development

[frosty gale]

actually i wonder if i convert the buffer to a string then reverse the string would it be faster than reversing a second copy of the buffer

[neon leaf]

no

that was bens solution

remember

[frosty gale]

was it

[quartz kindle]

strings dont have a built in reversing method

so you'd have to make your own

[frosty gale]

oh right

[neon leaf]

yeah he split into array, reversed and joined

[quartz kindle]

i also tried reversing a string with a for loop, not much difference

[neon leaf]

chloe: 67.301ms
best_possible_optimizations: 40.922ms
best_possible_optimizations_2: 42.505ms
tim2: 30.67ms

 robert  ~/projects  bun sha1.js 
[33.85ms] chloe
[25.81ms] best_possible_optimizations
[24.78ms] best_possible_optimizations_2
[29.72ms] tim2```

interesting

[frosty gale]

using it makes the speed fluctuate a lot

it fluctuates between 62 and 70ms

[quartz kindle]

bun uses JSC, which basically means slower raw js, and faster non-js built ins

such as c++ addons and internals

[neon leaf]

yeah

suprised your code is so slow compared to the other in bun

[quartz kindle]

:^)

[neon leaf]

robert  ~/projects  deno run sha1.js
error: Uncaught (in promise) ReferenceError: require is not defined
const crypto = require('crypto');
^
at file:///home/robert/projects/sha1.js:1:16

ugh

[quartz kindle]

xDDD

[neon leaf]

not gonna bother

v8 anyway

[quartz kindle]

deno uses ESM only i think

try import from "crypto"

or "node:crypto"

idk

[neon leaf]

robert  ~/projects  deno run sha1.js
chloe: 198ms
best_possible_optimizations: 134ms
best_possible_optimizations_2: 116ms
tim2: 16ms

💀

welp now i know what to not use in prod

chloe: 188ms
best_possible_optimizations: 131ms
best_possible_optimizations_2: 113ms
tim2: 19ms

 robert  ~/projects  bun sha1.js 
[41.48ms] chloe
[31.50ms] best_possible_optimizations
[34.47ms] best_possible_optimizations_2
[37.81ms] tim2

 robert  ~/projects  node sha1.js 
chloe: 69.046ms
best_possible_optimizations: 42.996ms
best_possible_optimizations_2: 43.665ms
tim2: 38.974ms```

[frosty gale]

by reusing a buffer like 0x7 and using buffer.copy into the preallocated buffer i got 20-30ms faster, huge gains

function chloe() {
    const reversedSet = new Set();
    let buffer = Buffer.allocUnsafe(32);
    for (let i = 0; i < entries.length; i++) {
        const entry = entries[i];
        
        entry.bytes_buffer.copy(buffer);
        const reversed = buffer.reverse();

        if (reversedSet.has(decoder.write(reversed))) {
            console.log("Duplicate found!")
        }

        reversedSet.add(decoder.write(entry.bytes_buffer));
    }
}

turns out buffer.from is a bit slow

[neon leaf]

uh that code wont detect properly

use decoder.end

[frosty gale]

oh thats why it fluctuates so much

[quartz kindle]

.set is also as good as .copy

if not slightly better

[neon leaf]

robert  ~/projects  bun sha1.js
[19.79ms] chloe
[27.96ms] best_possible_optimizations
[24.66ms] best_possible_optimizations_2
[21.96ms] tim2

[frosty gale]

buns clearly doing something right

[neon leaf]

with .write

robert  ~/projects  bun sha1.js
[27.10ms] chloe
[29.48ms] best_possible_optimizations
[27.56ms] best_possible_optimizations_2
[32.50ms] tim2

with .end

[frosty gale]

oh no you dont need to call .end on it

[neon leaf]

well you should to avoid allocating tons of memory

your function is using over 5gb ram on 10000000 objects

tims is definitely the best solution

uses barely any ram and is 2s faster

[frosty gale]

tims solution hurts my eyes

[quartz kindle]

mitata, one of the best benchmarking libs i've tried so far

very consistent results, unlike most other benchmarking libs

[neon leaf]

how many entries?

[quartz kindle]

100k

should try 1mil?

[neon leaf]

yes

and 10mil

(if you have 8gb ram free)

[quartz kindle]

10mil too much lmao

[frosty gale]

using binary encoding for some reason with textdecoder i got it down another 20ms to Chloe Function: 34.52ms

no idea why thats faster than ascii but this is js after all

gonna add tims too for comparison

[quartz kindle]

thats 1mil per iter

[neon leaf]

i want to try it on my pc

can you send that reversebuffer.js file

[frosty gale]

@quartz kindle your one runs slower than mine for some reason

Chloe Function: 35.194ms
Tim Function: 57.827ms
Hex Username Man Function: 95.63ms

i used this one

[quartz kindle]

hacks

[frosty gale]

i still dont get why using binary as a key is faster than ascii

ascii is literally just str[i] = buffer[i]

[neon leaf]

no thats binary

[frosty gale]

oh wait

[neon leaf]

ascii has some extra stupid validation afaik

[frosty gale]

its actual binary

'\x01\x02\x03'```

thats why

[neon leaf]

yeah

[frosty gale]

i thought it meant something like 0111010010101

then yeah thats gonna be always faster

[neon leaf]

mmmm

[quartz kindle]

ah yeah, i forgot about that

binary is alias for latin1, not ascii

binary/latin1 is the one that does no checks

[frosty gale]

i dont like this language

[neon leaf]

best_possible_optimizations seems to be 2x slower on both machines compared to tim2

[quartz kindle]

i still dont know if mine is actually reliable/accurate lmao, i did add one duplicate entry to see if it finds it and it did, but im not confident in this rocket science code

[frosty gale]

now your function runs slightly faster than mine lol

fluctuates though

[quartz kindle]

yeah, js fluctiations are normal, cant really deal with them

[neon leaf]

I really need to fix my cooling

I am using an aio with the pump in the radiator but mounted at the top

barely any water going through

[frosty gale]

using set instead of copy for buffer had no impact

yeah this is my end result and its almost as fast as tims, without finding a way to avoid using a set or converting buffer to string dont think theres a way to make it faster

function chloe() {
    const reversedSet = new Set();
    let buffer = Buffer.allocUnsafe(32);
    for (let i = 0; i < entries.length; i++) {
        const entry = entries[i];
        
        entry.bytes_buffer.copy(buffer);
        const reversed = buffer.reverse();

        if (reversedSet.has(decoder.write(reversed))) {
            console.log("Duplicate found!")
        }

        reversedSet.add(decoder.write(entry.bytes_buffer));
        decoder.end();
    }
}

i wonder what happens if i use an object instead of a set

[quartz kindle]

i thnk object will be faster

but im not sure

[frosty gale]

much slower around 30ms more

function chloe() {
    const reversedSet = {};
    let buffer = Buffer.allocUnsafe(32);
    for (let i = 0; i < entries.length; i++) {
        const entry = entries[i];
        
        entry.bytes_buffer.copy(buffer);
        const reversed = buffer.reverse();

        if (decoder.write(reversed) in reversedSet) {
            console.log("Duplicate found!")
        }

        reversedSet[decoder.write(entry.bytes_buffer)] = undefined;
        decoder.end();
    }
}

[quartz kindle]

hmm

what about Map?

should be similar to set i guess

[quartz kindle]

i got this with yours

[frosty gale]

yeah its about 2ms faster than using map

i checked v8 internals and maps and sets are almost identical in implementation

except set doesnt set any value

[quartz kindle]

ye

[frosty gale]

anyhow i guess this concludes our annual "try to optimise this js function and work around javascript quirks for no reason other than fun"

[neon leaf]

but does it

[quartz kindle]

more like monthly

xD

[frosty gale]

unless you can write something more cursed than tims rope function i dont think so

[neon leaf]

help 34 ms/iter (32 ms … 54 ms) 183 ms 189 ms 189 ms

(just kidding

[frosty gale]

wonder how it would perform with native bindings

takes ages to setup a nodegyp project though so i wont bother

[neon leaf]

I have an idea

[quartz kindle]

i dont think the data is large enough to make it faster

you would need to concat all the buffers into a single large buffer, pass it to c++ and then decode everything there in one go

otherwise it wont be worth the overhead

[frosty gale]

ill write a patch for v8 just to implement this single use case into arrays

[quartz kindle]

xD

[frosty gale]

