DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.tw/irclogs/ Connected with DDNet's IRC channel, Matrix room and GitHub repositories — IRC: #ddnet on Quakenet | Matrix: #ddnet-developer:matrix.org GitHub: https://github.com/ddnet

Between 2022-12-14 00:00:00Z and 2022-12-15 00:00:00Z

[ddnet-maps:master] 1 new commit

8c42369 M Supernova - ddnet-maps

is there a way to create a vertex buffer or something similar in ddnet code for rendering my own list of triangle vertices? (like doing glBindBuffer and then glBufferData & glBufferSubData for updating, in opengl) (edited)

Generally you can add it to the backend. But ddnet is 2d don't expect any helper functions for projection matrices or similar.

06:53

Graphics threaded should have vertex buffer object. But eg index buffer is backend only So u defs have to write own logic

welp

07:00

thanks anyway

matrix bloat kek

You can't.

12:04

@noby wrote an antibot functionality for his FNG server, together with DDNet devs they factored that out into a shared library so it can be loaded in official DDNet server too. In order not to tell bot authors how we detect their bots the source code and binary for antibot are not available

12:05

You would have to implement the interface at https://github.com/ddnet/ddnet/tree/master/src/antibot to implement your own antibot solution

12:06

or ask noby nicely if you have a server you want to use it for

12:06

Basically you can use the incoming information to try and detect suspicious looking patterns of movement/actions

12:08

yeah, by default we ship the "null" implementation of antibot, which does nothing

12:08

and should be optimized away by your compiler

1

deen

@noby wrote an antibot functionality for his FNG server, together with DDNet devs they factored that out into a shared library so it can be loaded in official DDNet server too. In order not to tell bot authors how we detect their bots the source code and binary for antibot are not available

>drag it into IDA Pro and look at eyported functions

12:19

No one will know how it works. Yes. That's how this world works.

default

>drag it into IDA Pro and look at eyported functions

You don't have the binary though

Never say never. We have talented people around this community. I'm sure some of them could figure out how it works

Uhm actually, they didnt say never

3

Learath2

Never say never. We have talented people around this community. I'm sure some of them could figure out how it works

i would try for the luls

12:43

but probs wouldnt have results xd

the C code is already unreadable

12:43

best approach would be blackbox testing

its noby C

12:43

xD

12:44

im still learning erlang and i kinda have mixed feeling about functional programming

12:44

i think i prefer to think iteratively and imperatively tbh

12:44

too much brain usage

Some problems are easier to think about in a functional way. Some aren't. Which is why I'm not a fan of overly ideological languages

yeah

12:45

rust takes best from both

12:45

Gimme the tools for both and I'll pick the one that fits my situation

Learath2

Some problems are easier to think about in a functional way. Some aren't. Which is why I'm not a fan of overly ideological languages

a RPN calculator

12:52

xd

12:52

-module(calc).
-export([rpn/1]).

rpn(L) when is_list(L) ->
    [Res] = lists:foldl(fun rpn/2, [], string:tokens(L, " ")),
    Res.

to_num(N) -> 
    case string:to_float(N) of
        {error, no_float} -> list_to_integer(N);
        {F,_} -> F
    end.

rpn("+", [A,B|S]) -> 
    [A + B|S];
rpn("-", [A,B|S]) -> 
    [A - B|S];
rpn("*", [A,B|S]) -> 
    [A * B|S];
rpn("/", [A,B|S]) -> 
    [A / B|S];
rpn(X, Stack) -> [to_num(X)|Stack].

12:53

https://en.wikipedia.org/wiki/Reverse_Polish_notation

12:53

fits stack perfectly

Data oriented programming fits functional programming well too. E.g. ECS

Learath2

Never say never. We have talented people around this community. I'm sure some of them could figure out how it works

I mean you guys use sqllite, demos could be stored in the database (who knows). and if 80%+ of the position matches someones else's movement. -> sus

[ddnet/ddnet] Pull request opened: #6132 Make tileflag names consistent and intuitive

TILEFLAG_VFLIP -> TILEFLAG_FLIP_HORIZONTAL
TILEFLAG_HFLIP -> TILEFLAG_FLIP_VERTICAL

According to the native editor, the "Tiled" editor and image search, a horizontal flip should be associated with switching left and right, modifying the x coordinate. I did not just switch the letters H and V to create compiler errors where the original constants are used. Whenever I was working with tileflags, the naming caused me to have no idea what I was doing. I mostly had to res...

ur clueless

12:55

also sqlite is rly powerful

Ryozuki

ur clueless

isn't that intended?

12:55

should I know how your anticheat works?

ddnet can use mysql or sqlite

I could try to break it if you want. Always needed a gore bot. (edited)

god

12:56

u just wanna be a hackermans or show off

12:56

the worst kind of guy

nah I just need games to fuck around with (edited)

cool and i dont care

default

I mean you guys use sqllite, demos could be stored in the database (who knows). and if 80%+ of the position matches someones else's movement. -> sus

This sort of detection is actually surprisingly hard to implement. We store multiple terabytes of race data. It's not so easy to match

People do it for Trackmania

12:57

and they aren't even the devs of the game

12:57

but just part of the community

I didn't say it's impossible. I said it's surprisingly hard

Osu got the same stuff to detect replay bots

way less inputs

inputs don't matter at first glance for me in the case of teeworlds

12:58

I would go for position

12:58

and then input

You first need to come up with a way to index the data. An efficient tree like structure where you can look for substrings of "gamestates"

current demo time/tick? (edited)

Which is actually a huge issue. You can't recreate the gamestate at a given point without running through all previous ticks

then how do you play the ghost data while a player plays a map

12:59

isn't it the exact same thing

12:59

just that you now look at the position difference factor throughout the run (some math guy probably can come up with the same stuff they use for chess anticheats - how far away you are from the "best" move, or also called "accuracy"). but I believe you that it is hard to index the database. (edited)

Learath2

You first need to come up with a way to index the data. An efficient tree like structure where you can look for substrings of "gamestates"

quad tree?

