development

[sharp geyser]

I see

Anyway

is it safe to just call _env.Dispose() in my own Dispose method

[scenic kelp]

i think so?

if you're writing a wrapper then presumably you have ownership of the resource the whole time so there's no way for it to escape

like your wrapper won't give other parts of the code access to the raw db underneath

[sharp geyser]

fair enough

Alrighty well, thanks a lot!

[scenic kelp]

ofc i love yapping about c# <3

[sharp geyser]

I am honestly getting more interested in C#

dk if you saw me yapping about it, but I have recently taken on the hobby of writing rust plugins (the game not the language)

so C# has become my new cup of tea, only thing is my code is hardly performant

I wrote something that while worked, was definitely the worst possible way to code it

It used multiple loops, some even nested, a lot of if statements that caused other loops

[frosty gale]

i dont think c# has a borrow checker from what ive heard

[wheat mesa]

C# is great

I’m going to be primarily a C#/ASP.NET developer for my new job and I’m so glad they don’t use dogshit frameworks

[naive pulsar]

const Topgg = require("@top-gg/sdk")
const express = require("express")

const app = express()

const webhook = new Topgg.Webhook("your webhook auth")

app.post("/dblwebhook", webhook.listener(vote => {
  // vote will be your vote object, e.g
  console.log(vote.user) // 395526710101278721 < user who voted\

  // You can also throw an error to the listener callback in order to resend the webhook after a few seconds
}))

app.listen(80)

how i can get my webhook auth

no

don't ask me for click unknow links please !

I don't know, your profile is strange, your server is strange, everything is strange. Thank you for your help

No, I just have a mind that thinks someone sent me a server with about 70 people, the same pfp and the same name, and you also sent a strange link asking me to enter the server? Look if you wanna help help

[shell tundra]

do not contact them @naive pulsar

@oak cliff scammer @gray plume

deleting your messages won't do anything you realise we have logs? @gray plume

[naive pulsar]

i will

[shell tundra]

they're a scammer, suit yourself

[sharp geyser]

Dont think thats what they meant

Also, i already pinged a mod about that exact person

how are they still here

💀

[shell tundra]

insane

[deft wolf]

crazy

[shell tundra]

reported internally, someone will see

[deft wolf]

We need more mods

[naive pulsar]

Only by reviewing his account do you find that he is a scammer of the degree of stupidity

[shell tundra]

lmao

[warm surge]

Let me scare them out of the server 👀

[shell tundra]

no

[warm surge]

Y

https://tenor.com/view/please-cat-pretty-please-begging-cute-gif-10675419

[shell tundra]

they'll be banned anyway no point

[warm surge]

Hm

[naive pulsar]

http://yourserverip:port/dblwebhook so if i using hosts service etc how i can get then server ip i'm still confused

[warm surge]

From your vps

but if you’re not using vps, you have to configure your windows host bs

[naive pulsar]

shi

[naive pulsar]

i can't use my api ?

like api.bot.xyz/dblwebhook

[warm surge]

you can

Mine is setup like https://oliverbot.xyz/api/dblwebhook

My nginx points to my api

[naive pulsar]

do you have docs or something to how i can do like this

[warm surge]

currently not at this moment because im currently at school and i don’t have my laptop with me

Wait

Wait

[naive pulsar]

np dude when you back or anytime you can send it take you time : )

[warm surge]

location /api/ {
proxy_redirect http://localhost:8000 https://oliverbot.xyz/api/;
proxy_pass_header Server;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Scheme $scheme;
proxy_set_header Host $http_host;
proxy_set_header X-NginX-Proxy true;
proxy_connect_timeout 5;
proxy_read_timeout 240;
proxy_intercept_errors on;

    proxy_pass              http://127.0.0.1:8000;
}

My nginx config

[naive pulsar]

this is

[warm surge]

Ye

If you want @real rose if he wants to, he can help you with nginx

[naive pulsar]

i see ill use youtube ig he will not respond

[warm surge]

He’s currently doing stuff in irl probably

[raven nexus]

this guy dm me also and ask me how to get top.gg api

[quartz kindle]

where is your api hosted?

[naive pulsar]

vercel

[naive pulsar]

yep i did't do anything wrong ig

also not top.gg

api

[quartz kindle]

then you can create a route there for topgg webhook

but your bot is not running there is it?

[naive pulsar]

yep

this make me wondreing

i must to move

all

daliy coins code

[quartz kindle]

what do you want to do with the vote?

[naive pulsar]

to

my api

[naive pulsar]

give the user some coins

[quartz kindle]

where is the bot's database?

[naive pulsar]

moongoose i can do it at api but api at vercel goes to sleep

after no request for like 5m

[quartz kindle]

local mongo or mongo atlas?

[naive pulsar]

atlas

[quartz kindle]

then you can connect to atlas from the api directly

topgg webhook -> vercel api -> mongo atlas

dont need to access the bot

[naive pulsar]

yeah but vercel host goes sleep after no request for 5m

so if you vote you will not get a coins

[quartz kindle]

you will

a vote is a request

[naive pulsar]

oh !

So I need to add my api to topgg webhook then by express.js get a post of vote ?

[solemn latch]

Does vercel even support express?

Turns out it does.
https://vercel.com/guides/using-express-with-vercel

neato

[quartz kindle]

you dont need express

you dont need topgg lib either

a vote webhook is an http request like a page view or route access

you just create a route in your vercel app, and put the url in topgg

and your vercel app will receive the request in that route

[naive pulsar]

I see now

Like

api.bot.xyz/vote

and but it at topgg

[quartz kindle]

yes

[naive pulsar]

Then this will be like post

So I can get info of users who posts ?

[quartz kindle]

ye

[naive pulsar]

Thanks dude

I'll try it

[quartz kindle]

export async function POST(req) {
  const voteData = await req.json()
}

[naive pulsar]

Yep ig it wasn't possible then I found a way to use it was so useful for me

[surreal sage]

ok so

uhh

if you member.roles.cache.get(guildId)

is that valid

or just no

[quartz kindle]

yeah, that should get the everyone role

[naive pulsar]

so now what happend if someone fake a post

app.post('/vote', async function(req, res) {
  const VoteDate = req.body
  console.log(VoteDate)
});

i mean now anyone can get a unllimit votes

use postman

then fake posts

how i can fix it

[deft wolf]

@solemn latch can you pls?

Or @harsh nova

[naive pulsar]

scam

[deft wolf]

I think that's what the "Authorization" field is for

[naive pulsar]

yeah ig this soo

but

how i can get it

amm

[deft wolf]

https://docs.top.gg/docs/Resources/webhooks

[naive pulsar]

i can't see ip while post

[lyric mountain]

you're the one who defines what the authorization is

[quartz kindle]

thats what authorization is for

nvm tons of people already said it lmao

[deft wolf]

Your message has an extra aura

[quartz kindle]

lmao

[lyric mountain]

tim to himself

tim to everyone else

[neon flicker]

Is it possible to cancel a collector?

Oh nevermind, it is .stop()

[naive pulsar]

the authorization by ip

or token

i didn't get it yet

[naive pulsar]

https://tenor.com/view/dragon-ball-dragon-ball-super-goku-goku-ultra-instinct-ui-gif-10017138207648182009

[eternal osprey]

hey guys my brain is literally not braining today,
// calculate timestamp of log,if it is within 24h, stop,
// if it is within 48h, give points. Outside 48, lose streak.
is this valid logic for achieving the following:

• user daily streaks.

idk, like suppose the timestamp is within 24h compared to the past one, it is still the same day right. So we just break.
But if the timestamp is within 48h, it is the next day, and we then update timestamp.
Buuutttt, if the timestamp is outside 48h the user forgot to maintain their streak right?