[].has_reverse(x)

[quartz kindle]

native bindings are very slow at going back and forth between c++ and js, sadly

[neon leaf]

ok i was gonna write tims stuff in assemblyscript but i truly dont understand wtf tim wrote to type it out

[quartz kindle]

lmfao

basically

lets say each buffer is like this [34,252,12,534,34,234,53,3463,64,554,2,352,3,23,...]
take first item 34, use it as an array index: [33x empty, something here]

after a few times doing this, you will eventually run into collisions

two buffers that have the same first number

when that happens, you create a new rope node/leaf, for example [33x empty, [251x empty, buffer that first time is 34 and second item is 252, ...], ...]

and slowly you build up a rope like a tree index, something like this:

{
  34: {
    252: {
      12: { ... }
    }
  }
}

which allows you to compare new buffers to old buffers byte by byte without having to compare all bytes

[frosty gale]

ill stick to developing an os kernel

[quartz kindle]

lmao

[neon leaf]

guess how long this takes with assemblyscript (ts to wasm ran in bun)

// The entry file of your WebAssembly module.

export function add(a: u8, b: u8): u8 {
  console.log(a.toString(2))
  return a + b;
}

function toString(data: Uint8Array): string {
  let str = '';

  for (let i = 0; i < data.length; i++) {
    str += String.fromCharCode(data[i]);
  }

  return str;
}

export function chloe(entries: Uint8Array[]): void {
  const reversedSet = new StaticArray<string>(entries.length);
  let buffer = new Uint8Array(32);
  for (let i = 0; i < entries.length; i++) {
      const entry = entries[i];
      
      buffer.set(entry);
      const reversed = buffer.reverse();

      if (reversedSet.includes(toString(reversed))) {
          console.log("Duplicate found!")
      }

      reversedSet[i] = toString(entry);
  }
}```

(this has been running 2m+ so far on 1mil)

[quartz kindle]

Lol

[quartz kindle]

btw this is how the full rope looks like at the end of the function (with a reduced test case of 10 buffers of 16 bytes each)

[neon leaf]

oh god

my 2am brain will not understand that

[past field]

hey tim

[quartz kindle]

here, easier to understand if you convert the arrays to objects

[neon leaf]

oh wait so its a binary tree

well almost

[quartz kindle]

yeah

[neon leaf]

ah okay now I get it

[quartz kindle]

its a rope data structure

its very similar to a tree

v8 uses ropes to handle string concatenation under the hood

[neon leaf]

why cant computers just be magic

[quartz kindle]

they are

but they are evil magic

:^)

[neon leaf]

brb gonna build a quantum computer

[quartz kindle]

idk why

but with objects instead of arrays

it gets 2x faster at 1mil iterations

but stays the same perf at 100k iterations

[neon leaf]

now put chloe0

for (const entry of entries) {
    const reversed = Buffer.from(entry.bytes_buffer).reverse();
    for (const e of entries) {
        if (e.bytes_buffer.compare(reversed) === 0) {
            console.log("Duplicate found!");
            process.exit(1);
        }
    }
}```

[quartz kindle]

cpu goes brrrr

[quartz kindle]

hi

[past field]

how’s it going

[quartz kindle]

still running

[quartz kindle]

good, doing some benchmarking

[past field]

nice💪🏾

[quartz kindle]

i dont think this is going to end any time soon

its a double loop

[past field]

i was gonna ask for a humongous favor

[quartz kindle]

essentially 100k x 100k

[past field]

and i understand if you say no

but

can you look at my game code and maybe give me any suggestions for optimizations or improvements that come to your mind? i can screen record the game to show you how it looks if you’re willing

again, completely understand if you say no

i know you’re a busy man

[quartz kindle]

i mean, i can, but like, if you really wanna fully optimize it, you're gonna need to rewrite the whole thing from ground up

its not that your code is bad

its just that your coding mindset can improve, different ways to think about things

and for that to work, the whole thing needs to start from the beginning

[past field]

ok i gotcha, over time i’m willing to learn! maybe not i’ll wait on the optimization lol

[quartz kindle]

they say that premature optimization is the root of all evil xD

[past field]

lol i want to get to that point!

[quartz kindle]

for starters, avoid using fixed timers

[past field]

wait

are you looking at it rn

[quartz kindle]

always think in terms of "when this ends, start that", never like "run this for x time, after this time run that"

[quartz kindle]

no, just from what i remember from before

[past field]

mind if i dm you the code? kinda scared to post it here again lol

[quartz kindle]

lol ok

but im not gonna look at it too deep rn

[past field]

whenever you can get to it, any feedback at all is appreciated!

[quartz kindle]

love how wildly js fluctuates dependign on what my laptop is doing in the background

exact same run

thats why you gotta run benchmarks a gazillion times, to find a good average

[neon leaf]

(or run on an empty vps/lxc)

[quartz kindle]

ye

ok i think this is the final version of the function, just a few adjustments, got rid of Buffer.isBuffer()

function tim2() {
    const rope = Array(255);
    for(let i = 0; i < entries.length; i++) {
        const buffer = entries[i].bytes_buffer;
        let current = rope;
        for(let n = 0; n < buffer.length; n++) {
            const entry = current[buffer[buffer.length - n - 1]];
            if(entry) {
                if(Array.isArray(entry)) {
                    current = entry;
                    continue;
                }
                if(n < buffer.length -1) {
                    const next = Array(255);
                    next[entry[n+1]] = entry;
                    next[buffer[n+1]] = buffer;
                    current = current[buffer[n]] = next;
                    continue;
                }
                // console.log("duplicate found");
            }
            current[buffer[n]] = buffer;
            break;
        }
    }
}

array is still better than object for integer keys, even though v8 claiming that objects have the same ooptimizations as arrays if all its keys are integers

got a sub 20 using continue instead of else, at 100k iter

might just be luck tho

ok i had enough of this xD

[neon leaf]

btw that assemblycode stuff was still running

I completely forgot

[quartz kindle]

LOL

[neon leaf]

moral of the story: dont use sketchy ts to wasm solutions

[surreal sage]

i need

that silly

pink colored

file share

website

thing

[neon flicker]

Why can't I set this option to "Discord Provided Link"?

I was trying to create a private test bot, and so I had to set the install link to "none" for once, however now I can't set it back to "Discord Provided Link"

But for my public bot, it is a valid option

[deft wolf]

Because private bots cannot have this button on their profile

[trim tartan]

+1

[neon flicker]

It won't let me generate an OAuth2 link for me to invite my bot to the Discord server where I will be doing the tests

[trim tartan]

The "Discord Provided Link" option in the Discord Developer Portal is typically used for redirect URLs that point to a Discord-provided endpoint, like the OAuth2 authorization flow. If you're having trouble setting this option, there are a few possible reasons:

1. Limited Use Case: This option is restricted to specific use cases where Discord provides a link for a particular purpose, such as OAuth2 authentication. If your bot or application doesn't fall under these use cases, this option might not be available.

2. Incorrect Configuration: If your bot's configuration doesn't align with the requirements for using a Discord-provided link, the option might be unavailable. For example, this option is usually relevant when setting up OAuth2 flows where Discord handles the redirection to a specified endpoint.

3. API Type: The option might only be available for certain types of applications or bots. For instance, it may not be applicable to bots that don't use OAuth2 or other specific Discord features that require a provided link.

4. Role or Permissions: Ensure that you have the correct role or permissions in your Discord server or Developer Portal account to modify this setting. If you don't have sufficient permissions, you might be restricted from changing certain options.

5. Region or Account Type: Some features or settings in the Discord Developer Portal may be region-specific or dependent on your account type. If the feature is not available in your region or for your account, you might not be able to select this option.