13:03

achtually

13:03

in chess its defined by centipawn loss

13:03

default

just that you now look at the position difference factor throughout the run (some math guy probably can come up with the same stuff they use for chess anticheats - how far away you are from the "best" move, or also called "accuracy"). but I believe you that it is hard to index the database. (edited)

we don't know the "best" move so far

If you only want to match position it is a bit more doable. We store entire server sessions. So no demo per individual run. Therefore the first objective would be to split those into sequences of positions per run. After that you need to find a data structure that can efficiently look for substrings

achtually we got someone that watched the leeman debate

13:03

called ryo

Learath2

If you only want to match position it is a bit more doable. We store entire server sessions. So no demo per individual run. Therefore the first objective would be to split those into sequences of positions per run. After that you need to find a data structure that can efficiently look for substrings

Something like ropes maybe

and we probably don't want to produce software to do it I guess ^^

default

achtually we got someone that watched the leeman debate

idk what debate is actually xd

oh I mean hans niemann

13:04

don't know how I got to leeman (teeman) (edited)

ah lol

13:04

well i actually play chess from time to time

13:04

so i know before this drama

I'm not even sure a sane best move could even be defined for a teeworlds run. Especially non-solo runs

I mean how it turned out

13:04

he didn't even cheat over the board

13:04

so magnus is just salty, like you, ryo

Learath2

I'm not even sure a sane best move could even be defined for a teeworlds run. Especially non-solo runs

but usually bot runs replay something that have been played before?

13:05

or how do they work

13:05

haven't hacked around with teeworlds yet

Learath2

I'm not even sure a sane best move could even be defined for a teeworlds run. Especially non-solo runs

I can easily define it

but calculating it is another story

just with an idea of using the hook function to stay at the same place in the air

okay, that's enough

13:06

please continue your cheat discussion in DMs

13:06

or stop them entirely, if possible

Anyway, problem with position only matching is map constraints. There are a lot maps that force a sequence of positions, so you need to discard any matches that are due to those. Which is where I start to not have anymore ideas on what to do

Learath2

Anyway, problem with position only matching is map constraints. There are a lot maps that force a sequence of positions, so you need to discard any matches that are due to those. Which is where I start to not have anymore ideas on what to do

90% match throughout the whole map would be insane

You could keep track of most common n-grams of symbols like they do in natural language processing, I think there are algorithms that do that

even with such meta paths to take, your height won't match, you are a bit faster or slower here and there (edited)

lets ask doctor Chairn for a research paper

heinrich5991

or stop them entirely, if possible

are you a Discord ToS fanatic? they changed how that works btw.

13:08

you can have talks about it now (edited)

I'm a "don't develop cheats for ddnet" fanatic

fair enough

or preferably for any multiplayer game

we wouldn't be a good match imo (edited)

default

even with such meta paths to take, your height won't match, you are a bit faster or slower here and there (edited)

There are truly constrained movements in some maps. E.g. hold d and jump when you hit the wall or speeder tunnels in freeze

yea but that still won't explain a 90% accuracy with anothers replay

default

we wouldn't be a good match imo (edited)

do you realize that some people derive more fun from a game when there are no cheats?

Anyway, it's hardly an unsolvable problem. It's just a hard problem

heinrich5991

do you realize that some people derive more fun from a game when there are no cheats?

like me playing ddnet just because it is fun together with friends? can't imagine. but then there are people that do break games to then report stuff to the game devs, or that do break games just to get better at it, or that break the game to profit off of it -> selling, boosting ego, etc. (edited)

Much harder when you consider the multiple terabytes of data you need to navigate through efficiently

13:10

Checking how much one run matches another is relatively easy in comparison

Ryozuki

lets ask doctor Chairn for a research paper

Chairn would probably know a data structure or two that might help

I assume Chairn wrote the anticheat then?

13:12

or is he just a known math/dev guy in this community? (edited)

default

like me playing ddnet just because it is fun together with friends? can't imagine. but then there are people that do break games to then report stuff to the game devs, or that do break games just to get better at it, or that break the game to profit off of it -> selling, boosting ego, etc. (edited)

I see that breaking games can be fun. in order to not destroy the fun of others, it'd be nice to make sure not to spread the resulting cheat

Monte Carlo

default

or is he just a known math/dev guy in this community? (edited)

Just a guy known to be good at this sort of theoretical stuff

heinrich5991

I see that breaking games can be fun. in order to not destroy the fun of others, it'd be nice to make sure not to spread the resulting cheat

https://secret.club/ doesn't do that though

Blog about reverse-engineering, hacking and breaking your software in every way imaginable.

Learath2

Just a guy known to be good at this sort of theoretical stuff

very cool, is he still active?

1

secret.club does not publish cheats that you can just use to destroy the fun of others

default

https://secret.club/ doesn't do that though

I don't see them peddling ready cheats. Messing around with intrusive anticheats and drm is a whole different thing

I also read that blog, it's very cool

13:14

in fact, I'd guess most people from there would agree that breaking multiplayer games is not cool. I know at least one person writing blog posts there

I mean obviously. Daax is helping anticheat solutions.