[quartz kindle]

check request headers

topgg sends an "authorization" header

the value of the authorization header is the string that you put in your topgg settings (url + auth)

for example:
url = api.nuvex-bot.xyz/vote
authorization = abc1234

code (pseudocode):

if(request.headers.authorization !== "abc1234") {
  return request.send("invalid request")
}

[naive pulsar]

i see

thanks dude

you helped

me so much

all respect 🫡

https://tenor.com/view/sigma-detected-respect-sigma-respect-patrick-bateman-american-psyco-gif-17523786110322592071

[quartz kindle]

sounds correct

but you want to count 24h from where? since last timestamp or something fixed like midnight UTC?

[eternal osprey]

EST

i log each time a user performs a certain action

that is when the "timer" starts.

[lyric mountain]

What are u trying to do? There might be a better way

[eternal osprey]

I have 4 cases, they all work based on a previous timestamp i saved:

user does action x at timestamp t2 -> owh no, t2 is still on the same day! -> error
user does action x at timestamp t3 -> t3 is the next day, Yaaay!
user does action x at timestamp t4 -> t4 is 2 days later, nooooo!

[lyric mountain]

why not store timestamp on database?

[eternal osprey]

i do

[lyric mountain]

then simply compare with current date

[eternal osprey]

yeah i calculate the difference in hours

i was just figuring if my logic was correct

[lyric mountain]

it's best if u work with millis, this way u dont go into float shenanigans

[eternal osprey]

ah yeah that's true

[lyric mountain]

to get length it's just current_millis - previous_millis

[eternal osprey]

yeah i got everything setup already, but indeed i might move to millis instead of hours

[past field]

question! so what’s a good way for me to get players feedback?

basically what i’m thinking

i have some responses that the bot sends based on certain keywords via event listener

but i want players to have a chance at submitting some funny responses

so i was thinking maybe i can send an embed at the end of my clickwar game for ppl to click and submit, but how can i do that and actually get their responses?

in the console log?

[lyric mountain]

ticket system

doing automatic feedback collecting would be against tos

[past field]

oh

[mossy zealot]

I have this to detect if someone votes on my bot, But it doesn't work..

I just added the topgg webhook, I don't have a custom website.

I just want to see who voted on my bot, and give them rewards

[quartz kindle]

where is your bot hosted? what did you put in your topgg bot settings?

also, why are you creating an Api instance if you're not using it anywhere?

[past field]

meaning link them to create a ticket in my server?

[quartz kindle]

no, like having a ticket command that people use to submit feedback

but you can always link them your server and tell them to go there

[mossy zealot]

I honestly don't know

[past field]

but how would i receive the feedback? in the console log?

[solemn latch]

webhook-topgg and having your own webhook listener doesnt work.
You can choose 1 or the other

[quartz kindle]

you are using a third party website that redirects votes to discord, not to your bot

[quartz kindle]

idk you decide

can be console.log, can be fs.writeFile, can be database

[past field]

fs.WriteFile?

[quartz kindle]

like if you want to create text files

or something

[past field]

hm

[quartz kindle]

you can also just have your bot post on some channel in your server

[past field]

i like this

[mossy zealot]

How do I redirect it to my bot?

[quartz kindle]

where is your bot hosted?

[mossy zealot]

It's hosted on a panel

[quartz kindle]

what panel?

[solemn latch]

The way webhooks will be handled will often depend on the host.

[mossy zealot]

it's a custom host

so idk

[solemn latch]

like, at home?

[mossy zealot]

https://panel.ranked.network/server

[quartz kindle]

contact the owners of that host

ask about ip address and port

[mossy zealot]

i have ip adress and port

[quartz kindle]

then your URL is http://ipaddress:port/vote

[mossy zealot]

how does that make sense if the url doesnt have a /vote

idk if u say itll work ig

[mossy zealot]

yeah but i saw this when i looked it up too

but i thought it was for websites with custom vote page

[solemn latch]

webhooks, and webhook urls are not for websites.

The way I like to explain it, is it uses the same communication methods as a website, but its for computer to computer communication.

[mossy zealot]

I see

[mossy zealot]

It's weird to explain but basically we have a online PC (i forgot what that's called) and it's just made into a nice panel

I also forgot what that panel is called

[quartz kindle]

is this pc in your home?

or in a server somewhere?

[mossy zealot]

server

we host MC servers and discord bots on there

for example

[quartz kindle]

alright then just do that, ip address and port, plus whatever path you have in your express listener

[mossy zealot]

how do I know where my express listener path is

i dont have access to the server

only the panel

[solemn latch]

/vote is the path

[mossy zealot]

the port in my code should be the same as my host?

[solemn latch]

The port in code, in your url needs to be the same.

If the server has any firewalls it must be allowed

[past field]

you're freaking awesome man. so i did this, i even added an "Accept" button that will DM the user who submitted the suggestion to tell them it's be accepted.

but now i'm thinking about when i have to restart the bot

the accept button wouldn't work anymore

[pine willow]

why shouldnt it work.

[past field]

so i'm guessing i can add this to the database, adding a table to keep it working

[past field]

because i didn't make a table for it in sql

it's just in the bots memory

so if i restart it wouldn't work (i think)

im new to databases so idk

[pine willow]

i can restart my bot. And the buttons still work after. No DB

[mossy zealot]

@solemn latch sorry for the tag

[past field]

let me try

[pine willow]

you give them custom IDs? right.

[solemn latch]

This looks like sftp, not http.

[pine willow]

Should be required

[pine willow]

Its saying SFTP

[solemn latch]

I've seen some panels say weird things due to poor configuration by the owner.

So I try not to assume 😄

[pine willow]

ouf

[past field]

yes

interaction failed when bot restart

[solemn latch]

I'm not sure what that is, since we dont know what panel software you're using.
It might work? I'm not sure.

[pine willow]

Doesnt look like ptero

i cant identify it

[solemn latch]

I'd suggest reaching out to whoever set it up tbh.

[pine willow]

Or reinstall it. When you are the owner

[solemn latch]

it sounds like the panel is a group thing, with multiple people using it. another person might know.

Sadly, its just a guessing game without more information.

[pine willow]

yessir

[mossy zealot]

I will try to find out what kinda host it is

@solemn latch omg after long searching

PufferPanel

VPS connected to PufferPanel

[quartz kindle]

that doesnt really help

you need to see which ports are open for you to use

and whether there is any internal redirection

does the panel show any "public url"?

or public ip port?

[sharp geyser]

Right so im using lmdb, but I am not sure the approach I want to go for.

I can either go nosql route like mongo and it basically just be a json table, or I can go the postgres route and have rows / columns and eventually support relations

[quartz kindle]

although if you have access to sftp, then it might be a full vps, and not an app hosting panel thing

if you have access to the full vps, then the port is the one you define inside your code, in express

you just need to check if there are any firewalls installed

[mossy zealot]

well the host is closed off since it's a panel you need to login to

the master url is the website

[quartz kindle]

try creating a basic express app

just a single app.get("/")

and run it on some random port, like 3000

then try to access it and see if it works

both with ip:3000 and with website:3000

[mossy zealot]

the website port is 8080 tho

[quartz kindle]

you cant use a port that is already used

[mossy zealot]

ohhhh

i see

does an offline port work?

If I take a port of one of the servers that is offline

[quartz kindle]

you can, but you can chose a port that is not used by anything

the program will create the port for you

unless there is a firewall that blocks ports that are not configured

[mossy zealot]

yeah there is a firewall which he had to manually add so it could work

[past field]

good website to upload terms of service for bot verification?