[deft wolf]

💀

[earnest phoenix]

what is bro doing

[neon flicker]

Apparently it will never let me turn off this option no matter what

So even tho it is currently a public bot with the all intents on, I still can't see the option "discord provided link"

Discord's developer portal is annoying

[deft wolf]

This option also prevents the "Add app" button from being enabled

[neon flicker]

Oh thanks

[radiant kraken]

rewrite it in assembly

js is too slow

[raven nexus]

i try to refix the vote seem it dont send to the webhook did i do anything wrong?

const { WebhookClient } = require('discord.js');
const dayjs = require('dayjs');
require('dotenv').config();

module.exports = (client) => {
  client.on("webhook_vote", async (voteData) => {
    try {
      const { user } = voteData;
      const webhookURL = process.env.WEBHOOKURL;

      // Create a Webhook Client
      const webhookClient = new WebhookClient({ url: webhookURL });

      // Get the current date and time
      const currentDate = dayjs().format('YYYY-MM-DD');
      const currentTime = dayjs().format('HH:mm:ss');

      // Send the message through the webhook
      await webhookClient.send({
        content: `:ballot_box: **Received a vote from ${user}!**`,
        embeds: [
          {
            color: 0x00A36C, // Customize the color as needed
            fields: [
              { name: 'Date', value: currentDate, inline: true },
              { name: 'Time', value: currentTime, inline: true },
            ],
          },
        ],
      });

    } catch (error) {
      console.error("Error handling vote event:", error);
    }
  });
};

WEBHOOKTOKEN=ZZZZZZZZZZZZZZZZZZ
WEBHOOKURL=https://discord.com/api/webhooksXXXXXXXXXXXXXXXXX

is the webhookurl is discord url? or no

[deft wolf]

Nope

[raven nexus]

ah

how to get the url?

[deft wolf]

I mean, it depends in what sense

Because you can use the discord webhook url to send a message but you can't use it to receive a request from top.gg

[raven nexus]

ah

so what url should i use?

it runing as 3001 for port

[quartz kindle]

did you create a custom event for vote?

because client.on("webhook_vote") does not exist

also, that code does not have any webhook server in it

show trhe code where the webhook server is

[raven nexus]

i found this
https://webhook-topgg.com/

[quartz kindle]

if you use that, you dont need any code

[raven nexus]

owhhhhhhhhhh

[quartz kindle]

with that the vote goes directly to discord, not to your bot

which also means you cant do rewards and stuff like that

[raven nexus]

its ok my bot dont have any function for give reward or anything

because it mostly flightsim used

[quartz kindle]

ok so you can delete all the vote code, and use the webhook-topgg url

[lament rock]

Been working on a program to automatically take material properties from a 3D program (in this case Unity) and bake all of the settings the material has into a texture atlas. This includes stuff like tinting the texture, repeat amount and offsets like specified in the material for each texture slot. Also supports normal maps and scaling on that. (I hate math) and the offsets of the texture where the texture overflows the edges to the other corner is what I'm especially proud of (I hate math)

[quartz kindle]

awesome

[lament rock]

Did I mention that I hate math

I also hate how when resizing a canvas, all of the image data is lost

Also needed to create clamp and lerp functions

Need to optimize as well since offset in both directions calls for 4 slices to reconstruct the image meanwhile offsetting in one direction only calls for 2.

Really makes me wonder how things look under the hood of a shader

HLSL makes things really easy

I mean like Unity shader internals btw since you need to sample the image textures with a special function if you want tiling and offsets

[spark flint]

please dont use chatgpt to help others

its inaccurate 99% of the time and theres no point answering if you aren't writing it yourself and can actually back what you're saying

[lyric mountain]

regarding that, you can generate a link without that button, simply click on OAuth and select your desired perms

the button is merely a convenience for others to invite your bot

[sage bobcat]

One message removed from a suspended account.

[sage bobcat]

One message removed from a suspended account.

[surreal sage]

DANKE

[sage bobcat]

One message removed from a suspended account.

[surreal sage]

such bullshit goal

[sage bobcat]

One message removed from a suspended account.

[quartz kindle]

how much does this dude host to have bills that high

storage is so cheap these days

[surreal sage]

yesss

its like

0.01 cent for 10 gig now no?

or am i too far in the future

~6$ per tb

[quartz kindle]

depends where

google/amazon its stupid expensive

the volume storage products at least

the personal 2tb cloud storage is cheap

but you cant really use that for this, is would be stupid slow

[neon leaf]

cloudflare r2 costs 1 cent per 10gb for fast storage

[quartz kindle]

plus processing fees and what not

[neon leaf]

no other fees really

[quartz kindle]

but still very much cheaper than almost everything else

[neon leaf]

class a is pretty hard to get to the free tier limit

[surreal sage]

https://diskprices.com/

34$ for 900gb SAS is pretty good

but i have 14 600gb ones laying around gg

[neon leaf]

sas is pretty annoying

[rigid maple]

https://svelte.dev/repl/440c652622ff44a882a78d34156fa825?version=4.2.18
CSS did not work fully due to repl
I made fullpage scroll snap animation using gsap. I made a sidebar (it appears at the top left in repl, normally it is right on the left of the page). I want to switch between sections using the sidebar, but I can't figure out how to do this. When the section was changed, I also changed the page url hash. Using this, I can understand which section it is in, but I couldn't figure out how to show that section when I go to that hash.

[ruby ermine]

Hii

[lament rock]

Having a bit of trouble offsetting an image. My logic is to get the sides and swap them around but that produces this funny little fella. Any help would be appreciated. Using the HTML5 canvas

const right = partInAtlas.getImageData(shiftAmountX, 0, sizeInAtlas, sizeInAtlas); // to left
const left = partInAtlas.getImageData(0, 0, shiftAmountX, sizeInAtlas); // to right

ctx.putImageData(right, 0, 0);
ctx.putImageData(left, shiftAmountX, 0);

[lyric mountain]

wdym "swap them around"?

like flip the image?

[lament rock]

This is what I want the end result to look similar to where the black line is where the end of the image actually was before. The left side is now on the right and the right side is now on the left of the black line

but for some reason, there's some overdraw

[quartz kindle]

does getimagedata copy the bytes?

or is it just a reference?

[lament rock]

I don't know, but I tried before with another Canvas instance with the same pixel data using drawImage from ctx.canvas

Produced the same results unless the data is also just a reference

The image data is stored in a uint8array

4 * width * height for RGBA

[quartz kindle]

did you try this?

var original = ctx.getImageData(0,0,300,150);

var copiedData = original.data.slice();
var copied = new ImageData(copiedData, original.width, original.height);

[lament rock]

I can try that when I get home from work. Just gonna bookmark this convo and I'll get back in like 7 hours :)

[quartz kindle]

xD

[naive pulsar]

guys for unknon reason when some one vote to bot its send many post not just one post ?

[neon leaf]

Are you returning status 200 correctly

[naive pulsar]

oh this was the err

thanks

[round cove]

Whenever a shard dies / fails to start, what do you guys recommend doing instead of restarting the whole bot over? It's kinda tedius to continue to do that if something dies on startup.

[real rose]

iirc you don't have to restart the whole bot?

I'm not certain of the process myself, but I wouldn't imagine that's the case

do you have respawn true enabled?

[round cove]

To be fair I'm not sure how else to get the shard to retry once its process exits.

Yeah it retries.

Sometimes it just dies at random, but I seldomly ever restart the bot

Waiting for 30 shards to spawn takes so long, I can't imagine being Iara lmao

[quartz kindle]

djs?

