DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.org/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 2024-05-06 00:00:00Z and 2024-05-07 00:00:00Z

[ddnet/ddnet] Pull request opened: #8315 Rename chat Say() to SendChat()

See also https://github.com/teeworlds/teeworlds/pull/3239

Checklist

[x] 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 valgri...

[ddnet-web:master] 1 new commit

7fa007c Update POL cost - def-

[ddnet/ddnet] Pull request closed: #8315 Rename chat Say() to SendChat()

02:24

github-merge-queue[bot]

[ddnet:master] 3 new commits

33baeb8 Rename CChat::Say to CChat::SendChat - ChillerDragon 44d0658 Rename CChat::SayChat to CChat::SendChatQueued - ChillerDragon c727b87 Merge pull request #8315 from ChillerDragon/pr_chat_cleanup - def-

Why does the master server report 2 formats for skins?

Image attachment

02:30

First is mine, and second is of a fellow player whose skin data looked different (edited)

this one is 0.7 skin format

this one being?

02:32

the latter im assuming

with body marking and stuff

ah

02:32

and what is the first one then?

0.6 skin format xd

02:33

seems without any color use

Yea I dont use custom colors, anyway how am I on 0.6, and how are those numbers representing color? (I'm trying to make a skin renderer)

so what skin do you render in first

02:35

first one or second one

So far I'm just trying to make sense of it, havent done anything yet

ig there's some skin renderer for both 0.6 and 0.7 written in typescript

I've seen some yea, I could use those

02:36

my use case is, I have a findPlayer function, and it returns whatever it finds after a request to the master server

02:36

from there I have that skin data

02:36

im kinda stuck on getting from skin data to an actually rendered image

afair master should show information about which protocol player uses

02:37

you can render skin by player protocol

02:37

like if player uses 0.6, render 0.6

02:37

if 0.7 -> 0.7

¯\_(ツ)_/¯

Sans

im kinda stuck on getting from skin data to an actually rendered image

ah its a magic xd

Image attachment

02:38

I can tell which is which now that you've told me

02:39

but as far as I can see, there's nothing specifying a protocol

wait wtf

02:42

how did you get 2 different skin formats in one response

Master server gives an array of servers, I filtered them for the server I'm in, from there I went to server.info.clients and that's another array of all connected (and connecting) clients

ah

each client may have a skin property, and if it does, it's in 1 of those 2 formats

its two different players

02:43

got it

yea xd

you can just check if there name in structure you fetching

02:44

so if it has name, its 0.6 one

02:44

otherwise 0.7

yea I know, problem is how do I get from that to an image buffer or at least a url

by yourself xd

why_god_why

there's couple skin database you can check by knocking some fancy url

I'd rather only support the official one by default, with some other way of adding other db's

ddnet uses https://skins.ddnet.org/skin/community/ for community database and https://skins.ddnet.org/skin/ for official ones

02:46

just append skin name to the end and send a request

I'm guessing there's no way to check from which db any skin is?

https://api.skins.tw/database/skins/ also usable

02:47

these all are 0.6

02:47

except of latter one (i don't remember actually)

02:48

yeah these all are 0.6

02:49

idk if there 0.7 db online available

02:49

better make one yourself

zhn

just append skin name to the end and send a request

oh btw, for 0.7 skins, which one do I use? I'm guessing marking but then what are all the other keys for?

Image attachment

they're all parts of one skin

02:50

0.7 skin is constructed by rendering all of them combined

02:50

also 0.7 color is different from 0.6

how so?