[mossy zealot]

github is a free website host

You could technically use pastebin

[frosty gale]

free meaning that github can cutely feed your source code to train an LLM

[rustic nova]

Yummy data

mmmmmmmmmm

Yummy api secrets I'm going to use as examples for others

mmmmmmmmmmmm

[eternal osprey]

   console.log("current", timestamp)
    console.log("previous", user[0].last_logged_timestamp)
    const differenceInMs = timestamp - user[0].last_logged_timestamp;
    console.log("difference", differenceInMs)```
is this an overflow that is happening?
numbers don't make sense

[pearl trail]

timestamp is Date.now() ?

[sharp geyser]

It depends on the type of timestamp it is

Sometimes they are in seconds, other milliseconds

that 000 one is a unix timestamp

Its in milliseconds

As you can see

current & difference are in milliseconds

previous is in seconds

Its not overflow its just different formats of timestamps

[small tangle]

Depends if they really want to train their model on my shitty code

[quartz kindle]

the more shit code the better, so that ai cant steal our jobs

:^)

[pine willow]

I just created this Auth page for my program and then i noticed that i cant use it because of Spotify copyright and stuff

[lyric mountain]

cant u just remove the logo?

[pine willow]

Isnt the name copyrighted too?

[neon leaf]

thats not how copyright works

[pine willow]

sure i can

[neon leaf]

it just means the name is "claimed", you can still use it yourself as long as its clear its not related to you

[pine willow]

okay okay

Anyways. Does it look good ??

[lyric mountain]

I mean, spotify also has a trademark, you cant use the name Spotify anywhere without permission

BUT this only applies to things you make

[pine willow]

I wanted to make the program public. Thats why i ask

[lyric mountain]

if you're using spot's oauth2 then this doesnt apply to that, because it's their site

[pine willow]

i use that. But the auth page that you see is coded by me

Not by spotify

[lyric mountain]

just to make it clear, you're not using oauth2?

[pine willow]

I use this.

this is oauth2 i guess

and when you link my app with the program

you get redirected to the page youve seen

[lyric mountain]

yes that's oauth

[pine willow]

I added it because otherwise nothing would be displayed

[lyric mountain]

you can make references to their name as you've just came from their site

[pine willow]

okay

[lyric mountain]

basically, you're only allowed to use their logo/name if it's linking to their site in some way

or if they give you permission for other purposes

[pine willow]

Okay. Thanks for you help!

[quartz kindle]

i mean

just publish it

the most it will happen, if anything, is they ask you to remove it

[pine willow]

https://tenor.com/view/nervous-pudgy-penguins-tensed-gif-10064957258394862850

[quartz kindle]

once i was working on a website for a friend, he had a shop and sold some products

and wanted to mention that his shop accepts payment throguh payoneer

and he got a c&d email from them

lmao

[pine willow]

o.o

How do they even find these small projects using their assets

😭

[earnest phoenix]

or you get a c&d

[quartz kindle]

i replied to that email asking for more details about it, but they never responded. after a month of waiting we just removed the payoneer payment method from the shop

they never replied, nor followed up

they just use automated bots and scanning shit for copyrighted words

and send automated email threats

[pine willow]

I see.

Just use Spotfy

✨

[quartz kindle]

spoofify

[deft wolf]

Spot!fy

[pine willow]

Goofify

[naive pulsar]

What is the best you to make OTP for my website ? Use nodemail I just wanna a way to block him to send many OTP also if he write many wrong Password and rech limits got ban for like 5m I just will use a Cokies to save times of send and last send was when but if you guys have a better way please say it

[quartz kindle]

i wouldnt use cookies for that

use the user account in the database

track login attempts there

[past field]

music bot.. is it hard to do? 👀

[pine willow]

when you want a bot with cool features

it can be hard

[earnest phoenix]

"cool features"

https://tenor.com/view/clone-clones-buzz-lightyear-toy-story-gif-21609842

[deft wolf]

Unique*

[vagrant veldt]

this is Fr3e server

[deft wolf]

Nah, we pay Tim monthly to send messages on this channel

[quartz kindle]

Lol

[pine willow]

yes

[earnest phoenix]

same thing

[past field]

if i defer a button with deferUpdate, could a channel message be the interaction acknowledgement?

[lyric mountain]

the acknowledge is the deferUpdate itself

[past field]

so what would i use to have the button acknowledged and a channel message send as a result of that button tap?

basically in click war, the players have lives. i have a “swap lives” button but i don’t want it to send an ephemeral.i want them to tap it and the bot sends a channel message who the user swapped with. it does that, but it also gives interaction failed, no errors in console logs. so i assume the button works fine, just dk the best way to have it respond without interaction failed.

[surreal sage]

import EventEmitter from 'events';

export interface PlaneWebhookEvent {
  event: 'issue';
  action: 'deleted' | 'updated';
  webhook_id: string;
  workspace_id: string;
  data: Record<string, unknown>;
  activity?: Record<string, unknown>;
}

const webhookEvents = new EventEmitter<{
  'eventHere.actionHere': [PlaneWebhookEvent];
}>();

export default webhookEvents;

How can I make the types on eventemitter like that?

dynamic..

ok wait

got it nvm

[sharp geyser]

Im confused

wdym dynamic

[quartz kindle]

you're using a followup? you have to use at least one reply

[past field]

ohhh

so i don’t need to defer

because follow-up

[quartz kindle]

you just have to remember this rule:
interaction -> reply -> followup
interaction -> defer -> reply -> followup

[past field]

ok i got you

tim u da goat

[warm surge]

@quartz kindle hey tim i need your help here

may i slide you a invite real quick, i need you teach a friend how ports 443 and 80 works

[sharp geyser]

uhm

port 80 = http

443 = https

not much more than that tbf

[sharp geyser]

What are they having problems with

[warm surge]

he think you cant use same port for different websites

[sharp geyser]

Then he needs to know how nginx works

not port 443 and 80

[sharp geyser]

Nginx works based off a technique called "virtual hosting"

Specifically name based virtual hosting

Right I got it partially wrong

When a user visits a domain example.com the browser will perform a dns lookup to resolve the IP address behind that domain, it will then send a HTTP request to that ip with the domain in the Host header. Nginx then examines this header to determine which server block to route the request to. It does this by comparing the value in the Host header and the one defined in server_name in your blocks. Once it finds the right one it will serve the content at that location

[warm surge]

thanksss

[sharp geyser]

The reason everything is able to seemingly run on one port is because all of those aren't actually running on a single port

Well they are, its just not in the sense you think

nginx as the service is listening on port 80 & 443

otherwise there would be no need to proxy pass your stuff, and what not.

when you are doing listen 80; you are essentially saying have this website listen on the nginx port 80

[warm surge]

thanks misty

[sharp geyser]

👍

Some of my explanation may be a bit wrong, but the gist of it is correct

[quartz kindle]

thats what a webserver software does
listen to X port, examine requests that come in, and do somegthing with them

[sharp geyser]

oh tim is here to fill in the holes I left

[quartz kindle]

you could literally run nodejs on port 80 and check headers yourself, and handle multiple websites with a single nodejs program

or redirect them elsewhere if you prefer

[sharp geyser]

yuh

but

nginx better

it does it all for me

[quartz kindle]

yeah nginx is better at it because it was made expecially for it

[sharp geyser]

ye

btw tim

while you are here

I has a question regarding my db

Since I am using lmdb, I figured why not just try and make a relational database.

Problem is, how exactly should I keep track of table structures & relations properly

should I just use indexing and what not?

Create an additional file alongside the db files that keep track of relations?

[random willow]

anyone looking to get paid for making me a elo bot lmk

dm me

[quartz kindle]

many ways you could do that, but in general you use and abuse lmdb's key-value

store everythign as a key value

table structures, relations, indexes, etc

[sharp geyser]

How would that work?

Seems like it would be a pain in the ass to make sure everything stays where it should & look it up appropriately

[quartz kindle]

tables -> array of table names
tables:name -> structure for table
index:table:col:value -> index for table col
index:table:col:value2 -> index for table col
index:table:col:value3 -> index for table col

etc

you can also create seperate lmdb databases for some of those

[sharp geyser]

interesting

That wont impact performant will it?