13:15

He is obsessed with Byfron. Maybe even works for them atm.

13:15

And surely had one or both hands taking part in or on Vanguard. (edited)

Besides, our flavor of anticheat is a different approach anyway. Most of these people are interested in clientside anticheat solutions, which is a losing battle anyway

Train ai with bad and good gameplay

Jupstar has been banned for 5 minutes (bad play)

unclear if "AI" would be good at this

13:17

it's good at distinguishing known classes from one another

heinrich5991

unclear if "AI" would be good at this

not in realtime, but letting it go through the whole dataset you got could be worthwhile (edited)

I'd bet it's one of the best fits for ai

Anyway, fun discussion. I need to get back to studying the most mind numbing topic known to mankind

Jupstar ✪

I'd bet it's one of the best fits for ai

my bachelor thesis was about using AI and related methods for intrusion detection

13:18

it's not a good fit

You want to tell me you are not at work right now chatting in the ddnet discord server?

I'm indeed not at work

Ryozuki

i think i prefer to think iteratively and imperatively tbh

I was a functional programming fanboy with Haskell, then imperative with Nim. In the end they are just tools with different tradeoffs, none offer magic bullets.

13:19

And offering all kinds of paradigms and features also causes problems with huge complexity and teams not understanding the shared code base they are working on

[ddnet:master] 2 new commits

a4867d2 Make tileflag names consistent and intuitive - Patiga 609178d Merge #6132 - bors[bot]

13:23

[ddnet/ddnet] Pull request closed: #6132 Make tileflag names consistent and intuitive

heinrich5991

it's not a good fit

Vacnet cries

it's probably okay for detecting known cheat patterns, like spinbotting, etc.

13:24

but it's not good for detecting things you haven't thought of

That's for all cheats xd

heinrich5991

but it's not good for detecting things you haven't thought of

The story of AI so far, in other areas too I guess

Are humans good at it?

13:27

At least the potential of ai

Jupstar ✪

Are humans good at it?

human brain is amazing

Ryozuki

human brain is amazing

Team machine

13:38

It's the logical step in nature. It built us to built something better (edited)

a human can optimize the algorithm a computer uses and cut lot of computation time, but a computer can only do computations fast

13:39

erlang rly do be ugly

13:39

but luckily there is elixir

13:39

which ill check sooner than later

13:39

uses the same VM BEAM

13:40

oh yesterday i saw this

13:40

https://www.ponylang.io/

13:40

its a serious project btw

13:40

maybe safer than rust

heinrich5991

but it's not good for detecting things you haven't thought of

Have you asked the AI how to detect something you say an AI can't detect? If you describe your problem good enough, it could deliver results. (edited)

https://www.ponylang.io/discover/#what-is-pony

Discover - Pony

@Ryozuki do u think humans are the peak? ^^

no

13:42

well right now

13:42

on earth yes

13:43

as in the most developed creature

13:43

being aware of ones own death

Without bacteria we would die

and

13:43

a human to me involves all the microcosmos needed for a human to be

13:43

xd

13:43

without the cell the human would die

13:43

and without the human the bacteria would be dead

13:44

ironic

default

and without the human the bacteria would be dead

?

13:45

How so

i think some bacteria are only found in humans

13:45

stomach

13:45

or mammals

13:45

idk

Yes

13:46

Have u tried stop to eat?

the fact u can discuss this

13:46

proves we are the peak

13:46

kek

I mean it is impressive how this world works innit bruv

13:47

the world is basically (whatever) black (really is) without our eyes and only reflects light (edited)

and "this" is being discussed using a advanced tech (edited)

So the tech is the impressive thing?

well to me the fact we able to do the tech is ye

But we were only able to be born by bacteria

AI is just automating automation

13:48

its still made by us

13:48

i think so

So it's ultimately made by whatever was before humans

i wonder

13:48

gotta know if big bang is true

13:49

what was truly there

13:49

before the big bang

13:49

since big bang is the making of "things"

13:49

itself

world 1.0 from our perspective

13:50

and that world was world 2.0 of some other world

13:50

or that one egg-theory is true, and we all are the same person, evolving into a god once we lived through every life

kurzgesagt

it is a serious thing

13:51

they just animated it

i know the story

13:51

the animation is dope

We will never know, because we are nothing to this timeline

13:52

Whatever time even is.

relevant and epic video

13:53

https://www.youtube.com/watch?v=uD4izuDMUQA

TIMELAPSE OF THE FUTURE: A Journey to the End of Time (4K)

I find it plausible though that you can't create any power, nor destroy it

13:53

and that this one dude that explained that an apple would recreate itself in a closed chamber with only its own particles given long enough time hopefully says the truth, based on that thesis (edited)

13:54

i am happy to relive this life in the future after an incredible big time window in world 2.0 (from our perspective) (edited)

u wont remember this

I know

so this will only be your "only" life

that's the fancy stuff

hence why ppl turn to religion

14:01

nothingness after death is scary

relgion is only mankinds invention

14:02

but you shall let the people believe in whatever they want

i mean lot of stuff is mankind invention

14:03

but i guess u wanted to mean its not real

nah that would be an insult

14:03

maybe there is something because us typing here wouldn't be logical

im not religious btw

14:03

xd

because out of nothing, not everything can exist. but then you can ask again what created whatever was before it. so maybe life is just illogical. (edited)

i do believe stuff that happens is bound to happen

14:04

hard to explain

14:05

but everything is fixed on the timeline

14:05

since its based on previous inputs

14:05

maybe smth https://en.wikipedia.org/wiki/Chaos_theory

Chaos theory is an interdisciplinary area of scientific study and branch of mathematics focused on underlying patterns and deterministic laws of dynamical systems that are highly sensitive to initial conditions, and were once thought to have completely random states of disorder and irregularities. Chaos theory states that within the apparent ran...