02:50

I still dont get how those numbers represent color XD

uh, enable custom color on your skin

Image attachment

zhn

uh, enable custom color on your skin

yea

https://github.com/AlexIsTheGuy/TeeAssembler-2.0/blob/main/dist/color.js is readable enough i think

02:57

anyway afair this string is just concatenated 4 bytes

02:57

represented in 16 base

Sans

Click to see attachment 🖼️

try color&255, color/256&255, color/256/256&255

deathgood

try color&255, color/256&255, color/256/256&255

?

yeah literally

i assume it's uh

02:59

structured like

boolean algebra dont fail me now

RRRRGGGGBBBB

he showed you the way how you can extract numbers out if this number

02:59

it could be hsla

03:00

its uint

03:01

https://github.com/tee-community/tee-skin-renderer/blob/5e6041534e6d67400a768934e2249970e9061882/src/color.ts#L5

03:02

this function is 0.6 only btw

03:02

because converts with color limiting

zhn

https://github.com/AlexIsTheGuy/TeeAssembler-2.0/blob/main/dist/color.js is readable enough i think

this works perfectly for 0.6 skins

03:03

with/without color

03:03

now i need something to handle 0.7 skins

03:03

nod

have never knew that js is so ugly uh

zhn

have never knew that js is so ugly uh

this is why we use typescript

03:04

also that's compiled js, from typescript

03:04

im assuming, at least, cuz it's in a dist folder xd

kek

03:04

outplayed myself

Sans

now i need something to handle 0.7 skins

anyhow this

03:05

know any 0.7 skin renderer?

[ddnet/ddnet] Pull request opened: #8316 Use SendChat() in DbgStress

Sans

know any 0.7 skin renderer?

there's no 0.7 skin renderer in public

03:11

i think so

03:11

you should figure it out yourself by lurking 0.7 skin renderer code in upstream

zhn

https://api.skins.tw/database/skins/ also usable

this one has the most skins because it mirrors all other services + has many skins of its own https://uniqueclan.net/proxy/skin/

zhn

have never knew that js is so ugly uh

i dont think this is an inherent issue in js

Image attachment

[ddnet/ddnet] Pull request opened: #8317 Don't access static ConCommandStatus with a member

Fixes this clang-tidy 19 warning

/home/chiller/Desktop/git/ddnet/src/engine/shared/console.cpp:739:2: error: static member accessed through instance [readability-static-accessed-through-instance,-warnings-as-errors]
  739 |         pConsole->ConCommandStatus(&Result, pConsole);
      |         ^~~~~~~~~~
      |         CConsole::

Ewan

this one has the most skins because it mirrors all other services + has many skins of its own https://uniqueclan.net/proxy/skin/

gm

did i ping u monkalaugh

monkalaugh

>:(

sry

04:03

i usually turn it off

no you didn't xd

04:03

im just curious if pinging online people is something off the road

some ppl get mad about it idk

btw did you ever see an atomic segfault with sigsegv

when storing or loading

04:05

im trying to think of how this could happen lol

04:05

but it doesn't sound unfamiliar

on loading

04:06

i saw a so question about it and that clearly stated that gcc had a bug with unaligned data in atomic

04:06

but its kinda 2015

damn

04:07

std atomic?

yup

bizarre

04:07

idk if i've ever encountered that

04:07

can't remember

04:08

if u attach debugger does it break on an intsinsic

its actually just CHttpRequests state from ddnet codebase being checked in another loop

04:08

maybe its the problem

04:08

id shouldn't check it every tick though

is the CHttpRequest valid

wdym

do u have a ptr to it

its under shared ptr, not a nullptr

ic

i couldn't use it any other way

i just wonder what the debugger says before segfault

04:13

also it's just being init like CHttpRequest ctor -> std::make_shared?

04:13

or some other tomfoolery

clean std::make_shared

04:15

btw seems its loop problem

i see

it didn't sigsegv this time

04:15

i should check httpstate a bit rarer

04:15

is this a real word xd

less frequently

my brain is malfunctioning

rarer isn't a word

got it

more rarely/less rarely

i fucked up my sleep schedule

04:16

again

sadge

04:17

im getting over that myself

04:17

woke up at like 2 today

04:17

but the weird thing is i went to bed at like midnight the day before yesterday and still woke up at 2 pm yesterday... (edited)

same!