As in looking up

[proven lantern]

create keys like this

const eloKey = ({rating_type, leaderboard, versus, teamSize, id}) => {
    if (rating_type === 'distinct_team') {
        return [rating_type, id, leaderboard, versus].join("|")
    }
    if (rating_type === 'combined_team') {
        return [rating_type, id, leaderboard, teamSize].join("|");
    }
    if (rating_type === 'distinct_player') {
        return [rating_type, id, leaderboard, versus].join("|");
    }
    if (rating_type === 'combined_player') {
        return [rating_type, id, leaderboard, teamSize].join("|");
    }
    if (rating_type === 'overall_player') {
        return [rating_type, id, leaderboard].join("|");
    }
    return 'error_invalid_rating_type';
};
module.exports = eloKey;

[sharp geyser]

[proven lantern]

store keys in the key
["settings", guildId, userId].join("|");

like the guild and user ids

foreign key

[sharp geyser]

No ty

[proven lantern]

if you store it this way it's the fastest lookup

since you lookup by key

it's like a composite key

[wheat mesa]

I just discovered datagrip

I’ve been living a lie my entire life

[prime cliff]

Try now done a bunch of improvements to design and performance tweaks including a neat loading section 🙂 https://devspace.fluxpoint.dev/

[quartz kindle]

looking great

loads pretty nicely now

[prime cliff]

Some stuff is a little broken with the design changes but it's gonna be a lot better in the long run

[lyric mountain]

Gets better when u start using intellij or one of the other ides

Cuz u get all of datagrip on top of the ide

I've been using it to edit in-db scrips with intellisense since forever, couldn't go back to other tools even if I tried

[ionic schooner]

IntelliJ stuff is bussin bussin, really has everything you need and doesn’t cost much

[sharp geyser]

little confused on this actually, what is this -> index for table col

Does that not defeat the point of the key being an index:table:col:row

[proven lantern]

it's so you can paginate through the table in addition to lookups by key

[quartz kindle]

yeah i didnt write it correctly, i meant you can use lmdb's key-value for the indexes

for example you want an index in table users, column id

you can do it like index:users:id:374958739487 -> table row

[sharp geyser]

wait

isn't the numbers after id the value of that row?

Im a little confused on your thought process

[quartz kindle]

the numbers would be the user id

the user id is indexed, so it has to be fast to query

so it needs to be part of an lmdb key

[sharp geyser]

[quartz kindle]

and the value for that key is the row you are querying

or some sort of unique row id

something that links to the desired row

[sharp geyser]

hm

So would I have this is a separate lmdb database

[quartz kindle]

basically, database indexes are just a key->value map

[sharp geyser]

and it links to keys in the other one

[quartz kindle]

you could create a separate database for this

but lmdb is made to handle data like this

similar to redis

[sharp geyser]

what exactly is index actually

you append it at the beginning of the key

[quartz kindle]

to separate it from other types of keys that are stored in the same database

[sharp geyser]

icic

[quartz kindle]

for example table structures

or whatever else shares the same db

you basically create namespaces inside a single db to separate the data

and lmdb can do range queries

so you can query all keys that start with said prefix, for example

[sharp geyser]

so
index:users(table_name):id(column):11231231233(valueOfId) -> 11231231233
index:users:name:11231231233 -> Jane Doe

[lament rock]

Actually working with it all will make sense

[sharp geyser]

I assume appending the id would make it faster to query since i can look it up based off id

[quartz kindle]

since its an index, the purpose of it is to quickly find a row in a table, to prevent having to scan the table to find it

so the value that is being stored has to resolve to the full table row

[sharp geyser]

so it has to resolve to everything for that entry?

[quartz kindle]

yes

[sharp geyser]

if I am storing a name, phone number, email, etc

if I look it up, it has to return all that?

[quartz kindle]

yes

thats what an index is for

[sharp geyser]

[quartz kindle]

you dont create an index in a db to find only that column

[sharp geyser]

I wonder how that will work code wise

[quartz kindle]

you create an index to find the row based on that column

[sharp geyser]

is that the point of primary keys?

[quartz kindle]

yes

[sharp geyser]

interesting

[quartz kindle]

but primary keys are also unique row ids

which by themselves alrewady are an index

[sharp geyser]

so they have to be unique

[quartz kindle]

just stored in a different namespace

for example

rows:999999 -> [john doe, 78798942342, [e-mail]]
index:users:id:78798942342 -> 999999

[sharp geyser]

Ah

So it would just be an array of all the values

[quartz kindle]

yes

[sharp geyser]

interesting

[quartz kindle]

the names of the columns would be defined in yet another namespace

[sharp geyser]

yknow, that makes sense if you look at a db viewer

[quartz kindle]

the full thing can be something like

[sharp geyser]

I assume the order of the values put into the array matter

else they will get mixed up

what should be the phone number column, ends up having a value from the name column

or smth like that

[quartz kindle]

tables:1 -> [id, name, email]
rows:1:79287349823 -> [111111, john doe, [e-mail]]
indexes:1:id:111111 -> 79287349823

[sharp geyser]

[quartz kindle]

yeah it has to match the structure of the columns

the columns are always ordered

[sharp geyser]

icic

Well, this gave me some guidance

thanks a lot

Hopefully I will have smth to show you at the end of the week

[quartz kindle]

good luck!

[sharp geyser]

Im still debating on whether or not to make my own sql language

or just pull an EF Core and make a wrapper around the database

not even sure where i'd begin writing my own sql language 💀

[sharp geyser]

Whats the difference between IAsyncDisposeable vs IDisposeable? I see some sources recommend using it together but not sure why.

[scenic kelp]

asyncdisposable is newer and there's a big warning on the page so

[sharp geyser]

I never thought to look at the actual page for it

[civic scroll]

did i fail

or did i success

[civic scroll]

real

[raven nexus]

HI, i have question possible can we make the embed like this

Name : Test
Detail : Hello world

[lyric mountain]

what do you mean?

[quartz kindle]

like this?

[raven nexus]

yea but it like got a square

around the word

or like use any custom character to do

[lyric mountain]

