bun uses zig for the wrapper, the js engine itself is JSC, webkit's engine

JSC is c++ like V8

[quartz kindle]

just use whatever your language offers for tcp, like some sort of built in tcp/net/socket module

[sharp geyser]

icic

ty

[quartz kindle]

tcp is basically just writing bytes

create tcp, connect to some ip address, once connection is open, use some write function to write bytes

[proven lantern]

so v8 is just like discord.js-light. makes sense

[quartz kindle]

lmfao

8v is like chrome, JSC is like safari

(quite literally, chrome uses v8, safari uses jsc)

[spark frost]

how can i obfuscate my website so noone can see the code from its elements ???

[proven lantern]

what do you mean by code?

[spark frost]

like index.html

if i use elements from dev tool it shows all the correct codes

is there any way to obfuscate them ?

[proven lantern]

is bun on the same level as v8 and JSC?

[proven lantern]

can you give a fake example of a code you want to obfuscate?

[quartz kindle]

Bun is a wrapper around JSC

node is a wrapper around V8

[quartz kindle]

there is no point in trying to do that, html, js and css are by default meant to be exposed

[spark frost]

<form action="action_page.php" method="post">
  <div class="imgcontainer">
    <img src="img_avatar2.png" alt="Avatar" class="avatar">
  </div>

  <div class="container">
    <label for="uname"><b>Username</b></label>
    <input type="text" placeholder="Enter Username" name="uname" required>

    <label for="psw"><b>Password</b></label>
    <input type="password" placeholder="Enter Password" name="psw" required>

    <button type="submit">Login</button>
    <label>
      <input type="checkbox" checked="checked" name="remember"> Remember me
    </label>
  </div>

  <div class="container" style="background-color:#f1f1f1">
    <button type="button" class="cancelbtn">Cancel</button>
    <span class="psw">Forgot <a href="#">password?</a></span>
  </div>
</form>

[quartz kindle]

at most you can use front-end frameworks, they usually have a very different source code from what the resulting code looks like

[spark frost]

ohh but like i saw a website that is using a huge line of codes while which i use it directly shows i used api to get the elements

here is the example of that obfucated website (like how i want)

[quartz kindle]

that is not necessarily obfuscation per se, some front-end frameworks do compile stuff in very weird ways that makes it almost look like obfuscated, but its just the way they work

but you can also google some "html obfuscator" and try using that lol

[spark frost]

oh ok

thanks for info

[quartz kindle]

none of that is very good at preventing people from reading the page tho, anyone with enough time and dedication can find ways to de-obfuscate

[spark frost]

kk so basically you cant hide your code of website ??

[quartz kindle]

thats why you should never put important things like api keys and passwords on the front end

because its impossible to protect

[spark frost]

oh wait yeah 1 more thing like sometimes my api give a little error it shows the link to it can i hide that ??

[quartz kindle]

you cannot hide network requests, they always show up in the dev tools network tab

even if the code is obfuscated, once the request executes, it will show up in the network tab

[spark frost]

ohh

[spark frost]

and how can i get SSL certificate for any url (like where my application which sends api post is on http but i cant use on my website which is https )

so how can i make that http to https ??

[quartz kindle]

you need to add SSL to your application

there are many ways to get free SSL, for example using letsencrypt or acme.sh

there is also cloudflare

[spark frost]

even if that host is not mine (its a vps)

[quartz kindle]

yes

you do need a domain tho

[spark frost]

ok ill buy one or can we get domain for free ??

[quartz kindle]

some are free

you can also get free subdomains

[spark frost]

how ?

[quartz kindle]

https://freedns.afraid.org/

https://www.noip.com/

some examples of services that offer free subdomains (yourwebsite.theirwebsite.com for example)

or you can create a subdomain on the website you already have

[pearl trail]

free domains!!;??!??!?!??