04:18

maybe it has something to do with weather

04:18

we had like +15-20C

04:18

and then

04:19

kabooom

04:19

20C drop

damn

04:19

it's colder here too cuz it's spring

04:19

but that really doesn't say anything where i live

04:19

colorado weather is very erratic

justatest

1

i can't say the same about my state

04:20

but last years

04:20

idk

04:20

nature goes crazy

climate change is real

04:20

can't wait until we get 120 degree summers and -20 winters

yappie

04:21

there's no need in nuclear winter, we'll have the winter another way

so true

i bet if global community will do anything in future

04:22

or will they go like

04:22

"climate cycle huh"

yea

04:23

record breaking temperatures? melting ice caps? crazy coincidence (edited)

xdd

nobody will act until it's too late

04:23

it's already too late

sad but true

04:24

i did a little presentation about ecologists

04:24

and should have been write a little research on this topic

04:25

found out that there's no actual impact behind these people since none of them are actual manufacturer xd

04:25

manufacturers*

04:25

factories go brrr

yea

04:27

small companies and individuals are already doing as much as they should but each plastic straw is nothing compared to the pollution being created by international transit and meat packing

04:27

and in that area it's all about execs and their wallets

04:28

this is where we need to depend on legislature to do what's best for the world but all of the first world countries have already proven they don't give a shit about the future

04:28

so it's probably over

had a read about pagers cancellation on a wave of drug-dealing in us

04:31

like

04:32

they really sent children to the prison instead of removing key figures of drug-dealing

04:33

and now they want to ban phones in us xd

04:33

for underages*

typically

04:34

law enforcement and incarceration in the u.s. is just nasty

04:34

not necessarily about putting away those who do badly but more about maintaining some political image or fulfilling politicians' interests

why lobbying is still a thing wtf

04:35

and its legal in us

$

our world is doomed i suppose xddd

yeah

04:35

usually the politicians are getting a cut of whatever the lobbyist is lobbying for

04:35

i mean not usually but often

04:36

and especially in the headline cases

04:37

doomed indeed

[ddnet/ddnet] Pull request opened: #8318 Add support for clang tidy 19

04:41

[ddnet/ddnet] Issue closed: #8297 warning: 3 uninitialized fields at the end of the constructor call

[ddnet/ddnet] Issue opened: #8319 CamelCase or SNAKE_CASE for constants?

I can not find a clear trend for static const ``` $ grep 'static.*const' src/ -r | grep -v '[()]' | grep -Ev '(textrender|external)' [..] src/engine/client/client.cpp:static const ColorRGBA gs_ClientNetworkPrintColor{0.7f, 1, 0.7f, 1.0f}; src/engine/client/client.cpp:static const ColorRGBA gs_ClientNetworkErrPrintColor{1.0f, 0.25f, 0.25f, 1.0f}; src/engine/client/client.cpp: static const int NETTYPES[2] = {NETTYPE_IPV6, NETTYPE_IPV4}; src/engine/client/sound.cpp:static constexpr ...

[ddnet/ddnet] Pull request opened: #8320 Enforce more styling with clang tidy

morning

@Learath2 https://github.com/greshake/i3status-rust

GitHub - greshake/i3status-rust: Very resourcefriendly and feature-...

Very resourcefriendly and feature-rich replacement for i3status, written in pure Rust - greshake/i3status-rust

08:41

>>> Emerging (6 of 6) games-action/teeworlds-0.7.5-r3::gentoo >>> Installing (6 of 6) games-action/teeworlds-0.7.5-r3::gentoo >>> Completed (6 of 6) games-action/teeworlds-0.7.5-r3::gentoo

08:41

tf

08:41

teeworlds got updated?

08:41

https://www.teeworlds.com/?page=news&id=14785

"Up to 16 Players" YEP

YEP

ah

08:42

its old

08:42

i guess gentoo patched it

kek

nothing from tw xd

Is oy the only one with Wiki Access ?

Ryozuki

teeworlds got updated?

imagine

github-merge-queue[bot]

[ddnet:master] 2 new commits

3c606dd Don't access static ConCommandStatus with a member - ChillerDragon f9cb924 Merge pull request #8317 from ChillerDragon/pr_static_member_console - Robyt3

09:06

[ddnet/ddnet] Pull request closed: #8317 Don't access static ConCommandStatus with a member

09:06

[ddnet/ddnet] Issue opened: #8321 Client timeout code handling should be moved from engine client to gameclient

The engine client should not send game messages. It's currently the only use of CNetMsg_Cl_* in engine. See https://github.com/ddnet/ddnet/pull/5949#discussion_r1590462743.