[round cove]

yeah 😔

omg hi tim it's been so long

[quartz kindle]

afaik if you kill the child process, the manager should automatically recreate it

[round cove]

Oh hmmmmmmm

[lyric mountain]

Kill the child so the parent can have a new one, got it

[hidden gorge]

Any reason interactionCreate would be firing the ready.js event?

good to know using var instead of const casued this

[quartz kindle]

Lol

[lament rock]

lmfao

[hidden gorge]

i am a dumbass 😭

[lament rock]

This didn't work

[hidden gorge]

Database query result: { wins: 0n, losses: 0n }
Error fetching user stats: Error: No data returned from the database
    at Object.execute (/root/bots/Leon/commands/utility/userstatus.js:28:23)
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async Object.execute (/root/bots/Leon/events/interactionCreate.js:14:17)

const { SlashCommandBuilder, EmbedBuilder } = require('discord.js');

module.exports = {
    data: new SlashCommandBuilder()
        .setName('userstats')
        .setDescription('Shows your win and loss count'),
    async execute(interaction) {
        const userId = interaction.user.id;
        const db = interaction.client.db;

        try {
            const [rows] = await db.query(`
                SELECT
                    COALESCE(
                        (SELECT COUNT(*) FROM rps_games WHERE winner_id = ?), 0
                    ) AS wins,
                    COALESCE(
                        (SELECT COUNT(*) FROM rps_games WHERE 
                            (player1_id = ? AND winner_id != ?) OR 
                            (player2_id = ? AND winner_id != ?)
                        ), 0
                    ) AS losses
            `, [userId, userId, userId, userId, userId]);

            console.log('Database query result:', rows);

            if (!Array.isArray(rows) || rows.length === 0) {
                throw new Error('No data returned from the database');
            }

            const result = rows[0];

            if (typeof result !== 'object' || result === null || typeof result.wins === 'undefined' || typeof result.losses === 'undefined') {
                throw new Error('Unexpected data structure');
            }

            const wins = Number(result.wins);
            const losses = Number(result.losses);

            const embed = new EmbedBuilder()
                .setColor('#00FF00')
                .setTitle('User Statistics')
                .setDescription(`Here are your statistics, ${interaction.user.username}:`)
                .addFields(
                    { name: 'Wins', value: `${wins}`, inline: true },
                    { name: 'Losses', value: `${losses}`, inline: true }
                )
                .setTimestamp();

            await interaction.reply({ embeds: [embed] });
        } catch (error) {
            console.error('Error fetching user stats:', error);
            await interaction.reply({ content: 'There was an error fetching your statistics. Please try again later.', ephemeral: true });
        }
    },
};

It's failing to pull the data for some reason

[frosty gale]

why are your results returning bigints?

that is a performance killer

[lament rock]

@quartz kindle I figured out that param indexes 2 and 3 are for the size of the rectangle not the coordinates on the canvas

Dont need to duplicate the data either

works in all directions, x, y and xy

It also handles numbers greater than 1 and less than 0

[rigid maple]

How do I send the function in onMount to the component? sveltekit

[neon flicker]

I'm back to dashboard my first dashboard development attempt, and I'm facing with a strange issue

quick.db works outside my Next.js app folder, however inside the app folder, it keeps throwing the following error:
Module not found: Can't resolve 'write-file-atomic'

It worked to execute npm quick.db write-file-atomic

[frosty gale]

bro used a clash royale emote 💀

[lament rock]

Gaming.

[neon flicker]

Lol, it expresses my current reaction so

[neon leaf]

holy fuck

amazon is so stupid

[naive pulsar]

Oh this emoji from clash royale this remember me when I was playing it it's gold days

[sharp drift]

free amd for u what a lucky guy

[rigid maple]

😃

[hidden gorge]

What should I be doing?

[dusky idol]

Okay I need some help with verification of my new discord bot, this discord update is quite tragic. It's asking me to do identity verification and when I'm submitting my id it's just rejecting it. What can I do?
This isn't me trying to verify my first bot. I've a bot which got verified 2 years ago and like 4-5 months ago I verified my 2nd bot. When I verified my 2nd bot, it didn't ask for any id. I thought as I'm already verified once, I don't have to do it anymore.
Now after 4-5 months, I'm trying to get my 3rd bot verified. This time the verification page is completely changed for whatever reason and it's asking me plus not accepting the same document. BRUH

[quartz kindle]

they use stripe for the document verification, you can check on the stripe website which documents they accept from your country

maybe they changed something there and stopped accepting certain types

[dusky idol]

I checked that and it's Pan card only, I'm using that

[quartz kindle]

then you need to contact support or something

[frosty gale]

not sure why your db library is doing that but maybe there is a setting to disable using bigints

you probably dont need unlimited number precision

[dusky idol]

I opened a ticket on discords support a hour ago, now waiting for them to answer. Hopefully it helps
Not sure how they can handle it or assist because as you said stripe manages it mostly

[quartz kindle]

there is nothing wrong with your query, its not failing to pull the data, you're just mishandling the variables

rows is an array, [rows] is the first item of the array

[hidden gorge]

Honestly its my first time using MariaDB

[quartz kindle]

its not about the db, its just js

[hidden gorge]

Ah

[quartz kindle]

const rows = ... this is an array, so you can do Array.isArray(rows) and rows.length

const [rows] = ... this is not the array, this is the first item of the array, same thing as rows[0]

[hidden gorge]

Oooo

[frosty gale]

by the way you might want to set this config to true

not sure why the library by default returns the numbers as bigints

[quartz kindle]

probably because he set the columns as int(11)

but idk

[frosty gale]

oh right he did

???

no i think the library just returns all numbers as bigints

ah i think i see

it doesnt actually affect the size of the number but it affects the return precision

so if you have something like int(5) for a 0 value mysql will return 00000

not sure though

[quartz kindle]

this would not affect js tho, unless js converted it to string

[hidden gorge]

Ton of words

[neon leaf]

aggregation/counting is always in bigints by default

[quartz kindle]

INT(5) ZEROFILL with the stored value of 32 will show 00032
INT(5) with the stored value of 32 will show 32
INT with the stored value of 32 will show 32

[frosty gale]

that is so confusing

why is this a feature

[hidden gorge]

So can someone quickly simplify this for me 💀

I don't understand anything

[quartz kindle]

well, it doesnt really matter

numbers or bigints will not make any measurable difference in this case

your particular issue was just the rows / [rows] thing

[hidden gorge]

bigints means big integers right?

[frosty gale]

it depends how the library does it but id imagine since it returns 11 0's it tricks the library into thinking it should return a bigint? i dont know how mysql returns data on the low level though

does look like it does it all by strings though

and im not sure if you can check whether a number string would exceed the max number limit accurately

[quartz kindle]

it wouldnt make sense tho, because max safe integer is 16 digits long

but 0x7 said aggregation/counting is alreays done in bigints

[frosty gale]

so it is

probably that then

not sure why thats a bigint by default though

not much different than returning a number column

from the librarys perspective

all just tables

[quartz kindle]

the db itself has no concept of bigint since its purely a js thing, so the this particular db library must be doing something

[neon leaf]

postgres returns actual bigints

idk for mysql

[frosty gale]

if it returns numbers in binary it can just return the number it has as-is

if its a string it should just convert it to a string

none of these really require any special conversions in this case

[quartz kindle]

the js lib for postgres you mean? or what

[neon leaf]

no the actual db

[quartz kindle]

the actual db doesnt know what a bigint is

it only knows 32bit/64bit ints

[neon leaf]

well yes sure

but the type is called bigint

[quartz kindle]

ah

[frosty gale]

thats just a 64 bit number no?

ints by default are 32

[quartz kindle]

i guess js will always convert them to bigint without checking how big the number is

if the column is defined as int64

[frosty gale]

yeah bigint looks like its just 64 bit numbers

though if you want "bigints" theyre really just strings in js anyways so you might as well store them as strings in the database

best performance that way

[quartz kindle]

cpu-wise maybe, but storage-wise bigint is better

[frosty gale]

how about we just dont use bigints in database and instead store massive numbers like private key prime numbers as blobs/binary in the database

that way we retain blazingly fast performance

[quartz kindle]

fun fact, a js bigint is the most space-efficient method of storing short amounts of binary data in js

[frosty gale]

its js im not even going to try dispute this because its probably true

[quartz kindle]

i measured it in dev tools memory dump

xD

[frosty gale]

this language has you rethink everything you know about optimising code because it doesnt apply to it

[neon leaf]

tim when optimize my ip and subnet parser

[quartz kindle]

too much rocket science for me

[neon leaf]

yes

i hate ipv6

[jovial palm]

is there any qrcode module recommand to generate ?

[prime cliff]

https://davidshimjs.github.io/qrcodejs/

[jovial palm]

it give the undefined error

[prime cliff]

Oh you wanted a qrcode to show on discord...

That was for a website

[jovial palm]

yea

i want it like i scan the qr code

and let it open the strip payment page

[prime cliff]

@jovial palm use this https://github.com/ushelp/EasyQRCodeJS-NodeJS
You can output to a base64 or image stream that you can upload as a file

If you want to set an embed image/thumbnail with the attached file you can set the embed image url to attachment://file.png for example if you upload it as file.png

[timber hatch]

Does anyone kow the best way to check for NSFW content in images using JS which balances out resource intensiveness and cost.