Name   : Test
Detail : Hello world
``` like this?

[raven nexus]

nope wait

[quartz kindle]

?

[raven nexus]

like this but the box outlike replace with -----------------

[raven nexus]

yea

[lyric mountain]

that's what I did lul

[raven nexus]

something like this

[raven nexus]

not the XX

[lyric mountain]

it's the same thing, it's just that embeds make the black bg smaller

[raven nexus]

what can i decoration with it?

[lyric mountain]

well u could add a blank field after CTAF frequency

to align the fields

[quartz kindle]

lol

[lyric mountain]

le olde ascii graphics

[raven nexus]

how to do it like that then

[lyric mountain]

if you're on windows, press Win key and search "charmap"

[quartz kindle]

https://onlinetexttools.com/draw-box-around-text

[lyric mountain]

oooooor there's a site for that ofc

[quartz kindle]

https://message.style/app/editor

[quartz kindle]

theres always an app for that

:^)

[raven nexus]

it still not show for the bot did i miss something?

how could i put in the middle line?

[lyric mountain]

you cant do that

[raven nexus]

uh

[lyric mountain]

if you want to make it in a single block, you must use description

fields must always have a name and a value, and they are disposed in a 3-column (1-column for mobile) way

[raven nexus]

ok

[neon flicker]

Does Discrod API support the new guild system?

Like is it possible to fetch somebody's current guild tag?

[pearl trail]

oooo browserling

i love opening trojan and malware sites there

[neon flicker]

Okay for anybody wondering in the future, neither Discord API or any Discord.js versions support it yet

[lyric mountain]

it'll probably support eventually, it's just that it's still in thanos-beta stage

banners took a long time to be documented in the api

[raven nexus]

i sometime got bug

some time no error

is tere a way to fix? the problem only start with when the command was use for 2 time

the first run notthing problems

is there like completely remove the cooldown? if remove does it will impact anything

[lyric mountain]

yes, the impact will be that it'll have no cooldown

who wrote that code?

[raven nexus]

my friend do it for me

after i edit the last part which is the photo i send before then it crashes

[lyric mountain]

ask your friend to fix the error, you can't really attempt to fix something like that without understanding what the code is doing

ah, fortune, I remember you

[raven nexus]

usally like for discord api what is the minimum timeout for the bot

[lyric mountain]

cooldown isn't about the api

it's to prevent people from spamming commands

[night lagoon]

replace command.cooldown to command?.cooldown

[eternal osprey]

async function calculateBet(data: any, sportsBook: string): Promise<number> {
  const probability_winning: number = 1 / data.previous_price;
  const probability_losing: number = 1 - probability_winning;
  const bet_return: number = data[sportsBook + "_price"] - 1;
  const base_price: number = (
    (probability_winning - probability_losing / bet_return) *
    100
  ).toFixed(2);
  const recommended = (base_price * (1 / 3)).toFixed(2);
  return recommended;
}

huh base_price is a string according to ts?

[lyric mountain]

toFixed is a string

numbers dont have definite decimal count

[eternal osprey]

owh damn that's intersting

[lyric mountain]

for example, 2.0 doesn't have 1 decimal place

cuz 2 == 2.0 == 2.000000000000

it only has a fixed decimal count when viewed as a string

if you want to truncate to a certain precision, you have to multiply, truncate and then divide

((value * 100) | 0) / 100 will truncate to 2 decimal places for example

add/remove zeroes to 100 to change how many decimals are kept

[neon leaf]

why dont you type the data argument

async function calculateBet<Book extends string>(data: Record<'previous_price' | `${Book}_price`, number>, sportsBook: Book): Promise<number> {

[lyric mountain]

holy hell that's big

[neon leaf]

thats not big

[eternal osprey]

i see!

[eternal osprey]

the data is a big object

way too big to parse it so i just defined it as any

[neon leaf]

yea this accepts any object that contains these things

[eternal osprey]

wait really?

[neon leaf]

yes

[eternal osprey]

new to ts so this is awesome to learn

[neon leaf]

thats how ts types works

ts interfaces dont allow this

[eternal osprey]

i like ts as it combines 2 things i like, js and java

[lyric mountain]

eh, the typing system is closer to kotlin

which is why I dislike both

[eternal osprey]

okay not really java but idk i find it similar tbf

like the whole type system, the objects and classes

[frosty gale]

@quartz kindle how would you speed this up?

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);
        }
    }
}

this function basically goes through each entry, clones & reverses the buffer (since you cant reverse without making a copy of the buffer) and then goes through each entry to see if the reverse matches any other buffer

with 100k elements this takes very very very long

im guessing its mainly from the buffer copying with Buffer.from and the reverse method

i'll try to substitute that first

[neon leaf]

const reversedMap = new Map()

for (const entry of entries) {
    const reversed = Buffer.from(entry.bytes_buffer).reverse()

    if (reversedMap.has(reversed.toString('hex'))) {
        console.log("Duplicate found!")
        process.exit(1)
    }

    reversedMap.set(entry.bytes_buffer.toString('hex'), true)
}

could help

(will help)

[eternal osprey]

sender.ts:

module.exports = { sendData };```
receiver.ts:
```tsconst { sendData } = require("./sender.js");```

is this allowed/possible?

i know ts mainly uses import from

[neon leaf]

its a war crime

[frosty gale]

yeah that is much faster lmao

[eternal osprey]

[frosty gale]

only in javascript can you convert bytes to a string and use a map and it becomes much faster

although its probably because this only iterates 100k times as opposed to 100k*100k

[neon leaf]

I mean previous solution was O(n²)

[eternal osprey]

i see, but is it possible? Cuz when i hover over to it it doesn't display the function signature

[neon leaf]

this is O(n) yeah

[eternal osprey]

also, should i require from the .ts or the compiled .js

[neon leaf]

typings wont work

why would you want to use require

[eternal osprey]

cuz import from gives me this problem:

sender.ts(7, 16): 'sendData' is declared here.
import sendData

[neon leaf]

why are you importing the js file

[eternal osprey]

module.exports = { sendData };```
i might be ugly and retar3ded though

[neon leaf]

just import from ./sender

no file extension

[eternal osprey]

really? I never knew that was a thing

[neon leaf]

and remove module.exports

either do export default sendData

[eternal osprey]

wow i deadass suck

[neon leaf]

or export { sendData }

[eternal osprey]

i see, that is both js and ts syntax right?

just making a note here

[neon leaf]

yes, its esm

(modern js)

[eternal osprey]

i always though that you MUST use module.exports = {...}

instead of just export {..}

[neon leaf]

well if you are using cjs yes

but cjs is a curse that nodejs brought upon us many years ago

[proven lantern]

const reversedSet = new Set();

for (const entry of entries) {
    const reversedHex = entry.bytes_buffer.toString('hex').split('').reverse().join('');

    if (reversedSet.has(reversedHex)) {
        return "Duplicate found!";
    }

    reversedSet.add(entry.bytes_buffer.toString('hex'));
}

return "No duplicates found.";

[neon leaf]

this seems pretty slow

splitting a string into an array

[frosty gale]

i did the same thing just now but without the splitting

const reversedSet = new Set();
for (const entry of entries) {
    const reversed = Buffer.from(entry.bytes_buffer).reverse().toString();

    if (reversedSet.has(reversed)) {
        console.log("Duplicate found!")
        process.exit(1);
    }

    reversedSet.add(entry.bytes_buffer.toString());
}

you can actually just convert the buffer to a string and you can use it as a key for the set/map

no need to convert it to hex or anything special

it will still hash the same

[neon leaf]

did you test if thats faster?

[frosty gale]

didnt see if its faster but i can benchmark

[sharp geyser]

Im late to the party, but why exit the process if a reversed has been found.

Also, sets can't take in duplicates anyway right, so why check if it already exists?

[neon leaf]

I could imagine sha1 hashing the buffer may also be fast

since map lookups need less chars

[frosty gale]

this is "code to fulfil a specific purpose for only this instance which i will never use again" type of code

and theres code later on that i dont wanna comment out

[sharp geyser]

I see

fair enough

[frosty gale]

which two do you want to benchmark

[neon leaf]

i want to benchmark four

[proven lantern]

hold on, let me rewrite it in zig

[frosty gale]

great

[frosty gale]

no

[neon leaf]

ben,me,you,me but with sha1

[frosty gale]

sha1 should be much slower since it also calls openssl and adds an extra step but sure

plus theres a very small chance of collision which would make this unfit for purpose because of the risk