14:06

Chaos theory states that within the apparent randomness of chaotic complex systems, there are underlying patterns, interconnection, constant feedback loops, repetition, self-similarity, fractals, and self-organization.[2] The butterfly effect, an underlying principle of chaos, describes how a small change in one state of a deterministic nonlinear system can result in large differences in a later state (meaning that there is sensitive dependence on initial conditions).[3] A metaphor for this behavior is that a butterfly flapping its wings in Brazil can cause a tornado in Texas

14:06

Small differences in initial conditions, such as those due to errors in measurements or due to rounding errors in numerical computation, can yield widely diverging outcomes for such dynamical systems, rendering long-term prediction of their behavior impossible in general.[7] This can happen even though these systems are deterministic, meaning that their future behavior follows a unique evolution[8] and is fully determined by their initial conditions, with no random elements involved.[9] In other words, the deterministic nature of these systems does not make them predictable.[10][11] This behavior is known as deterministic chaos, or simply chaos. The theory was summarized by Edward Lorenz as:[12]

14:06

interesting stuff kek

yea you insulting me before will cause huge events chaos theory ~~proves~~ explains it. it's the little things that matter. (edited)

14:07

Ray

BASEDDEPT

14:09

u already blended to the #developer

14:09

it was always about discussing metaphysics

14:09

than programming

14:09

TOOBASED

14:10

talking about interesting stuff

14:10

https://oimo.io/works/life/

14:10

enter here

14:10

and zoom out

14:10

beautiful

14:10

but dont just zoom out a little

14:10

keep zooming out

14:10

and keep

14:11

infinitely

14:11

mesmerizing

14:13

https://conwaylife.com/wiki/OTCA_metapixel

14:13

it uses this btw

yea if this is game of life, all that little logic zoomed out will just create the same game again, first whole pixels, then those ships, reappearing logic, etc. (edited)

In the beginning I didn't think it was possible without cheating something but finally found a way to "build everything perfectly" Due to the inifinite recursion, each level is aperiodic for both space and time, and you'll never see the same pattern no matter how much you scroll

Ryozuki

but everything is fixed on the timeline

not quantum mechanically

Ryozuki

lets ask doctor Chairn for a research paper

definition is easy, move that grants you on the overall race (not just the part) the best time

yea but that isn't enough

14:30

people can pick the replay of their friend

14:31

and stuff like that should get detected as well, if not already

now, it's just a probably NP hard problem as the sum of optimal moves for each parts isn't probably the best time for the whole map

default

very cool, is he still active?

hello there

Chairn

now, it's just a probably NP hard problem as the sum of optimal moves for each parts isn't probably the best time for the whole map

don't give bot devs ideas to break world records

14:31

kek

TIL

14:32

https://gleam.run/

The Gleam programming language

14:32

The power of a type system, the expressiveness of functional programming, and the reliability of the highly concurrent, fault tolerant Erlang runtime, with a familiar and modern syntax.

14:32

this might be more for me xd

14:32

i love types

then you will love object oriented programming languages (edited)

14:32

if you are only stuck in web dev languages (edited)

OH

14:33

implemented in rust even

14:33

https://github.com/gleam-lang/gleam

GitHub - gleam-lang/gleam: ⭐️ A friendly language for building type...

⭐️ A friendly language for building type-safe, scalable systems! - GitHub - gleam-lang/gleam: ⭐️ A friendly language for building type-safe, scalable systems!

default

don't give bot devs ideas to break world records

well, if they can solve this problem, they can probably help to more important business

I mean someone probably does it already

14:34

selfhost a server so data doesn't get sent to ddnet

14:34

let some AI figure out the fastest strat for every challenge

@Mr.Gh0s7

justatest

2

but the fastest strat for each challenge is not the whole map best time. example: when there's a cut

Ryozuki

i love types

Try typescript troll

troll

Chairn

but the fastest strat for each challenge is not the whole map best time. example: when there's a cut

Yea. You are right. Then make it go through all the open spots and see if it finds some cheese. Or let the pathfinder find potential shortcuts. By not eliminating freezes by default. Shouldn't it find an even faster way through thousands of generations eventually? If you are not limiting it to a set path (through all the challenges in a row). That isn't a problem I want to deal with though. I would just improve upon the best existent time. Possibly by replay editors. Similar to the one I made for Trackmania. But reusing other players' plays should be detected already. This is why I thought I was smart when I decided to improve my own replay (unsent to the server) with my own plays (for another game). (edited)

noby

@Mr.Gh0s7

I'm waiting for xmas vacation to continue working on it :D

oo

default

if you are only stuck in web dev languages (edited)

bruv

default

then you will love object oriented programming languages (edited)

and more bruv

isn't it one?