Because i could use NSFWJS but it's extremely resource intensive, or i could outsource it to a micro-service but that would become extremely expensive?

The use case is an emoji listing site which gets around 3000 new listings per day.

We also allow people to change the content of their emoji, in case they make revisions to the graphic. And we don't know whether it would be more appropriate to re-check on every edit OR re check on an interval.

[deft wolf]

You can simply add emote verification for accounts that are new/untrusted and you can also add an emote reporting option

[quartz kindle]

you could also limit/downscale the image size before checking, if you dont already to that

[frosty gale]

i love how they have a camera option for testing

[earnest phoenix]

https://tenor.com/view/rock-sus-gif-21568224

[lament rock]

@quartz kindle https://github.com/PapiOphidian/MaterialList2Atlas if you were curious at all about my struggles for the past few days. It's considered complete

[quartz kindle]

reads 2 lines from readme
10/10

[lament rock]

Thank you

[surreal sage]

hey chat

[earnest phoenix]

bye chat

[surreal sage]

i'm working on vehicle culling and i'm not sure what if statements and stuff to do to make it work properly

i'm updating the "state" of a vehicle on every frame render

the data I have is:

• is the vehicle within the player camera fov
• the distance to the vehicle from the camera
• if the vehicle is blocked by terrain or not (and so the camera cant see it)

and I can either hide the vehicle all together, or just hide the chassis (chassis is e.g. wheels)

I'm just not sure what order of things I have to do

[lyric mountain]

There's a tradeoff between correctness and resource cost

You'll need to sacrifice one or the other

[frosty gale]

either less boobies or less images

in brainrot terms

[rigid maple]

I'm trying to make fullscreen snap scroll animation using fullpage.js, but all sections appear on a single page.
(I'm trying to rewrite a project I got from github using svelte)

The height of all sections is 100vh

[young pawn]

does anyone know how i would apply classes / attributes to the body / html further down in the component tree in nextjs app router without using vanilla js to modify the dom since that's frowned upon (e.g. document.body)?

so i can provide theme previewing, rtl, and ltr support etc.

[spark flint]

tbf, i dont think nsfwjs is that intensive

[sharp geyser]

Ima be real, don't bother checking

If you really dont want it on there, have a report system

[spark flint]

An example of what I made for a Roblox game:
https://github.com/CWRBLX/asset-scanner

[sharp geyser]

Trying to guard against nsfw images is more work then its worth

[spark flint]

it can scan many a second and its running on the same machine as the website itself

[sharp geyser]

and often times the results vary in accuracy

[timber hatch]

It's not me so much as the registrar, even with a report system if some manage to slip by... domain registrars hate it.

In regards to the inaccuracy i think i will mark any with a score higher than 0.5 as awaiting approval by moderator. i would never STRAIGHT remove from an image model, only make the review process easier.

I would PROBABLY dictate a risk score PRIOR to making the image check too. i.e. emoji name, file name, etc. those cold be great indicators as to the content of the image and the likelyhood of it needing to be checked

[sharp geyser]

Why would your domain registratar care?

[timber hatch]

well we're on namecheap now but when we were on... godaddy 🤢 they would always be on our case

[spark flint]

because they have to step in for stuff

namecheap is good and gives 24h notice most of the time

[sharp geyser]

I mean, if its illegal content sure

but nsfw stuff isn't illegal 🤷

[spark flint]

godaddy is trigger happy for anything other than phishing lol

[timber hatch]

lol

[sharp geyser]

I honestly just don't care enough. You can't expect me to guard against everything perfectly

its better to just let it happen, and handle it as it gets noticed. As even with an AI model it is not perfect if you dont have enough data

[hidden gorge]

var { Events } = require("discord.js");

module.exports = {
    name: Events.GuildCreate,
    async execute(guild) {
        console.log()
    }
}

is there a way i could call the client from the guild to access the db or no?

[deft wolf]

Have you tried guild.client yet?

[hidden gorge]

yes returns undefined

found a way

[wheat mesa]

You can just pass your client as an argument to the function wherever you set up your event handler

That would be the “proper” way to do things

You could technically make your client a singleton if you wanted but that could be a little messy

[quartz kindle]

there is no reason for why guild.client would not work

[deft wolf]

✨ magic ✨

[quartz kindle]

unless guild is not a guild

[warm surge]

Why cant you add client like (client, guild)

[hidden gorge]

I was lazy and did this: event.execute(...args, client);

[warm surge]

Bruh

[hidden gorge]

WHAT IT WORKS

[quartz kindle]

thats not a good idea

[hidden gorge]

also u just muted me for 23 days thats a dick move

[warm surge]

I do it

[quartz kindle]

because different events can have different number of arguments

[warm surge]

Ill make it 24

[hidden gorge]

no

[warm surge]

it work it works

[quartz kindle]

if you do ...args, client
if event has 1 arg, it will be arg1, client
if event has 2 args, it will be arg1, arg2, client
so client is not always in the same position

so you have to remember that in your event handlers

[hidden gorge]

do not do that

[frosty gale]

bro thinks hes him

[warm surge]

Im definitely him

https://tenor.com/view/sponge-bob-patrick-star-laughing-lol-trying-not-to-laugh-gif-15636797

[harsh nova]

bro has learnt the art of mod abusing fron the best

[hidden gorge]

he muted me once for nothing

[warm surge]

I been doing it

[night lagoon]

Use for of loop

Or for in

[proven lantern]

what's a good library for making graphs? is D3 still good?

[lyric mountain]

In what lang?

[quartz kindle]

english

[proven lantern]

JavaScript in the browser or in nodejs building it serverside inside a sveltekit serverless function

like building a svg

im looking at d3-node and d3

[night lagoon]

A nice way to log

client.on("ready", (client) => console.log(client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.users.client.user.id))

[sharp geyser]

[earnest phoenix]

does it actually take 1-2 weeks to get my bot verified

on top gg

[night lagoon]

nope

i got it in 2 days

[earnest phoenix]

epic

mine is a user install

does that affect time

[night lagoon]

I don't think so

[lament rock]

Does anyone know how to convert an HDR color to SDR
For instance, I have a color of RGBI(191, 79, 6, 2.353308);
The color on the left is RGB(191, 79, 6); and the one on the right is the output target color which is RGB(255, 255, 93);

Using JS, but any programmatical way to do it without relying on some external deps/magic would be very appreciated.
I can understand other languages to a certain degree so don't hold back

[proven lantern]

ai says to do this

function hdrToSdr(r, g, b, intensity) {
    // Normalize RGB values to 0-1 range
    r = r / 255;
    g = g / 255;
    b = b / 255;
    
    // Apply intensity
    r *= intensity;
    g *= intensity;
    b *= intensity;
    
    // Apply a simple tone mapping (Reinhard operator)
    r = r / (1 + r);
    g = g / (1 + g);
    b = b / (1 + b);
    
    // Convert back to 0-255 range
    r = Math.round(r * 255);
    g = Math.round(g * 255);
    b = Math.round(b * 255);
    
    return [r, g, b];
}

// Example usage
const hdrColor = [191, 79, 6];
const intensity = 2.353308;
const sdrColor = hdrToSdr(...hdrColor, intensity);

console.log(sdrColor);```

[earnest phoenix]

never use AI to code

please 🙏

you need to make mistakes to learn

[proven lantern]

here's some more AI insight

Note that this is a basic implementation and may not give you exactly the output you're looking for (RGB(255, 255, 93)). The exact conversion depends on various factors, including the color space, the tone mapping operator used, and any additional color grading applied.
For more accurate results, you might need to consider:

The color space of your input (e.g., linear RGB, sRGB, etc.).
The target color space for your output.
A more sophisticated tone mapping operator.
Color grading operations to achieve the desired look.

If you need a more precise conversion, you might want to look into color management libraries or implement a more complex tone mapping algorithm. However, these often involve external dependencies or more complex mathematics.

[earnest phoenix]

what are you trying to do @proven lantern

[proven lantern]

help Ophidian

[earnest phoenix]

well dont use AI because its 99% wrong

most the time

[sharp geyser]

I mean

its not wrong in this case from what I remember

[earnest phoenix]

yall cant be serious

😭

[sharp geyser]

[earnest phoenix]

yall dont use ai and call yourselves developers right?