[frosty gale]

people abuse them though which is why they don't tend to last

one bad website can get the entire domain flagged

the domain owner is also on the hook with the registrar and ICANN

[soft laurel]

Are the new SKUs for bot monetization the same as Entitlements?

[frosty gale]

lol

[dense quarry]

how do I get a discord bot token

[proven lantern]

yeah
https://discord.com/developers/docs/monetization/entitlements#list-entitlements

[soft laurel]

How are the differing types mapped? Like, how would I know which SKU/Entitlement the user purchased?

[proven lantern]

look for entitlement_sku_ids

[soft laurel]

Thanks! I see that now.

[proven lantern]

if you are working with consumables it's a little bit different

[soft laurel]

On the discord developer portal you create an application. It'll be under the Bot tab. Be sure to save it somewhere as you won't be able to see it without refreshing it. You should be shown it during or imediately after Bot creation.

[soft laurel]

How so? We're talking about the One Time use purchase?

[soft laurel]

lol sorry, lots of things. I'm also having trouble creating a test SKU. What scope on the Oauth token do I need? I keep getting a 401 unauthorized with body

{
    "sku_id": "my id",
    "owner_id": "my guild",
    "owner_type": 1 // guild subscription (1) user subscription (2)
}

with url https://discord.com/api/v10/applications/< bot-id >/entitlements
with scopes applications.store.update applications.builds.read applications.entitlements applications.commands.update

[proven lantern]

create the skus here.
https://discord.com/developers/applications/<BOT_ID>/skus

[soft laurel]

Even the test skus? Does the entitlements not create the test skus?

I'll give it a try

bit confusing on discords end

[proven lantern]

this endpoint creates a test consumable. you need to create the consumable sku before you can create a test consumable
https://discord.com/developers/docs/monetization/entitlements#create-test-entitlement

create-test-entitlement does not create a sku

[soft laurel]

If I created the SKU in the development portal is that not the same thing?

Like I have a store and everything

The create test entitlement takes a guild subscription or user subscription owner_type which isn't a consumable? Question mark?

i'm not fighting you, just making sure I understand the terminology

Maybe I should clarify and say that I'm creating a test entitlement to test my implementation of checking skus in my bot.

[proven lantern]

yeah, you can make it look like a guild/user purchased a subscription

[soft laurel]

That's the issue i'm having is a 401 on the test entitlement endpoint

was wondering if I was missing a scope in my token request

[proven lantern]

you need to pass an auth header

Authorization: Bot TOKEN

[soft laurel]

I'm sending a bearer token, does it have to be the Bot API Token?

[proven lantern]

yeah, can you even make a bearer token for a bot on discord?

[soft laurel]

I used the client_secret and client_id of the bot to generate the token. Maybe that wasn't doing what I thought it was

worked like a charm. Had to use the Bot < bot token >

[proven lantern]

bearer tokens are used for oauth and oauth is user auth. just use the bot token

[soft laurel]

ty ty

[proven lantern]

nice

[crystal wigeon]

ok so

anyone there

what is the best resolution i should use for a canvas to send on discord as an image?

for example:

https://cdn.discordapp.com/attachments/1004615731431874600/1269258381865254974/profile.jpg?ex=66af687b&is=66ae16fb&hm=d0a4e3b2d38137b5adbd0b777fde28103b08ca59e37df012e3c73381cee8c075&

im using 708x550. but the icons are not legible at all

[pearl trail]

more like the color

bright bg + bright icon =

[crystal wigeon]

even if i use dark bg.

the badges are too small i feel

no?

the text on it is not visible. i tried to like zoom in and all

i need a bigger reso

[lament rock]

@crystal wigeon Your problem is contrast. The icons themselves are very low contrast and your choice of colors isnt the best against the background. You can create a higher contrast while retaining both the background color and the icon colors by having a container for the icons.