14:42

all that bs sweet talk sounds like any for web dev developed language

you mean gleam?

for example

14:43

before that you had one starting with l or i

i mean there is only one web lang and its js

14:43

but i guess u dont like high level perfomance languages?

which people made typescript for, for example which was worth using until js got type-safe themself, recently (edited)

gleam is a functional language that runs on BEAM, the erlang VM

14:44

and types are not just"object oriented programming"

14:44

actually OOP gets many things wrong

14:44

composition over inheritance

14:44

rust gets it right imho

how bad is C# then

i try to avoid it whenever i can

14:45

but sometimes i gotta use it to mod ff14

14:45

sadge

sadge indeed

14:45

that you don't like it besides that (edited)

trash kek

14:45

why would i use that when i have my beloved rust

wdym, it can get compiled to native now

its only famous cuz m$$ is behind it

which allows me to write internals

14:45

in fucking c# did you imagine that a few years ago (edited)

and

14:46

c# unsafe code is unsafe

14:46

i can do the same in safe rust code

Ryozuki

c# unsafe code is unsafe

same with C++

not same with rust

you don't know that something is messed up until you run it

i never mentioend c++

yea but that's hand holding

lmfao?

14:46

XD

14:47

i need to tone myself down

14:47

i get too heated

14:47

brb

Ryozuki

i never mentioend c++

I did now. Because Rust was something a lot of C++ devs moved to.

14:47

and them not allowing you to code unsafe stuff is hand holding for me

u definitly dont know shit about rust

14:48

cuz what u said isnt true

14:48

i recommend u learn rust

14:48

it will teach you a lot

14:48

its incredible

nah

bad mentality

14:48

dont say stuff about it then

I won't mate with that ~~crab~~ crap

and arguing with people who havent even tried it is the worst

true

its like a profesional doctor talking with a person who heard x on the streed about some medicine

on the topic of rust is there a more elegant way to do this? cur += i32::from_str(line).unwrap();

street*

14:49

cur += 32 line.parse().unwrap();

14:49

if ur parent function returns a result

14:49

u can change unwrap() for ?

14:50

cur += line.parse()?; (edited)

is that possible by default

14:50

or do you have to write extension functions in rust? (edited)

14:50

to use functions on an object

in rust u only have structs

14:50

and you "implement them"

14:50

its half like oop but without the bad stuff

so whatever line is you had to implement yourself?

line here is a &str

14:51

a string

then how do you write a function for that struct if it is not yours

.parse() is a function you get if your struct implemented the trait From

14:51

you make a trait with that function

14:51

and implement the trait for the struct

aha

a trait is like a interface

14:52

for example, this is my trait for api endpoints

Image attachment

14:52

Image attachment

14:52

and implement it like this

14:53

Image attachment

14:53

and my api client can execute any struct implementing Endpoint

and you think this is better than for example BetterEndpoint : Endpoint or some function in a class that got a this parameter? (edited)

composition is the best

14:54

it falls in place so well

14:54

try rust

i get the first being worse. but the second is nothing else. from the outside. (edited)

and u will know why

14:54

the rust book has a chapter on OOP programming

14:54

ur style

14:54

https://doc.rust-lang.org/book/ch17-00-oop.html

Object Oriented Programming Features of Rust - The Rust Programming...

14:54

good read

14:55

rust type system is rly powerful

14:56

and more everyday

14:56

since it takes stuff from functional languages

14:57

but the best is the fearless concurrency and safety

14:57

kek

14:57

thanks for coming to my ted talk again

thanks for teaching me how to read rust code, but won't gonna code in it

u should learn it

14:58

if ur passionate about programming

14:58

u can only get positives

14:58

learning never bad

I rather code in asm than learn rust and code the same thing in it (edited)

ok

14:59

clown_gang

woah

15:01

I will use Rust actively when Microsoft starts to use it

15:02

But for me, it is nothing different than other programming theories combined into a language that its users think is better than some other language that the people using that language think is good enough for their needs. Under Windows, C-family code will return the same thing as Rust. I hope. Otherwise, Rust creates tons of overhead, or even worse, runs in a VM. (edited)

default

I will use Rust actively when Microsoft starts to use it

they do

sauce

They part of rust foundation

that, maybe

What's ur primary language?

any C language goes

C#?

pure C, never

15:07

C#/C++ works for me

If u ever encounter using c# use rust instead

C# for external stuff until they introduced native compiling since NET 7.0. C++ for whatever C# can't do. Which is hard to find something for nowadays. Because C# can even go Kernel using a Wrapper VollRagm introduced in our community. (edited)

Microsoft build a os in c# that does not directly use a kernel. But c# too much bloat

What type of bloat

Well garbage collection is bloat for example

that is an issue when you run C# in its little JIT VM

15:12

and I never had issues with Garbage Collection. Only when it comes to graphics stuff. Because graphics stuff gets excluded by it.

15:12

Which makes you have to set everything to null to force it onto those objects. Or otherwise you will leak memory most likely, as soon as you reinitialize something graphical. (edited)

U had no issues bcs it's made for easy use

15:13

It eats ram and is slower in most cases than normal allocation deallocation tho

thing is, you don't have to use GC. you can use it. (edited)

15:15

if that is a nightmare for every memory management freak

What are u even developing

home or at work?

Wherever u have to use it

default

sauce

i think they use it in some kernel parts, they also made the windows crate

15:20

a crate is a library

15:21

https://learn.microsoft.com/en-us/windows/dev-environment/rust/rust-for-windows

Rust for Windows, and the windows crate

Using the windows crate, and calling Windows APIs.

Yea you know why

@Jupstar ✪ he is a c# user

15:21

no point in talking

15:21

cant convince em

because Rust wouldn't be able to communicate with the OS like at all

they usually are m$ bootlickers

that's why Linux had to create a crate too

not at all xD

15:21

by that i mean that microsoft themselves made the crate

15:21

previously it was a user made one

15:22

as in a random person made the library

default

that's why Linux had to create a crate too

nevermind. but they actively added support for it in the kernel in version 6.1.