[sharp geyser]

I think you have a misconception

Nothing wrong with using AI to get a direction to go in

[earnest phoenix]

i get that, just dont rely on AI to do all your coding or you will never learn

[warm surge]

Same

[sharp geyser]

At this point it is impossible not to use AI

[earnest phoenix]

i dont

[sharp geyser]

Its literally built into every browser now

[warm surge]

#general you got question to answer

[earnest phoenix]

lets be friends fr 🔥

[sharp geyser]

I use AI to explain vauge errors

Its better for it to pull sources that I can fact check myself then to scour for them

I hardly ever ask AI for code, it will get it wrong 70% of the time

[warm surge]

If it works it works

[sharp geyser]

Yea, well your bot works, does not mean its good

[sharp geyser]

yes

that will evaluate ofc

djs everything has the client attached to it

guild object, channel object, user object, role object

if it doesn't, it surely will have something that does

[warm surge]

true

[prime cliff]

There's a big difference between using AI to suggest or auto complete stuff (in the sense of completing things like next number in a sequence or last used variables based on context)
Compared to writing the entire code or generating images for you.

Pretty much a lot of big developer tools/services, even Discord, has used some level AI techniques before the whole "AI is supreme" era came in.

---

---

Even stuff like message reply features which some social apps, keyboards and android os have had for a long time to suggest what to reply back with "hi" "ok" based on context

Meanwhile apple is flaunting message reply features in iphones like it's some kind of crazy new technology

[proven lantern]

here's a better version

// Adjustable parameters
let exposure = 1.0;
let contrast = 1.0;
let saturation = 1.0;
let highlightCompression = 1.0;
let shadowLifting = 0.0;

function linearToSRGB(x) {
    if (x <= 0.0031308) return x * 12.92;
    return 1.055 * Math.pow(x, 1/2.4) - 0.055;
}

function sRGBToLinear(x) {
    if (x <= 0.04045) return x / 12.92;
    return Math.pow((x + 0.055) / 1.055, 2.4);
}

function hdrToSdr(r, g, b, intensity) {
    // Convert to linear space
    r = sRGBToLinear(r / 255);
    g = sRGBToLinear(g / 255);
    b = sRGBToLinear(b / 255);
    
    // Apply intensity and exposure
    r *= intensity * exposure;
    g *= intensity * exposure;
    b *= intensity * exposure;
    
    // Apply contrast
    r = Math.pow(r, contrast);
    g = Math.pow(g, contrast);
    b = Math.pow(b, contrast);
    
    // Apply ACES-inspired tone mapping
    const tonemap = (x) => {
        const a = 2.51;
        const b = 0.03;
        const c = 2.43;
        const d = 0.59;
        const e = 0.14;
        return (x * (a * x + b)) / (x * (c * x + d) + e);
    };
    
    r = tonemap(r * highlightCompression) + shadowLifting;
    g = tonemap(g * highlightCompression) + shadowLifting;
    b = tonemap(b * highlightCompression) + shadowLifting;
    
    // Apply saturation
    const luminance = 0.2126 * r + 0.7152 * g + 0.0722 * b;
    r = luminance + saturation * (r - luminance);
    g = luminance + saturation * (g - luminance);
    b = luminance + saturation * (b - luminance);
    
    // Clamp values
    r = Math.max(0, Math.min(1, r));
    g = Math.max(0, Math.min(1, g));
    b = Math.max(0, Math.min(1, b));
    
    // Convert back to sRGB space
    r = linearToSRGB(r);
    g = linearToSRGB(g);
    b = linearToSRGB(b);
    
    // Convert to 0-255 range
    r = Math.round(r * 255);
    g = Math.round(g * 255);
    b = Math.round(b * 255);
    
    return [r, g, b];
}

[eternal osprey]

this is a vps right?

like tf is a kvm

this shit is so fucking cheap for a vps

def changing my vps

[radiant kraken]

what are u making

[proven lantern]

im not making this

idk what it's for

[radiant kraken]

what

[proven lantern]

PapiOphidian asked how to conver hdr to sdr

[pearl trail]

ye

i've been happy with it

12gb for < $5

tho just dont trust that "turbo" clock speed

it's running at base speed

[proven lantern]

the host system has a bunch of VMs running on it right?

[pearl trail]

ofc

[proven lantern]

so you dont get all that processing power

[pearl trail]

all other vps providers are like that too

unless you buy the "dedicated vps"

which is a lot expensive

[proven lantern]

aws uses vCPU as a unit of measurement for processing power

oh, it says 2 vCores

[radiant kraken]

sheeeesh

[civic scroll]

that will cost you 30 vBucks

[sharp geyser]

kernal-based virtual machine

https://www.redhat.com/en/topics/virtualization/what-is-KVM

[ashen rivet]

i want change avatar bot in top.gg

[eternal osprey]

Which one are you using

[eternal osprey]

I just bought this for 5$

Seems like a good deal to me, and i can always upgrade to the 8gb versions soon

I am using galaxygate now, 4gb ram, 1.4ghz cpu, no ddos and 30gb storage. Overprices as FUCK

Will directly cancel them

[deft wolf]

This one probably (or the second version with 15TB traffic)

https://datalix.eu/2024-sale
-# pls don't ban

[eternal osprey]

Owh damn

[eternal osprey]

Owh damn i bought this one but i can ofc still switch

Whatchu think

[deft wolf]

I personally use the one with 12GB of ram for my bot

[eternal osprey]

I see, i do find my vps maximized a lot at 4gb ram so i might buy the 12 one

[civic scroll]

5dollar????

[late crest]

tell me how is it going? great?

[neon leaf]

Datalix is good.

[late crest]

are you using it too?

[neon leaf]

Yes, I recommended it for everyone here

I've been using it for over 1 year

[late crest]

This looks like a gold for me. So cheap and pretty powerful for this price

[neon leaf]

It's not thaaat good pricing

It's definitely very good though

[late crest]

nah it's probably better than some 'Discord Bot Hosting' sites

[neon leaf]

Well yes but noone can compete with hetzner dedicated servers

50€ for i9-9900k 2x1tb 128gb ram

[late crest]

this sounds also very cheap

for ds

[neon leaf]

That is the cheapest in the entire world

So yes

[late crest]

Thanks, I'll add it to bookmarks

[eternal osprey]

Datalix is the best

I just got this one for 5$!!

[eternal osprey]

This one lmao

But i upgraded the ram to 16gb

For 7$

Honestly datalix so far has been great. Their dashboard etc is also 10x better than galaxygate

[late crest]

guys, how long does it took to get your discord bot verified here?

[harsh nova]

-upto

[gilded plank]

@late crest

When will my bot be reviewed?

Currently our average bot reviewal time is a few days or more.

Because of this — and because some bots take longer to review than others due to their features — we can't guarantee your bot will be reviewed as quickly as someone else's in the past and we also can't guarantee your bot will be reviewed within that timeframe. There is no exact time for how long bot approval can take. There is no way to check your bot's position in our reviewal queue, but remember you're not first and you're not last!

You may edit your bot's page as much as you like both before and after it's reviewed and this will have no impact on its place in queue.

You can read more about our bot reviewal process in this support article: How the Bot Reviewal Process Works.

In the meantime, please make sure your bot follows all of our Bot Guidelines for a quick and smooth approval!

[late crest]

nah, I was just asking how long it took for others

[real rose]

#logs and #mod-logs in that case

[frosty gale]

ddr3 in 2024 is diabolical

[lament rock]

I got a server PC with 8GB of DDR2 running Windows 11

Server PC as in 2 CPUs

Dont ask why I put Windows 11 on it

[outer tide]

e

[late crest]

💀 why?

isn't like DDR2 too slow for Win11? Also what CPUs does it have?

[pearl trail]

trimmed debloated win 11 maybe

[late crest]

I hope so. Debloated win11 is good

[quartz kindle]

win11 LTSC ftw

[neon flicker]

Is there a limit for slash command option choices?

[deft wolf]

I think the limit is 25

[deft wolf]

So if I get scammed can I blame you?