[neon leaf]

hold up lemme just write an sha1 function in pure js

ok this is not sha1 but its good enough i think

function rotateLeft(n, s) {
    return (n << s) | (n >>> (32 - s));
}

function toHexStr(n) {
    let s = "", v;
    for (let i = 7; i >= 0; i--) {
        v = (n >>> (i * 4)) & 0x0f;
        s += v.toString(16);
    }

    return s;
}

function sha1(buffer) {
    let words = [];
    let message = new Uint8Array(buffer);
    let messageLen = message.length;

    for (let i = 0; i < messageLen - 3; i += 4) {
        words.push(
            (message[i] << 24) |
            (message[i + 1] << 16) |
            (message[i + 2] << 8) |
            message[i + 3]
        );
    }

    let remainingBytes = (messageLen % 4);
    let lastWord = (remainingBytes === 3 ? (message[messageLen - 3] << 16) : 0) |
        (remainingBytes >= 2 ? (message[messageLen - 2] << 8) : 0) |
        (remainingBytes >= 1 ? message[messageLen - 1] : 0);
    words.push(lastWord << 8 | 0x80);

    while ((words.length % 16) !== 14) {
        words.push(0);
    }

    let h0 = 0x67452301;
    let h1 = 0xefcdab89;
    let h2 = 0x98badcfe;
    let h3 = 0x10325476;
    let h4 = 0xc3d2e1f0;

    words.push(0);
    words.push(messageLen * 8);

    for (let i = 0; i < words.length; i += 16) {
        let w = words.slice(i, i + 16);
        for (let t = 16; t < 80; t++) {
            w[t] = rotateLeft(w[t - 3] ^ w[t - 8] ^ w[t - 14] ^ w[t - 16], 1);
        }

        let a = h0;
        let b = h1;
        let c = h2;
        let d = h3;
        let e = h4;

        for (let t = 0; t < 80; t++) {
            let temp;
            if (t < 20) {
                temp = (rotateLeft(a, 5) + ((b & c) | (~b & d)) + e + w[t] + 0x5a827999) | 0;
            } else if (t < 40) {
                temp = (rotateLeft(a, 5) + (b ^ c ^ d) + e + w[t] + 0x6ed9eba1) | 0;
            } else if (t < 60) {
                temp = (rotateLeft(a, 5) + ((b & c) | (b & d) | (c & d)) + e + w[t] + 0x8f1bbcdc) | 0;
            } else {
                temp = (rotateLeft(a, 5) + (b ^ c ^ d) + e + w[t] + 0xca62c1d6) | 0;
            }

            e = d;
            d = c;
            c = rotateLeft(b, 30);
            b = a;
            a = temp;
        }

        h0 = (h0 + a) | 0;
        h1 = (h1 + b) | 0;
        h2 = (h2 + c) | 0;
        h3 = (h3 + d) | 0;
        h4 = (h4 + e) | 0;
    }

    return toHexStr(h0) + toHexStr(h1) + toHexStr(h2) + toHexStr(h3) + toHexStr(h4);
}```

[frosty gale]

ill have to benchmark it manually because the benchmark tool i use is only browser js and browser js doesnt have buffer

it has uint8array but the .toString() method works differently and likely much slower

[proven lantern]

same thing written in zig

[frosty gale]

this is the improvised benchmark

function chloe() {
    const reversedSet = new Set();
    for (const entry of entries) {
        const reversed = Buffer.from(entry.bytes_buffer).reverse().toString();

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

        reversedSet.add(entry.bytes_buffer.toString());
    }
}

function ben() {
    const reversedSet = new Set();
    for (const entry of entries) {
        const reversedHex = entry.bytes_buffer.toString('hex').split('').reverse().join('');

        if (reversedSet.has(reversedHex)) {
            console.log("Duplicate found!")
        }

        reversedSet.add(entry.bytes_buffer.toString('hex'));
    }
}

function hex_username_man() {
    const reversedMap = new Map()
    for (const entry of entries) {
        const reversed = Buffer.from(entry.bytes_buffer).reverse()
    
        if (reversedMap.has(reversed.toString('hex'))) {
            console.log("Duplicate found!")
        }
    
        reversedMap.set(entry.bytes_buffer.toString('hex'), true)
    }
}

chloe();
console.time("Chloe Function");
chloe();
console.timeEnd("Chloe Function");

ben();
console.time("Ben Function");
ben();
console.timeEnd("Ben Function");

hex_username_man();
console.time("Hex Username Man Function");
hex_username_man();
console.timeEnd("Hex Username Man Function");

@neon leaf surprisingly your map function performs faster consistently

Ben Function: 128.123ms
Hex Username Man Function: 82.781ms
Chloe Function: 94.153ms

i have no idea why, just doing toString and using a set which has less overhead should be much faster than using a map and converting it to hex twice

[proven lantern]

slowest?

[neon leaf]

yeah what thats the slowest

[frosty gale]

@quartz kindle please explain

[neon leaf]

makes perfect sense

[frosty gale]

what do you mean slowest

[proven lantern]

biggest number = bad

[frosty gale]

yes i posted the benchmark

[frosty gale]

bens split function is slowest

[neon leaf]

yes ben is clearly 50ms above me

yes

[frosty gale]

oh fuck

meant to tag you

[proven lantern]

const crypto = require('crypto');

// Generate test data
function generateTestData(size) {
    return Array.from({ length: size }, () => ({
        bytes_buffer: crypto.randomBytes(32)
    }));
}

// Original function
function originalFunction(entries) {
    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) {
                return "Duplicate found!";
            }
        }
    }
    return "No duplicates found.";
}

// Optimized function
function optimizedFunction(entries) {
    const reversedSet = new Set();

    for (const entry of entries) {
        const reversedHex = entry.bytes_buffer.toString('hex').split('').reverse().join('');

        if (reversedSet.has(reversedHex)) {
            return "Duplicate found!";
        }

        reversedSet.add(entry.bytes_buffer.toString('hex'));
    }

    return "No duplicates found.";
}

// Performance test
function runPerformanceTest(size) {
    const testData = generateTestData(size);

    console.log(`Testing with ${size} entries:`);

    console.time('Original Function');
    originalFunction(testData);
    console.timeEnd('Original Function');

    console.time('Optimized Function');
    optimizedFunction(testData);
    console.timeEnd('Optimized Function');

    console.log('---');
}

// Run tests with different sizes
[100, 1000, 10000].forEach(runPerformanceTest);

that's the test i ran

[neon leaf]

robert  ~/projects  node sha1.js
Chloe Function: 122.715ms
Ben Function: 144.211ms
Hex Username Man Function: 83.672ms
Hex Username But With SHA1 Function: 277.187ms

woops

[frosty gale]

everything about js is opposite

something that should logically be faster is actually slower

[neon leaf]

i mean the results make perfect sense to me

doing 2 array actions instead of 1 is obv slower

Buffer.from doesnt actually clone the data

it just points it to the old position and slices modified data

[frosty gale]

well bens benchmark yes but yours one convert the string to hex and uses a map which has extra steps but its much faster

[frosty gale]

i tried it in repl and using .reverse() on the new slice makes a clone and doesnt modify the original

but it doesnt matter here

[neon leaf]

well in js you wont see that

v8 doesnt clone though

thats why you always use the offsets when converting to ArrayBuffers

since those are not building on top of the original

[proven lantern]

hex_username_man wins

[eternal osprey]

helpers/parsing.ts:1:20 - error TS2732: Cannot find module '../config.json'. Consider using '--resolveJsonModule' to import module with '.json' extension.

1 import config from "../config.json";
                     ~~~~~~~~~~~~~~~~```
huh why error men waddafuck men

[neon leaf]

wait I just realized something

why is it even reversing

[frosty gale]

hold on ill convert yours to a set and see how it performs

[neon leaf]

my function is 10ms faster without reversal

I dont see why reversal is needed

[frosty gale]

about 2ms faster

[frosty gale]

how would you do it then without reversing

are you just comparing it backwards?

[neon leaf]

Hex Username Man Function: 8.431ms

🔥

[neon leaf]

i just directly compare

all are reversed anyway

so if none are reversed its the same effect

[frosty gale]

lmao i specified the encoding in my toString() and its faster than yours

not by much though

[neon leaf]

and if I dont reverse I can use the v8 object id

function hex_username_man() {
    const reversedMap = new Map()
    for (const entry of entries) {    
            if (reversedMap.has(entry.bytes_buffer)) {
                    console.log("Duplicate found!")
            }
    
            reversedMap.set(entry.bytes_buffer, true)
    }
}```

try this

[frosty gale]

9.908ms

oh wait you dont understand

it has to be reversed

[neon leaf]

but the value isnt used anywhere?