15:22

it wouldn't be able to do it by itself

nono

15:23

u are rly confused

15:23

what they added in the kernel 6.1 is support to use rust in the kernel itself

15:23

to help make drivers with it

15:23

u can use and call linux apis before 6.1 lol

pandaeyes

WorrySnailEyes

15:24

anyway no wonder you have that actitude

Remember ryos words. C is just a protocol xd

unwilling to learn

15:24

its pretty common in c# devs

15:24

i found this on the ff community

15:24

there is a marketboard website

15:24

they used c#

15:25

and its slow

15:25

so im making my own in rust

15:25

to flex

yea but it's not slow because of C#

15:25

it's slow because of wrong use of C#

nah its slow cuz they do wrong db management xD

15:25

but anyway its slow

15:25

and i will dab on them

15:26

altho ultimately

15:26

rust is way easier to paralelize

15:26

and do concurrent

15:26

simply because u dont need to be worried about race conditions

15:26

only deadlocks

15:26

and its as perf as C

@default even if i don't care what u use. Doesnt it suck that, no matter how hard u try, your code will execute at least twice as slow than the same code in a system level language.

ferrisBongo

15:27

i would use c++ before c#

again, I compile it into native machine code since C# 7.0

@default can u do runtime reflection in compiled c#?

I would have to do research on the terminology here first

Nice but how does that solve the bottlenecks that come with the design

I know what reflection is

15:28

but runtime reflection got me all dizzy

Jupstar ✪

Nice but how does that solve the bottlenecks that come with the design

name me a bottleneck

i mean reflection itself is runtime mostly i would say xd

Yea so probably no. Though you can call the API for that. I just never tried.

u can do compile time reflection in rust

15:29

https://github.com/dtolnay/reflect

GitHub - dtolnay/reflect: Compile-time reflection API for developin...

Compile-time reflection API for developing robust procedural macros (proof of concept) - GitHub - dtolnay/reflect: Compile-time reflection API for developing robust procedural macros (proof of conc...

I dereference an address to a function if I want to use a function from a game now Did you mention you hate unsafe code? Yea. That's peek performance unsafe code. (edited)

15:30

You crash if you don't get the types correct that the function expects from you

well ur mostly describing some kind of FFI

15:31

Image attachment

15:31

csharp 7 is slower

15:31

with AOT

15:31

i think its ur thing

15:31

native

15:31

@Jupstar ✪

15:32

looks like AOT uses less memory

15:32

but its slower

15:32

Image attachment

15:32

peak mem 1mb for rust

15:32

xd

Yeah look at that ram

https://programming-language-benchmarks.vercel.app/rust-vs-csharp

Rust VS C# benchmarks, Which programming language or compiler is fa...

benchmarks,benchmark,performance,fasta,helloworld,mandelbrot,nbody,spectral-norm,binarytrees,coro-prime-sieve,edigits,fannkuch-redux,http-server,json-serde,knucleotide,lru,merkletrees,nsieve,pidigits,regex-redux,secp256k1,Rust,Rust lang,C#,C# lang

15:34

a thing GC is good at is binary trees

15:34

but they still use shitload of memory

15:34

Image attachment

15:35

kek

15:35

Image attachment

15:36

https://www.mn.uio.no/ifi/english/research/groups/psy/completedmasters/2020/gaarde/masterthesis-gaarde.pdf

ItS tHe ImPlEmEnTaTiOn

lol there is a thesis

default

ItS tHe ImPlEmEnTaTiOn

i can tell u they did their best efforts

15:37

u can submit ur solution

15:37

btw

Gc is simply big problem sometimes.

its only good when dealing with trees

Most times xd

and even so its mostly not noticable

15:38

but they use more memory always

15:38

like 400% more

But in the end u can still use a gc in system level languages too

true

So i could imagine u can beat c sharp

i wonder if they used a arena allocator for the tree problems

If it really comes from this

why do you keep mentioning GC

Does it also beat c here?

when you don't even have to use it

15:39

what if there is no GC

@default but the non GC c# is slower

15:39

as u have seen

15:39

xD

default

what if there is no GC

rust is like a safe C++ without GC

15:40

its a game changer

Not sure actually if you can see it even from your sweet tables

15:40

because compiling AOT doesn't automatically mean GC is disabled

AloneBlu

in fact people still use GC for the native stuff for whatever reason

What do u use?

15:41

I didn't read all

15:41

Win api?

yes pretty much

The compiler probs not allowed to optimize them away

15:42

Can u access the stack for objects?

explain

Else this could be a place where compiler can optimize gc away

15:43

Bcs the lifetime is known potentially

15:43

Store objects on stack instead of heap

never done that

15:44

no experience

But with winapi calls it might not do that

15:44

Bcs of side effects

@Jupstar ✪ i think its mostly heap based

15:44

unless its a non boxed primitve

15:44

primitive

default

never done that

another reasonn to learn C or Rust

15:45

not knowing what the stack or heap is

i know what a stack is

15:45

just not if my program uses it or the other one

most of the time if u type new in c# its heap

15:45

iirc

^^

yea but then you could force use stack

MyStupidOOPClass x = new MyStupidOOPCLAS()

surely

HEAP

15:46

oops

15:46

capslock

default

yea but then you could force use stack

hard on a language such as c# or java

i mean give me an example in your language

15:47

i then can think if I can implement it in my language

MyStupidOOPCLAS a;

15:47

No new

yea that is possible

@Jupstar ✪ a string being on the heap or stack in c# is an implementation detaiñl

15:47

detail

but it will scream that it was never intitialized

thats how much they care about perf

so it is null or the default object

https://stackoverflow.com/questions/12727821/net-stack-and-heap-what-goes-where-when-i-declare-a-string

.NET stack and heap, what goes where when I declare a string?