[sharp geyser]

if you get scammed by datalix, its a reseller not datalix itself

datalix allows reselling and even promotes it

[quartz kindle]

welcome back lol

[sharp geyser]

sup tim

[quartz kindle]

brb gonna make my own vps business with datalix machines

[sharp geyser]

go for it

they really do not care I dont think

[quartz kindle]

building a vps business isnt a bad idea tbh, but you will be paying from your own pocket for quite a while until you have enough clients

i dont have that much pocket money rn

otherwise i would totally give it a try

[deft wolf]

Tim's VPSs

We need a catchy name and it might work out

[quartz kindle]

"my fucking VPS"

per-user disk encryption so datalix cant see any thing my users put in it

[sharp geyser]

well

they are still datalix's servers

I dont think they'd like having encrypted drives

💀

[quartz kindle]

well if i rent a dedi, do they have a say in it?

[sharp geyser]

I mean

you are renting it

you dont own it

If you stop paying for it

it goes back to them

[quartz kindle]

sure, they can just reformat the disks

but they cant have my data

[sharp geyser]

fair enough

doubt theyd take it anyways

[frosty gale]

how do i upgrade

[late crest]

I think it is still in insider preview program

and you will need Win 10 LTSC to upgrade

[frosty gale]

i would be interested in changing my windows 11 pro installation to the ltsc version

because im sick of microsoft adding more bloatware and annoyances each update

and more things that are opt out by default instead of opt in

but if you cant i might just reinstall it from scratch

[quartz kindle]

idk if you can, clean install is usually the way

doubt windows would let you "upgrade" into something that removes a ton of shit

[late crest]

yup the best way

[late crest]

ltsc still has many unneeded services and telemetry, but it it much better than stock

[prime cliff]

Hell yea

[sharp geyser]

good job on whatever that is

[prime cliff]

A self-hostable dashboard for server/website/project/logging management and other dev related tools like image manipulation/convert, markdown editor, network tools like dns 🙂

[craggy pine]

Fancy seeing you here Builder. The dash looks nice.

[prime cliff]

Thanks and yea gonna use this to replace sentry and portainer when it's fully functional 🙂

[late crest]

Does anyone know why is it empty and why this exists?

[prime cliff]

a few of those pages have been like that for a while now

[late crest]

There are problems with some graphs too

like votes and invites

[prime cliff]

Vote credits got removed too a long time ago

[sharp geyser]

They are planning on doing a lot of that when they release the new site stuff afaik

[prime cliff]

🙃 Ever single time you navigate it just forgets everything and loads it all again, such a waste

Why are the navigation links also missing on loading...

[sharp geyser]

[late crest]

looks sad

[prime cliff]

Lol no wonder why that's broken as hell, the CDN service they use for those flags is just gone with ssl errors and http redirects to some sketchy website

[deft wolf]

It's been like this for months now

Credits for votes were removed about half a year ago and I still see people who weren't aware of it

#auctions-support message

[late crest]

Is there any way to show server count like it is on other bots? Or it will show after some time auto?>

[deft wolf]

https://docs.top.gg/docs/API/bot#post-stats

You can also use some autoposter depending on what language you use

[past field]

any tips for my relationship calculator?

https://pastes.dev/7XIIsoabva

[sharp geyser]

does it work?

[past field]

eh

not as i was expecting

[sharp geyser]

so its not drawing the second avatar?

[past field]

correct

but the logs say it is

wait

[sharp geyser]

I was reading that oop

[past field]

that's my bbygirl u cant seethat

[sharp geyser]

💀

[past field]

ima put it back lol

[sharp geyser]

rip people who can use snipe / have modded clients

[past field]

User 1: Maki https://cdn.discordapp.com/avatars/563434444321587202/3bedfc513ab9272409513297aa9265ea.png
User 2: bigtex3328 https://cdn.discordapp.com/avatars/197538444631605259/4d6dbf87e8fe581ab7ef300d6fd466c9.png
Loading avatars...
Avatars loaded successfully.
Canvas created. Drawing background...
Drawing first avatar...
First avatar drawn.
Resetting clipping region...
Drawing second avatar...
Second avatar drawn.
Canvas converted to buffer.
Embed created. Sending reply...
Reply sent successfully.

[sharp geyser]

It might honestly be drawing it out of bounds

what happens if you copy & paste that image in the embed

Discord limits the size of images in embeds and iirc has been known to "clip" images cutting stuff off

[past field]

nvm i fixed it

i changed the pixels to 350x175

and since i made it smaller i had to reposition it a little bit

i need to make the font a little smaller though

[sharp geyser]

yuh

so

_payload should be the json data

[sharp geyser]

Why

Can you detail your end result please

at this point in time, I am confused on what you want to accomplish

ok

so you need stuff from both the header & body correct?

So

you 100% can

Had to find a source, but you can do smth like this
https://github.com/arlyon/async-stripe/blob/master/examples/webhook-rocket.rs

It details how you can get something from the header as well as something from the body of a request

Assuming you are still using rocket

No problemo

I like axum tho

because its not as annoying

It has extractors built in that you can use that does all the impl stuff for you

https://docs.rs/axum/latest/axum/extract/index.html#common-extractors

In case you ever look at using axum

you'd literally just do

fn some_handler(headers: HeaderMap, Json(payload): Json<SomeStruct>)```

axum is nice

it has a concept you are familiar with I assume

middlewares

Its similar to express middlewares in a way

I swapped from using rocket to axum

and plan on rewriting my backend to my old project into it when I get off my ass and do it

💀

either that or I will use asp.net core

Going to flip a coin to decide tbh

[neon flicker]

Is there a character limit for modal text input labels?

[deft wolf]

I think the limit is 45 characters (at least that's what I understand from the documentation)

[sharp geyser]

yes

[neon flicker]

Thank you 😊

[sharp geyser]

45 characters

God damnit NyNu

😭

Discord didn't register your messages until after I tabbed back in and hit enter

[sharp geyser]

Does not need to be verified

[past field]

can i take it off?

[sharp geyser]

make the bot private

[past field]

ohhh

@pine willow 👀👀

[pine willow]

What

[past field]

i saw you joined my server and left lol

[pine willow]

Yes

Joined, added bot, left

Click war doesnt work single Player

Bs

[past field]

LOL

[past field]

you should try it with your server!

[pine willow]

Funny

My Server has 120 Members and 0 Messages in 3 weeks

Nothing can save it

[past field]

oh damn

come join us for a game lol

[sharp geyser]

Unable to create a 'DbContext' of type 'RuntimeType'. The exception 'Method 'GetDatabaseLock' in type 'Npgsql.EntityFrameworkCore.PostgreSQL.Migrations.Internal.NpgsqlHistoryRepository' from assembly 'Npgsql.EntityFrameworkCore.PostgreSQL, Version=9.0.0.0, Culture=neutral, PublicKeyToken=5d8b90d52f46fda7' does not have an implementation.' was thrown while attempting to create an instance.

What in the flying fuck does this mean

[proven lantern]

Did your dependencies auto update again?

9.0.0.0

seems pretty new

[proven lantern]

https://www.nuget.org/packages/Npgsql.EntityFrameworkCore.PostgreSQL#versions-body-tab

[late crest]

datalix for bot 1v core enough? On free hosting I have 50% cpu limit

Im thinking about getting ryzen one, since it is faster in single/multi-thread than xeon. And I dont need too much ram for my bot.

[lament rock]

1 vcore is not enough for a lot of applications especially at server cpu frequencies

If your app is small then yes

[sharp geyser]

depends on what you are running

also depends on how powerful that one core is

datalix has some pretty beefy cpus tho

[late crest]

ryzen core is pretty powerful even on single thread

[sharp geyser]

at least in terms of their ryzen

[late crest]

Right now it is very small, around 43 servers in. And usage is pretty low, but I am planning to advertise it.

[sharp geyser]

Will it be enough to run a discord bot? Sure, probably not the most efficient though

It also depends on what your bot is doing as well

[late crest]

I just want to get rid of "Application is not responding" on first try, and then work on second try. On free hosting there is 1 core limited to 50%, 512mb ram, 1gb storage

[late crest]

umm the thoughest function is combining pictures. Avatar and Other picture which takes time and resources.