[frosty gale]

with this youre just seeing if theres any duplicates

i want to see if the reverse of any of the buffer bytes matches each one

[neon leaf]

or are the entries themselves reversed

[frosty gale]

so if you have an array [[1,2,3], [3,2,1]]
the function should return duplicate found

because 1,2,3 reversed is 3,2,1

[neon leaf]

robert  ~/projects  node sha1.js
Duplicate found!

works fine

ah wait

i called the wrong fn

[eternal osprey]

why does my tsc add .default to each imported value?

[frosty gale]

function chloe() {
    const reversedSet = new Set();
    for (const entry of entries) {
        const reversed = Buffer.from(entry.bytes_buffer).reverse();

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

        reversedSet.add(entry.bytes_buffer.toString("ascii"));
    }
}

function hex_username_man() {
    const reversedMap = new Map()
    for (const entry of entries) {
        const reversed = Buffer.from(entry.bytes_buffer).reverse()
    
        if (reversedMap.has(reversed.toString('hex'))) {
            console.log("Duplicate found!")
        }
    
        reversedMap.set(entry.bytes_buffer.toString('hex'), true)
    }
}

Chloe Function: 71.578ms (from 95.693ms)
Hex Username Man Function: 80.589ms

all i changed was that in my .toString() i specified the encoding of ascii and now it performs faster than hex mans, does v8 try to guess the encoding by any chance if i dont specify it as a parameter which could make it slower? @quartz kindle

[neon leaf]

default encoding is utf8

[frosty gale]

specifying utf8 makes it much slower now

that makes sense then

although im not sure why utf8 is much slower

ascii and utf8 arent too different

both buffers should be able to be represented in a single byte with little conversion

anyways ive had enough of this language for a day

[neon leaf]

uh wtf

function chloe() {
    const reversedSet = new Set();
    for (const entry of entries) {
            const reversed = Buffer.from(entry.bytes_buffer).reverse();

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

            reversedSet.add(entry.bytes_buffer.toString("ascii"));
    }
}

function hex_username_man() {
    const set = new Set()
    for (const entry of entries) {    
        // also check for reverse
        const reversed = Buffer.from(entry.bytes_buffer).reverse().toString('ascii');
            if (set.has(reversed)) {
                    console.log("Duplicate found!")
            }
    
            set.add(entry.bytes_buffer.toString('ascii'))
    }
}```


 robert  ~/projects  node sha1.js 
chloe: 1.376ms
hex_username: 1.233ms

 robert  ~/projects  node sha1.js 
chloe: 1.189ms
hex_username: 1.207ms

 robert  ~/projects  node sha1.js 
chloe: 1.009ms
hex_username: 0.832ms

 robert  ~/projects  node sha1.js 
chloe: 1.162ms
hex_username: 1.689ms

 robert  ~/projects  node sha1.js 
chloe: 1.299ms
hex_username: 0.873ms

this is cursed

[frosty gale]

if all of this was written in Rust™️ we wouldnt be having this discussion

[lyric mountain]

function reverse() {
  const arr = entries // ffs
  const len = arr.length

  for (let i = 0, l = len / 2; i < l; i++) {
      const reflex = (len - 1) - i

      arr[i] += arr[reflex]
      arr[reflex] = arr[i] - arr[reflex]
      arr[i] -= arr[reflex]
  }
}

[lyric mountain]

try this

oh wait, we're not benchmarking reverse algos?

[neon leaf]

correct

finding duplicate buffers while reversing

[lyric mountain]

what for?

btw a set and a map are exactly the same

[neon leaf]

idk ask chloe

[frosty gale]

set cant store a value though, its a few ms faster than using a map in this case but yeah not a big difference

[lyric mountain]

sets can store a value in fact

[frosty gale]

since when

[lyric mountain]

sets are just an abstracted map

[lyric mountain]

since always, sets are literally maps that have a fixed value

usually null

[frosty gale]

still faster though since you dont deal with the value anymore from the javascript perspective so the native code can do all of that stuff

but performance suggests so

[lyric mountain]

the performance difference is probably from js typecheck

[frosty gale]

@lyric mountain ive had a look at v8 internal code and v8 does differentiate between them

they are technically 2 separate classes but they do extend from the same classes

both extend from JSCollection which extends from JSObject so basically both are objects

i feel deceived

[lyric mountain]

is there anything in js that isn't an object?

[neon leaf]

isnt undefined not an object

[frosty gale]

hashmaps in js are clearly a scam theyre just objects

[frosty gale]

yeah undefined isnt an object

checked the internals its a standalone class

[lyric mountain]

dw we can fix that

[frosty gale]

null also isnt an object

NaN isnt one either but its a very unique type

theres different variations of NaN which appear identical in js

[neon leaf]

typeof NaN === 'number' // true smh

[frosty gale]

v8 source code is so modular you have to go through like 20 different functions to find what you need

[frosty gale]

from the looks of it v8 determines the result of typeof using "tags"

each thing in js is assigned a tag so the nan type is probably assigned a tag of number

which is why you need to use isNaN method

hold on NaN might actually be a number

isNaN checks if something is a number and then does the c++ isNaN on its value

[eternal osprey]

hey is my logic correct?

function getDatePart(timestamp) {
    const date = new Date(timestamp);
    return new Date(date.getFullYear(), date.getMonth(), date.getDate()).getTime();
}

const timestamp = estDate.getTime();  
let previous = Number(user[0].last_logged_timestamp);
const differenceInMs = timestamp - previous;
const currentDatePart = getDatePart(timestamp);
const previousDatePart = getDatePart(previous);
const oneDayInMs = 86400000; 
const differenceInDays = (currentDatePart - previousDatePart) / oneDayInMs;

if (differenceInDays >= 1 && differenceInDays < 2) {
  // it is the next day!
    }else if (differenceInMs < 1){
// it is still the same day
     
    } else if (differenceInMs >= 2){
//it is 2 days later..
    }

[quartz kindle]

what is going on lmao

[eternal osprey]

i don't know 😭

[quartz kindle]

what is the data input for these? like what is entries

[quartz kindle]

sorry i was reading 5 hours of chat where i wa smentioned multiple times exdee

[lyric mountain]

...why

actually why are u constraining to the same day, instead of 24h?

timezones exist, you'll lock your bot to a specific region of the globe

switch ((differenceInMs / oneDayInMs)|0) { // to int
  case 0: // within 1 day (not inclusive)
    ...
    break;
  case 1: // within 2 days (not inclusive)
    ...
    break;
  default: // anything else
    ...
    break;
}

[quartz kindle]

yeah ascii is faster than utf8 since you can skip all checks and just convert charcode to char

[neon leaf]

still love this overly complex ipv6 code https://github.com/0x7d8/rjweb-utils/blob/aeb1d348e58937bf9acc07553743d771ecd8165e/src/network.ts#L415

[scenic kelp]

why wouldn't it be

it's an IEEE 754 thing

[wheat mesa]

yeah NaN is just a magic value

Same with "Infinity"

[frosty gale]

i dont agree with this

who do i complain to

[frosty gale]

overexaggeration it was only 3 times

[quartz kindle]

lmao

but what exactly were you doing

i still dont get it

why reverse the buffers?

[neon leaf]

3,2,1 should also be a duplicate of 1,2,3

[quartz kindle]

why?