Robyt3

Even better would be if there was a downloadable skin index json which contained the hashes of the current skin files, so downloading the entire skin would be unnecessary to check if it's unchanged

we can use http caching for that. there's an etag header and cache-control: max-age

$ curl -I https://ddnet.org/skins/skin/cammo.png
HTTP/2 200 
date: Mon, 06 May 2024 09:05:57 GMT
content-type: image/png
content-length: 5863
last-modified: Tue, 27 Jan 2015 00:48:35 GMT
etag: "54c6e063-16e7"
expires: Sat, 27 Apr 2024 07:53:08 GMT
cache-control: max-age=86400
access-control-allow-origin: *
cf-cache-status: HIT
age: 75851
accept-ranges: bytes
server: cloudflare
cf-ray: 87f7cfc00e533bc4-WAW
alt-svc: h3=":443"; ma=86400

$ curl -H 'If-None-Match: "54c6e063-16e7"' -I https://ddnet.org/skins/skin/cammo.png
HTTP/2 304 
date: Mon, 06 May 2024 09:08:23 GMT
last-modified: Tue, 27 Jan 2015 00:48:35 GMT
etag: "54c6e063-16e7"
expires: Sat, 27 Apr 2024 07:53:08 GMT
cache-control: max-age=86400
access-control-allow-origin: *
cf-cache-status: HIT
age: 75997
server: cloudflare
cf-ray: 87f7d34e4ab63542-WAW
alt-svc: h3=":443"; ma=86400

github-merge-queue[bot]

[ddnet:master] 2 new commits

28f3ac3 Use SendChat() in DbgStress - ChillerDragon d4e3f6f Merge pull request #8316 from ChillerDragon/pr_use_sendchat - Robyt3

09:10

[ddnet/ddnet] Pull request closed: #8316 Use SendChat() in DbgStress

Daniel

especially for me

like im playing since 2013-2014 and want to know how many k hours i spent in this lovely game (im not ready to install it on steam because i know the hours will never match the truth)

just start now so that you'll know the game hours in the last 10 years in 2034

Patiga

In Rust, which crate would you suggest me to create a thread pool, that limits the amount of pending tasks? creating a new task when the task-list if full should block or smth

https://docs.rs/rayon/1.10.0/rayon/iter/trait.ParallelBridge.html#tymethod.par_bridge does that nicely

ParallelBridge in rayon::iter - Rust

Conversion trait to convert an Iterator to a ParallelIterator.

Patiga

thanks :) I found that rayon offers an API to turn non-parallel iterators into parallel ones, that works out nicely https://docs.rs/rayon/latest/rayon/iter/trait.ParallelBridge.html#tymethod.par_bridge

ah, you already said that

heinrich5991

we can use http caching for that. there's an etag header and cache-control: max-age

$ curl -I https://ddnet.org/skins/skin/cammo.png
HTTP/2 200 
date: Mon, 06 May 2024 09:05:57 GMT
content-type: image/png
content-length: 5863
last-modified: Tue, 27 Jan 2015 00:48:35 GMT
etag: "54c6e063-16e7"
expires: Sat, 27 Apr 2024 07:53:08 GMT
cache-control: max-age=86400
access-control-allow-origin: *
cf-cache-status: HIT
age: 75851
accept-ranges: bytes
server: cloudflare
cf-ray: 87f7cfc00e533bc4-WAW
alt-svc: h3=":443"; ma=86400