[earnest phoenix]

oracle free tier is love

[late crest]

commands are pretty light

and optimized (I think)

[earnest phoenix]

unless they delete it out from the existance because yes

[late crest]

I tried to register, but it does not accept any of my credit cards to verify 🙂

same as AWS

[earnest phoenix]

worked for me and got a x86 2 vcpu 1gb ram vm

i couldve gotten that juicy 4vcpu 24gb ram arm server though but

getting that is hard

its out of capacity all the time

so you either break the tos or be lucky

[late crest]

how long are you working with oracle free tier?

[earnest phoenix]

its been a few months since i set it up

it now has a purpose aka running analytics server for my site

[late crest]

ah I see

[late crest]

will I be able to upgrade vcpu there in Datalix?

without recreating VM

[lament rock]

Idk. I don't use datalix. You can look at how the price to performance is and scale based on your current needs

[pearl trail]

yes

you can upgrade, but you cant downgrade

[late crest]

ok, great

I will give it a try

[frosty gale]

oracle free tier has extremely poor fraud detection

i once had my free trial registrations denied consistently and oracle support wouldnt tell me why

i once tried registering on another wifi connection and it worked, flagged my ISP apparently

[late crest]

I heard many stories like that

about oracle

[frosty gale]

they make you out like some criminal when you try contacting them asking why

"we are not allowed to give any details"

[late crest]

thats crazy customer support

did you tried AWS free tiers tho?

[frosty gale]

yeah its okay

give you a reasonable amount of free stuff but most of it is the limited first 3 months or whatever

they tend to give you 12 months free on most of their flagship products like S3 but the rest either no or its always free

on a lot of these you can probably run them outside of the free trial anyways since cloud providers can be really cheap if you know what youre doing

on their VMs or VPSs though you have no hope of competitive prices they tend to be extremely costly

[earnest phoenix]

free tier / trial sucks

[proven lantern]

lambdas have a free quota every month

[frosty gale]

yeah aws lambdas do but to use them for anything meaningful you need to have some "aws router / proxy" instance which is behind the 12 month free trial

[proven lantern]

they have api gateways and it's like $1 for 300 million requests

pay per use

[frosty gale]

yeah its api gateways

[proven lantern]

also they have function urls

those are newer

Function URLs are included in Lambda’s request and duration pricing. For example, let’s imagine that you deploy a single Lambda function with 128 MB of memory and an average invocation time of 50 ms. The function receives five million requests every month, so the cost will be $1.00 for the requests, and $0.53 for the duration. The grand total is $1.53 per month, in the US East (N. Virginia) Region.

5 million request for $1

not bad

300 million for $1 is better

[neon leaf]

30 Mil for 0$

[proven lantern]

where should i move my domains to now that google domains shut down?

[sage bobcat]

One message removed from a suspended account.

One message removed from a suspended account.

[proven lantern]

they moved us over to squarespace

no thanks

[quartz kindle]

i used to run 2000 servers on google's old free tier, which is like 0.25 vcpus and 512mb ram

[quartz kindle]

oracle terminated my account without explanation and deleted all my files

i was running a minecraft server on their free ampere vms

[quartz kindle]

if you're lucky, xd

[deft wolf]

You must have done something wrong, they wouldn't have deleted your account for hosting a minecraft server... right?

[quartz kindle]

i literally had nothing else in there

only a single minecraft server, which i setup according to their own guide as well

lmao

and the server would be inactive 99% of the time

maybe they didnt like that 24gb ram were being used to run nothing 99% of the time

or that i didnt upgrade to any billing plan after 1 month

i also had issues signing up, had to contact their support about it because it wouldnt let me register

[deft wolf]

Or maybe they thought you were making money from the minecraft server without sharing a penny with them

[quartz kindle]

oh yeah, making money from a server that had at most 4 people on

xD

15/06/2022 - account created

12/08/2022 -

[deft wolf]

How can they serve you in the future if they just terminated your account?

[quartz kindle]

:^)

[dusky idol]

Hey I'm experiencing slowness in my bot's performance after it runs for a long duration. Like upto 16 hours or more.
I've figured that it's being caused by the usage of Mongodb database, overtime the amount of threads being used increases to upto 120 or even 150 sometimes. But normally it shouldn't go upto 50 I'm assuming.
When I restart the bot, the count goes back down to normal obviously but builds back up over span of time. Currently I just re-start the bot once a day and it's not that big of a deal as restart time is quite acceptable. But I was wondering how can I manage the threads properly and just permanently sort this issue, read some docs not that helpful. I can't switch from mongodb to sql now as there's approx 50k lines of code

If anyone ends up responding, feel free to ping me cus im not looking at channel

[quartz kindle]

what language is your bot written in?

[wheat mesa]

Yeah it’s going to depend on the language, but most langs have a way to profile their performance. I’d highly recommend looking into a profiler, especially if it’s written in a compiled language

But assuming it’s in js I believe there is a js profiler out there somewhere

[quartz kindle]

also depends what he means by threads, actual cpu threads or mongodb connections?

in any case if its increasing like that its a code mistake that needs to be corrected

[wheat mesa]

Also, 50k lines is a MASSIVE number for a discord bot

Going to hope it’s not 50k lines of js 😭

[small tangle]

49k of them is a handwritten isEven :^)

[wheat mesa]

Classic

[lyric mountain]

[wheat mesa]

Yours is different

You use actual software patterns and such for your bot

Plus it’s java so boilerplate

(And afaik your bot is also fairly complex in its features)

[lyric mountain]

oh, wait, I'm actually below 50k

[quartz kindle]

hax

[small tangle]

Imagine writing comments

My code is clearly self-explanatory

[quartz kindle]

whats the estimated cost to develop?

should be around half a million lmao

mine is saying 200k for 10k lines

[wheat mesa]

I used that tool on 188 million lines of python I autogenned and it said something like 6 trillion dollars

[quartz kindle]

xDDD

its 71 million if i include node_modules

[wheat mesa]

holy fuck

Least bloated node modules folder

[quartz kindle]

its actually a small one

compared to many projects out there

wasnt there someone here with 600mb node_modules folder?

mine is 77mb

[wheat mesa]

Most of mine end up being at least 200mb

Installing one package immediately installs 674 package deps

[quartz kindle]

most of that is devDependencies

idk why dev deps are installed by default

for production you have to specifically run npm i --omit=dev

[wheat mesa]

Most of my projects also never make it to prod lol

[quartz kindle]

xDDD

[wheat mesa]

Deploying is hard man

I’m so scared of deployment that I overthink it before I finish a project

[quartz kindle]

my issue is finishing the product, ie packaging the code, writing docs, desigining user interface/website

[late crest]

what is this app called?

[quartz kindle]

https://github.com/boyter/scc

hahaha funny

[radiant kraken]

my bot was over 10k lines of python in one file

🤌

probably the longest if chain ever written

[late crest]

crazy and not funny

[late crest]

ty

[quartz kindle]

107k json lines wat

json database?

lmao

[late crest]

lmaao

nah, it's HSK dataset for learning Chinese vocabulary. Quiz command

[quartz kindle]

jeez

[late crest]

it is pretty hard to maintain tho, cuz it's have 5000 hieroglyphs there with meanings and other things.

[dusky idol]

Python

[quartz kindle]

sounds hard

[late crest]

not that hard, but time consuming yes

[quartz kindle]

do you use a library for mongodb? or do you use some raw api?

do you have a local offline mongodb or do you use the atlas service?

[lyric mountain]

I forgot which tool was that, I use cloc now

ah right, scc

permission denied, for some reason

yikes, snap

[late crest]

windows or linux?

[lyric mountain]

it's...linux ofc

[late crest]

even with sudo?

[lyric mountain]

it's a hard pass for me to have a windows server

[lyric mountain]

yes, snap has those issues

[late crest]

bruh 😦

[frosty gale]

only businesses meat riding microsoft or with serious technical debt use windows server

rarely any good reason to use it

[lyric mountain]

aight, got from binaries

[pearl trail]

[frosty gale]

Windows is the last OS I see as being lightweight and getting out of your way

[frosty gale]

how does this estimation thing even work 💀

[lyric mountain]

idk, there's probably an explanation on their git

[lyric mountain]

not to mention gui-based windows is awful

if you use cmd you lack a lot of tools bash has, if you use powershell you cant use && to chain commands (yes, they broke it) and Everything-Is -In-Weird -Kebab-Case