Inside the icons, I can see what is supposed to be identifying features for what the icon is, but unless I zoom in, I can't make it out and it isn't a resolution issue as on the client, the images are covered to a container size. The only benefit would be from super sampling at the detriment of lower connection speed download rates and higher transmission size and memory footprint while compositing and sending the image

[crystal wigeon]

so you're saying if i added some dark bg you will be able to identify the icons?

the container bg

lemme try it out

[lament rock]

https://media.discordapp.net/attachments/497161350934560778/1269273425327816788/profile.png?ex=66af767e&is=66ae24fe&hm=6537208784024a4b5ea934cec19560f04e973ea684241505ee4b17def19e3fa1&

https://media.discordapp.net/attachments/497161350934560778/1269273465886605352/profile.png?ex=66af7688&is=66ae2508&hm=de7bf218df1d92f4e813dcace748344462bec733e13dfa42b945c4f25bd9934a&

I have both a dark and light mode and while I don't change the icons themselves, they've been crafted to be high contrast so you can still tell what they are

Dont mind the fact that the text is overflowing

[crystal wigeon]

how do i make them high contrast

in figma

the icon color itself needs to be high contrast right? nothing to do with icon resolution?

i was thinking of increasing the icon reso from 24px to like 32

https://cdn.discordapp.com/attachments/1004615731431874600/1269274341938429992/profile.jpg?ex=66af7759&is=66ae25d9&hm=2c13427b5f25fe57d205f8d052e36df64a9353314d349fa150546916c541e8b5&

the icons are better visible now hmmm. or is it not?

[lament rock]

Colors. Black and white is the highest contrast possible and white text on a black background is the most legible. You need to look into what colors contrast nicely with what colors you already have and deem to be non negotiable like if you have brand colors you arent gonna change those

Slightly, but that can be an issue of exposure/glare on the light theme

[crystal wigeon]

so like i want to allow the user to change the background.

right

if i added a color on the container would that help?

[lament rock]

On these, I'm using a custom background and patrons can do the same. But everything for info is in its own container specifically for contrast which I can control

you cannot control user content contrast

[crystal wigeon]

right, so i put a container bg

for the icons

https://cdn.discordapp.com/attachments/1004615731431874600/1269275663332937729/profile.jpg?ex=66af7894&is=66ae2714&hm=5c56c9ecadbcdf1c75e54547a70a82163a2da2592ba55f30433e9bb5765cc9d0&

like this?

[lament rock]

Yes

https://media.discordapp.net/attachments/475602890430611456/1269275377260560507/profile.png?ex=66af784f&is=66ae26cf&hm=e8dda557b79973975159ea7d4621ef62c72b7560f36df6568788373c9df2c488&

Some more images

[crystal wigeon]

gotcha

now i just need to make the icons itself more light

[lament rock]

I have two profile styles. You can have some control over user background contrasts but it's limited

Still need a container just with a semi transparent background

[crystal wigeon]

gotcha

thanks

[lament rock]

https://media.discordapp.net/attachments/497161350934560778/1269278604232228935/profile.gif?ex=66af7b51&is=66ae29d1&hm=8f0dfdf7a60390b3c20fab4f2c14ad3e7d31259646f750ec937c524f40668386&

The bot's profile

Need to add a custom bg :(

Gif quantization is also highly unstable

[crystal wigeon]

yoooo teach me how to add gifs on canvas

[lament rock]

https://github.com/AmandaDiscord/Amanda/blob/4afb19865cbbe275cdcf604b12a9e74225c5a3a1/packages/runtime-worker/src/commands/money.ts#L966

My method only requires me to build the profile background once and then just add the next frame on top of the previous one. If you need transparency, you need to dispose of the previous frame after adding it to the encoder and rebuild it

nvm I rebuild the whole frame anyways

:3

[crystal wigeon]

pro

[lament rock]

You should probably make use of a worker_threads thread pool since this will exhaust the thread even though it's async

[pine willow]

Good Ping?