[neon leaf]

idk

[quartz kindle]

what about 3,1,2 or 2,1,3?

omg, this dude keeps dming me about how to use may api, he claims to be a developer who's building a system for a client

but everything he does is ai+nocode

"developer" smh

[neon leaf]

smhsmh

do you even have openapi docs

[quartz kindle]

nop

[neon leaf]

time to make some then

those keep everyone away from asking you

[quartz kindle]

i have docs

just not in openapi

[neon leaf]

well

the benefit of openapi is that ai coders can feed it to the llm

and get real output

[quartz kindle]

i dont know if i wanna support ai devs

:^)

[neon leaf]

and nerds can generate their sdks using openapi

so they dont need to use any website

[frosty gale]

it was me checking for possible duplicates in reverse over a dataset in case there were duplicate ids but in reverse

my initial no sleep solution was to do it in a squared time complexity because i overestimated the computing capacity of my cpu

[quartz kindle]

what is entries exactly tho?

how large are the buffers?

[rustic nova]

Lmao

The client probably goes "yeah we need a bit of ai and some of that API"

Both equally stupid

[frosty gale]

an object which contains a 32 byte buffer

[quartz kindle]

im benchmarking this :^)

[neon leaf]

oh not again

[quartz kindle]

btw, i copied the functions you guys posted so far, but apparently they are not finding any duplicates? let me douple check my data set lol

[neon leaf]

whats your data

make sure its not a buffer array and not a pure 2d number array

function generateTestData(size) {
    return Array.from({ length: size }, () => ({
        bytes_buffer: crypto.randomBytes(32)
    }));
}```

[quartz kindle]

nvm its just ben's function is not detecting duplicates

[neon leaf]

okay

this is the best I can do

function best_possible_optimizations() {
    const set = new Set()
    const buffer = Buffer.allocUnsafe(32)
    for (const entry of entries) {
        for (let i = 0; i < 32; i++) {
            buffer[i] = entry.bytes_buffer[31 - i]
        }
        const reversed = buffer.toString('binary')
        if (set.has(reversed)) {
            console.log("Duplicate found!")
        }
        const str = entry.bytes_buffer.toString('binary')
        set.add(str)
    }
}```

hex_username: 65.929ms
best_possible_optimizations: 42.636ms```

this only works if the data is proper though (32 bytes)

[frosty gale]

bros converting it to text binary

[neon leaf]

7ms faster

[frosty gale]

if js sets/maps supported buffers as keys i wouldnt even convert it to a string at all but ascii encoding is the closest i can get

[neon leaf]

they do support any object as key

[frosty gale]

i tried using a buffer and it seems to be by reference

[neon leaf]

but comparing only works if its the same reference

[frosty gale]

the keys should just be hashes under the hood so it shouldnt matter whether you give it a string or number or anything

but js just chooses to hash the reference address likely

[neon leaf]

robert  ~/projects  node sha1.js
chloe: 62.837ms
hex_username: 62.182ms
best_possible_optimizations: 19.6ms

mmmmmmm

looks like i found a way

function generate_unique_int_from_32_bytes(buffer) {
    let result = 0;
    for (let i = 0; i < 32; i++) {
        result = (result << 8) | buffer[i];
    }
    return result;
}

function best_possible_optimizations() {
    const set = new Set()
    const buffer = Buffer.allocUnsafe(32)
    for (const entry of entries) {
        for (let i = 0; i < 32; i++) {
            buffer[i] = entry.bytes_buffer[31 - i]
        }
        const reversed = generate_unique_int_from_32_bytes(buffer)
        if (set.has(reversed)) {
            console.log("Duplicate found!")
        }
        set.add(generate_unique_int_from_32_bytes(entry.bytes_buffer))
    }
}```

ok this is probably actually it

[lament rock]

Its the same as why [] !== []

[neon leaf]

I dont see any reason why anything could be more than 5ms faster than that

@quartz kindle

[frosty gale]

reusing the same buffer is a good idea

but your generate unique int i would imagine wouldnt be guaranteed to be unique 100% of the time

[neon leaf]

it is always unique

unless its over 32 bytes

it just bitshifts a number from the 32 bytes

[quartz kindle]

function generate_unique_int_from_32_bytes(buffer) {
    let result = 0;
    for (let i = 0; i < 32; i++) {
        result = (result << 8) | buffer[i];
    }
    return result;
}

wont that simply cut off most of the bytes?

[neon leaf]

lemme look at it again

[quartz kindle]

you're just taking the last 4 bytes basically

[neon leaf]

o wait yeah

function generate_unique_int_from_32_bytes(buffer) {
    let result = 0;
    for (let i = 0; i < 32; i++) {
        result = result * 256 + buffer[i];
    }
    return result;
}```

[quartz kindle]

thats basically a hashing approach, you need a good number hasher for that

[quartz kindle]

that will exceed number limit

[neon leaf]

hex_username: 1.029s
best_possible_optimizations: 661.657ms```

1000000 entries

[neon leaf]

hmm

[quartz kindle]

32 bytes into a single number is basically a bigint

256bit bigint

[frosty gale]

now it looks like the output hash can exceed the js 64 bit limit

[quartz kindle]

but you can try a number hasher like djb2 or fnv

[frosty gale]

so the numbers cant be reliable

[quartz kindle]

but its not guaranteed to be unique

[frosty gale]

yeah no hash is

with something like this id want it to be guaranteed to always be unique

if one hash happens to collide with one entry and have a false positive thats already not acceptable for 90% of use cases

[neon leaf]

then this is the fastest i have rn

[frosty gale]

looks like a set supports bigints too

so you can should be able to feed the buffer into a big int

[neon leaf]

bigint was 100% slower than my other code

[frosty gale]

thats guaranteed to always be unique

yeah because they use strings under the hood

theyre very slow

[neon leaf]

🤔

[quartz kindle]

not that much difference

but still

[neon leaf]

what did you change

[quartz kindle]

different approach, no strings

[frosty gale]

strings suck

what is your code

[quartz kindle]

function tim2() {
    debugger
    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(Buffer.isBuffer(entry)) {
                if(n < buffer.length -1) {
                    const next = Array(255);
                    next[entry[n+1]] = entry;
                    next[buffer[n+1]] = buffer;
                    current = current[buffer[n]] = next;
                } else {
                    console.log("duplicate found");
                }
            } else if(Array.isArray(entry)) {
                current = entry;
            } else {
                current[buffer[n]] = buffer;
                break;
            }
        }
    }
}

basically a byte rope

[frosty gale]

cursed code detected

but this is the path to performance in js

[quartz kindle]

not sure how reliable it is, ropes are hell to work with (as you can see from the debugger statement lmao)

[frosty gale]

i got 3ms faster by converting my for const loop to for i

gonna see if i can get rid of converting to strings or sets

[quartz kindle]

for..of is bad, avoid at all costs

[neon leaf]

(unless you use fancy things that implement Symbol.iterator)

[quartz kindle]

if you dont mind having your code a bit slower :)

[neon leaf]

still faster than forEach

[quartz kindle]

tbh i do use for..of quite a lot

[frosty gale]

wonder what this does under the hood https://www.npmjs.com/package/buffer-map

[quartz kindle]

for non-performance critical work

[neon leaf]

converts to string

in binary format

[frosty gale]

yeah to binary

this npm package is useless

you can implement everything it does in just a few more characters of code

not even extra lines

this is basically the is-odd package

[quartz kindle]

if you wanna squeeze some extra mileage from converting buffers to strings

use node's built in string_decoder

[frosty gale]

does this relate to the browser TextDecoder

ill try it though

[quartz kindle]

sort of

node:string_decoder is basically the decoder that is used under the hood for both buffer.toString and TextDecoder

but you get better performance if you use it directly for course, less checks and overhead

although is mostly focused on utf8, since its a more complex format

[neon leaf]

tf

chloe: 69.182ms
best_possible_optimizations: 41.754ms
best_possible_optimizations_2: 24.044ms
tim2: 34.596ms

js just died for a sec

its now same as the first again

[quartz kindle]

fluctuations :)

[neon leaf]

yeah string decoder made it ~5ms faster