If I execute this line I create a string which is a reference. string mystring = "Hello World" Is variable mystring in the same context as the object I declare it? And the data "Hello World" on ...

Ryozuki

thats how much they care about perf

It's the fastest high level language

15:48

But still shower than rust cpp and c^^

Jupstar ✪

It's the fastest high level language

is it?

15:48

what about golang

15:48

kek

@Jupstar ✪ also modern java might outcompete c# iirc

15:48

with the new GC impls

Ryozuki

is it?

Maybe Julia if that's even considered high level it some less known. But pretty fast

15:48

Or

Ryozuki

@Jupstar ✪ also modern java might outcompete c# iirc

No java even bigger bloat xd

Image attachment

15:49

i like how u deleted ur comment

15:49

before i posted this

15:49

kek

15:49

modern java got optimized a lot

yea there is the program

15:50

and then the runtime

15:50

that is like 800mb

Image attachment

15:50

llook the ram on AOT

15:50

they clearly dont use GC

15:50

yet slower

peak mem for aot

15:50

impressive

15:51

well not really

15:51

was obvious

doesnt beat rust tho

15:51

more control = easier to optimize

15:51

u cant win with a gc

15:51

its simple

15:51

but u dont have to worry about losing ur job

15:51

since m$ is behind c#

do you want me to tell you that rust is indeed better

it would be good but not required

15:52

cuz i already know

15:52

BASED

yea but you lied to yourself

wdym

you actually kept the discussion about something you knew going

15:52

with someone who never used rust

15:52

check mate

oh cuz i love discussing about this

15:52

@Jupstar ✪ knows

15:52

blobwhistle

Rust is indeed better than C#. You proved it with various code tests.

nice

Xd

now u have to learn rust

15:54

and u will join the club with @ReiTW

Nah g

ferrisBongoHyper

My parents house is strict. No Rust.

u mean work?

15:55

dont limit urself cuz work

15:55

learn it cuz u want

not work

15:55

who cares about work

then what

family tree

does ur parent beat u

15:55

if u dont use c#

if I would use Rust

15:55

maybe

is he a microsoft employee?

kek

monkalaugh

nah but Siemens

boring corpo

15:56

i recommend this read

15:56

http://www.catb.org/~esr/faqs/hacker-howto.html

15:56

to get the right mindset

15:57

note that hacker is not used in the kids sense

15:57

as u may know

Maybe I know

15:58

Though Mr.Robot was a rather realistic TV show.

There is another group of people who loudly call themselves hackers, but aren't. These are people (mainly adolescent males) who get a kick out of breaking into computers and phreaking the phone system. Real hackers call these people ‘crackers’ and want nothing to do with them. Real hackers mostly think crackers are lazy, irresponsible, and not very bright, and object that being able to break security doesn't make you a hacker any more than being able to hotwire cars makes you an automotive engineer. Unfortunately, many journalists and writers have been fooled into using the word ‘hacker’ to describe crackers; this irritates real hackers no end.

>

The basic difference is this: hackers build things, crackers break them.

15:59

hackers are for example, people who contribute to ddnet

15:59

literally

default

Though Mr.Robot was a rather realistic TV show.

it isnt, but the serie is rly good

15:59

i watched it all xD

15:59

love the actor

15:59

dead stare

Ryozuki

it isnt, but the serie is rly good

it isn't?

realistic

like the end. yes.

16:00

unrealistic.

It's completely realistic when it comes to the techniques and the technology, but it's very fictional when it comes to stuff like two people planning and executing a huge hack in a matter of days. So no, in practice it's not possible for someone to be as good as Elliot, although with enough time, people, skill, managing and luck it's possible to a group of hackers to execute most of the stuff he and the rest of fsociety does.

16:02

actually NSA did more amazing stuff than what this series shows

16:02

they hacked air gapped turbines in a nuclear facility to not operate correctly in iran iirc

16:02

i can recommend u a rly good pdocast

darknet stories?

yeah

16:02

best onen tbh

yea

i used to listen to it at the gym xd

i listened it through once

16:03

stopped for whatever reason

16:03

couldn't wait anymore for the next episode

the thing is we only know about their failures

16:03

an they are still amazing

16:03

the successes will never be known

16:03

cuz by becoming known they failed

EZY

thats a good quote

16:03

i should become a writer

16:03

ezpepe

that's their quote probably

16:04

i heard or read it somewhere (edited)

in case there is probably lurkers reading this

16:04

https://darknetdiaries.com/

Darknet Diaries – True stories from the dark side of the Internet.

A podcast featuring true stories from the dark side of the Internet.

16:04

here is the podcast

16:06

and what i mentioned is

16:06

Stuxnet

16:06

https://darknetdiaries.com/episode/29/

Stuxnet – Darknet Diaries

Stuxnet was the most sophisticated virus ever discovered. Its target was a nuclear enrichment facility in Iran. This virus was successfully able to destroy numerous centrifuges. Hear who did it and...

16:06

https://en.wikipedia.org/wiki/Stuxnet

Stuxnet is a malicious computer worm first uncovered in 2010 and thought to have been in development since at least 2005. Stuxnet targets supervisory control and data acquisition (SCADA) systems and is believed to be responsible for causing substantial damage to the nuclear program of Iran. Although neither country has openly admitted responsibi...