[earnest phoenix]

rly bad

[fallow gale]

terrible

[rustic nova]

Me when pinging myself

[warm surge]

Shitty

[sharp geyser]

@rustic nova

[rustic nova]

@sharp geyser

What's your response time

Timed out: 5000ms

[sharp geyser]

2ms

[rustic nova]

Lie

[bitter granite]

Thats so bad

[pine willow]

Thats true

[warm surge]

👀

[rustic nova]

Guessing someone just did rubber duck debugging here

[warm surge]

Bru

[surreal sage]

does this code look broken

[warm surge]

yea send code?

[surreal sage]

whar

[warm surge]

send the code i can fix it

[surreal sage]

  if (message.attachments.size > 0)
    for (const attachment of message.attachments.values())
      if (
        attachment.contentType &&
        !attachment.contentType.startsWith("image/") &&
        !attachment.contentType.startsWith("video/")
      )
        return message.delete();```

[spark flint]

whats the problem with it

it looks fine

[surreal sage]

i don't know

for some reason it doesnt work

[spark flint]

try doing console.log(attachment) before the if

see if its definitely returning the right thing

[surreal sage]

wait

oops

okay

it wasnt broken

i was testing with a psd

[spark flint]

ah

[surreal sage]

not knowing its mimetype starts with image/vnd

[warm surge]

  if (message.attachments.size > 0) {
    for (const attachment of message.attachments.values()) {
      if (
        attachment.contentType &&
        !attachment.contentType.startsWith("image/") &&
        !attachment.contentType.startsWith("video/")
      )
        return message.delete(); 
}
} ```

[spark flint]

webster that looks worse

[surreal sage]

brackets are optional in javascript

[spark flint]

you also dont need { for the if

[warm surge]

Ahh

Im on my phone

[surreal sage]

vs

[spark flint]

if in doubt, js if (message.attachments.size && !message.attachments.every(a => a.contentType.startsWith("image/") || a.contentType.startsWith("video/"))) message.delete();

[surreal sage]

READABLE CODE PLEASE THANK YOU 🙏

[spark flint]

i mean

thats readable

it does the same job, if not faster

tested

const attachments = [
  {
    name: "image.png",
    contentType: "image/png",
    size: 1024
    // whatever
  },
  {
    name: "video.mp4",
    contentType: "video/mp4",
    size: 2048
    // whatever
  },
  {
    name: "file.exe",
    contentType: "application/exe",
    size: 1024
    // whatever
  },
  {
    name: "blahaj.shark",
    contentType: "shark/chomp",
    size: 1234
    // whatever
  },
  {
    name: "fish.png",
    contentType: "image/png",
    size: 1024
    // whatever
  }
];

// test performance
console.log("testing original version");
console.time("original version");
for (const attachment of attachments) {
  if (
    attachment.contentType &&
    !attachment.contentType.startsWith("image/") &&
    !attachment.contentType.startsWith("video/")
  )
    console.log("some attachments are not images or videos");
}
console.timeEnd("original version");

// test performance
console.log("testing short version");
console.time("short version");
if (attachments.some(a => !a.contentType.startsWith("image/") && !a.contentType.startsWith("video/"))) {
  console.log("some attachments are not images or videos");
}
console.timeEnd("short version");```

[quartz kindle]

the first one will keep looping even after it found one

should have a break in there

[quartz kindle]

figured out a hack to detect webrtc audio activity lmao

/** @type {Map<string, RTCRtpReceiver>} */
const ___receivers = new Map();

const ___peer = Object.getOwnPropertyDescriptor(window, "RTCPeerConnection");
Object.defineProperty(window, "RTCPeerConnection", {
    value: function() {
        const p = new ___peer.value(arguments);
        p.addEventListener("track", trackEvent => {
            if(trackEvent.track.kind === "audio" && ["6666", "6667", "6668"].includes(trackEvent.streams[0].id)) {
                ___receivers.set(trackEvent.streams[0].id, trackEvent.transceiver.receiver);
            }
        })
        return p;
    }
});