Is there an option to disable this, if not, can it be removed? It's not really useful at all

Image attachment

oh god, not again

Image attachment

nothing interesting

16:19

just a talk about the universe, something game related, and then two people rambling with me why Rust is good.

yup, just flew over, most of it was just you with Ryo or Jupstar on languages

could have gone into #off-topic. but some devs aren't that strict.

#developer is also #off-topic but reserved for developers

[ddnet/ddnet] Pull request opened: #6133 Remove spammy log (thanks Skeith)

Checklist

[ ] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-ad...

Skeith

Is there an option to disable this, if not, can it be removed? It's not really useful at all

done ^

Thank you

[ddnet/ddnet] Issue opened: #6134 Render tune settings when using entities mode

It's kinda annoying when there's a tune zone and you don't know what it does. I think we could render the tune settings when using entities mode in a not too annoying way rather than having to open the editor.

[ddnet/ddnet] Issue opened: #6135 On macOS

On macOS 13.1 different command line tools are mixed, leading to a compilation failure: ``` [...] -- Found ZLIB: /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX13.1.sdk/usr/lib/libz.tbd (found version "1.2.11") -- Could NOT find OpenSSL, try to set the path to OpenSSL root folder in the system variable OPENSSL_ROOT_DIR (missing: OPENSSL_CRYPTO_LIBRARY OPENSSL_INCLUDE_DIR) -- Could NOT find Crypto (missing: CRYPTO_LIBRARY CRYPTO_INCLUDEDIR) -- C...

17:09

[ddnet/ddnet] Pull request opened: #6136 Also ignore sqlite-shm and sqlite-wal files

See https://www.sqlite.org/tempfiles.html

Checklist

[ ] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed no physics that affect existing maps
[ ] Tested the change with [ASan+UBSan or valgrind's memcheck](htt...

[ddnet:master] 2 new commits

d5cf363 Remove spammy log (thanks Skeith) - def- 75d0735 Merge #6133 - bors[bot]

17:20

[ddnet/ddnet] Pull request closed: #6133 Remove spammy log (thanks Skeith)

[ddnet/ddnet] Pull request opened: #6137 Improve move_sqlite script to adapt timezone

Fixes #6105 This change does some more stuff:

change the filename to include the time with seconds to be able to run the script easier more often
Fix that _backup tables are only moved after 1h and not immidiately in UTC+1
Add a parameter to configure the number of minutes the backup
Add parameter to change timestamps to be in localtime
consider the timeout in num_transfer calculation
remove rows in _backup tables to not print in them in the log
rename tables to TABL...

https://svelte.dev/blog/announcing-sveltekit-1.0

Announcing SvelteKit 1.0

Web development, streamlined

17:41

i was using this on some projects

17:41

poggers

17:41

bye reactjs

[ddnet:master] 2 new commits

6581958 Also ignore sqlite-shm and sqlite-wal files - def- 601ebb5 Merge #6136 - bors[bot]

18:24

[ddnet/ddnet] Pull request closed: #6136 Also ignore sqlite-shm and sqlite-wal files

Image attachment

18:50

Lmao

it escalates xd

xd

19:03

i send that here

19:03

some day ago

19:03

sent

yes u always fastest

19:05

i dont even read news anymore, they slower than u

lulEZ

I wonder how they check if it is from ChatGPT. When it is complete bs?

maybe they simply ask it

I think it's to prevent question such as: "Why would AI write this code to do X when it could simply do Y?"

They nerfed openai too much

19:35

Before i could ask it to act as a browser console but now it cries out "im not a browser im an ai i cant do that"

19:36

"i was not trained for thaaat"

[ddnet:master] 2 new commits

60b4708 Improve move_sqlite script to adapt timezone - Zwelf 31fa4f4 Merge #6137 - bors[bot]

19:42

[ddnet/ddnet] Issue closed: #6105 sqlite timestamp entries don't contain correct timezone

19:42

[ddnet/ddnet] Pull request closed: #6137 Improve move_sqlite script to adapt timezone

how do i check if a packet is actually being sent?

Test on local server (edited)

Alexander

"i was not trained for thaaat"

Easy bypass

Image attachment

Anime.pdf

Test on local server (edited)

But actually if you dont care if it was received you can do it with some tool like wireshark

GitHub

Click to see attachment 🖼️

@Chairn well wouldnt that defeat the point of oco map

well, a small sacrifice

there should also be a way to toggle some tiles in entities since on some maps switches, speed up tiles and other things cover freeze tiles or tps

[ddnet/ddnet] Issue opened: #6138 Some tunes don't render right in Demo player

Most notably in minigames like Xpanic, which have unique tunes for grenade or shotgun bullet speeds, if you were to replay the video in editor, the objects would fire "normally", causing the bullets to fly off screen if they weren't even moving in game. An example of this is, if I set tune grenade_speed from 1000 to just 100. They fly off, and through the wall, but explode where they should explode in-game. https://user-images.githubusercontent.com/95713843/207716576-eeca55db-53fa-49...

Chairn

well, a small sacrifice

Maps with math challenges/pick a path had to die this death a long time ago kek

kek

(edited)

[ddnet/ddnet] Issue opened: #6139 Possible to join not started team but with practice enabled

First tee joins /team 1 and use /practice Second tee joins /team 1 and is stuck there because of practice Practice should probably act as if the team started

Exported 744 message(s)