setInterval(() => {
    for(const receiver of ___receivers.values()) {
        const contributors = receiver.getContributingSources();
        const active = contributors.filter(x => x.source > 100 && x.audioLevel > 0.001);
        for(const source of active) {
            document.dispatchEvent(new CustomEvent('@_@', { detail: { type: "audio", data: source.source } }))
        }
    }
}, 200);

google meet works in very weird ways, instead of having a dedicated audiotrack for each participant, like they do with videotracks, they have 3 fixed audio streams, each handling a certain number of participants

and i found out that the audio activity of each participant is temporarily held on the receiver and accessible via getContributingSources()

RTCRtpReceiver.getContributingSources()
Returns an array that contains an object for each unique CSRC (contributing source) identifier received by the current RTCRtpReceiver in the last ten seconds.

[timber hatch]

does anyone know a responsible way to stop people absolutely ruining my life with bandwidth? creating a UGC web app which allows user to input custom html to be displayed on their page. But obviously don't want them to abuse it, make their page extremely large, then get thousands of visitors to it

[quartz kindle]

character limit, disable js, sanitize html

[frosty gale]

but is this approach blazingly fast?

[quartz kindle]

probably not

but pretty sure its faster than using mutationObserver on the html

and listening to html changes

running receiver.getContributingSources(); 100k times took ~150ms

so its not that slow, should be no problem running it 5x per second

but it does scale with how many people are in the call

[vivid fulcrum]

yeah, i'd guess splitting all of your peers among 3 streams helps them have less load compared to one stream. but i'm confused why would it not be a dedicated stream for each peer

[quartz kindle]

no idea

there already is a dedicated stream for each peer, but it only has a video track, not an audio track

[vivid fulcrum]

cool, they must be buffering the audio instead of directly consuming it

[frosty gale]

you have no friends to test it with

[quartz kindle]

im testing it with myself

yes thats me on cam

[frosty gale]

google tries finding out who is reverse engineering their api and they just see this

[vivid fulcrum]

i literally cannot come up for a reason why they would do that lol. i'm thinking maybe it's an optimization for the client, it would be cheaper to consume already mixed audio streams than each one individually. but wouldn't that point stand for video too

[quartz kindle]

yeah i have no idea why either

[spark flint]

use CF

so it just caches their page

[frosty gale]

bun try not to suggest a proprietary service challenge (hard)

[sharp geyser]

CF is good

[neon leaf]

cf is good

[frosty gale]

@quartz kindle when creating a write stream in node to a file does it flush the data immediately on write?

or does it write to a buffer first then flushes once theres enough data written or its closed?

ideally id want it to flush only after some inactivity on the write stream just so i save some write cycles on the disk

[quartz kindle]

should flush immediatelly

if you want to pause it for a bit, you can cork then uncork

function createInactivityWriteStream(file) {
  const stream = fs.createWriteStream(file);
  let timeout;
  const writer = stream.write;
  stream.write = function(data) {
    if(!stream.corked) {
      stream.cork();
    }
    writer(data);
    clearTimeout(timeout);
    timeout = setTimeout(() => {
      stream.uncork();
    }, 1000)
  }
  return stream;
}
``` no idea if this works xD

[frosty gale]

ah okay, thanks

do you know what the size of the internal buffer is?

id also want it to flush if enough data is written but not sure if node will do that anyways

[quartz kindle]

around 64k i think

[quartz kindle]

not if its corked

[quartz kindle]

highWaterMark <number> Default: 16384

[sharp geyser]

@rustic nova have you ever gotten an net::ERR_CERT_AUTHORITY_INVALID when visiting your mailcow dashboard?

Its saying my certs are invalid but thats def not true

I haven't touched them at all

[frosty gale]

jeez come on node