$ curl -H 'If-None-Match: "54c6e063-16e7"' -I https://ddnet.org/skins/skin/cammo.png
HTTP/2 304 
date: Mon, 06 May 2024 09:08:23 GMT
last-modified: Tue, 27 Jan 2015 00:48:35 GMT
etag: "54c6e063-16e7"
expires: Sat, 27 Apr 2024 07:53:08 GMT
cache-control: max-age=86400
access-control-allow-origin: *
cf-cache-status: HIT
age: 75997
server: cloudflare
cf-ray: 87f7d34e4ab63542-WAW
alt-svc: h3=":443"; ma=86400

thanks, I'll add that to my TODO list

Ryozuki

@Learath2 https://github.com/greshake/i3status-rust

got it integrated

Image attachment

Ryozuki

got it integrated

looks pretty owo

owo

Ryozuki

got it integrated

Dudes PC is goated af

Robyt3

thanks, I'll add that to my TODO list

(actually, maybe expires: Sat, 27 Apr 2024 07:53:08 GMT is easier to parse than cache-control: max-age=86400 because we can use curl's date parser (which we already use elsewhere))

Ryozuki

got it integrated

gigachad

Image attachment

@heinrich5991 do u know if the -I option for includes only accepts absolute paths?

12:26

-I

also accepts relative paths

meloƞ

gigachad

Give me one of those screens pls owo

owo

Teero

Give me one of those screens pls owo

owo

No

meloƞ

No

pleeeaaaase

OK 300€

50€ and one hug?

12:34

+ my 4k points kog acc

Maybe the authors of status.tw have some old data around: https://status.tw/, you could ask Mr_Anderson on https://discord.gg/paavjRnH

Invite to join a server

announcements

Teero

50€ and one hug?

How Long is the hug?

meloƞ

How Long is the hug?

aipgores theoretical perfect time :D (edited)

Hmm, make it Times two and i might considere

so 67s?

13:00

sure

man

13:52

i hate depending on C libs

13:52

or cpp

13:52

i wasted 2 hours trying to link something

what's the current example?

heinrich5991

what's the current example?

you are asking me or?

yep

13:53

what C lib did you try to link?

so i forked a rust library that uses a llvm library

13:53

called tablegen

13:53

https://github.com/edg-l/tblgen-rs-alt

GitHub - edg-l/tblgen-rs-alt: tblgen-rs updated for llvm 18

tblgen-rs updated for llvm 18. Contribute to edg-l/tblgen-rs-alt development by creating an account on GitHub.

13:53

for some reason it fails on ubuntu, when running cargo test

13:53

but it works locally

13:53

and it also works on macos

ah

yea. depending on big C++ projects like LLVM is a pain

@Ryozuki when edlang-cranelift?

https://github.com/edg-l/tblgen-rs-alt/actions/runs/8969217906/job/24630181711?pr=5

Learath2

@Ryozuki when edlang-cranelift?

i wish cranelift was a optimizing compiler

13:55

but sadly they just want it to be fast for compile time and wasm focused

I mean the "optimizing compiler" use case is well-covered by LLVM

yeah, its what edlang uses

13:56

i dont even have a fully working language so multiple backends is too much work xd

13:56

but i would love to see something like LLVM but that its library friendly

13:56

a.k.a usable as a library outside the language it is made in (be it c++ like llvm or rust)

13:57

llvm is too much cpp focused

13:57

the C api is a after thought

13:57

i got a nice idea

I'd say the best bet for that would not be to rewrite LLVM from scratch, but enhance LLVM's ability to be a library

make a "language-server" but for library interface generators

13:57

kek

13:57

some standard protocol

the language server protocol is a bit too much "whatever vscode does" I've heard

xd yeah but i meant the idea of protocol to have easy codegen for the library in the language u want to use

13:58

i guess defining the api in a json file or smth

13:58

and use it to codegen

sounds inefficient, I don't like it

how is it ineficient?

heinrich5991

I'd say the best bet for that would not be to rewrite LLVM from scratch, but enhance LLVM's ability to be a library

this would be a gargantuan task anyway, you'd have to have a very very good reason

using JSON/HTTPS is slow for APIs

the json file is just to generate for example the rust library that provides a nice interface to say llvm

13:59

i think u misunderstanding me

ah

the file is used only at compile time to get a native library

14:00

which u use

I see

14:00

there are a couple of projects that do something like that

my aim is to have nice libraries for stuff made in other languages

e.g. https://github.com/mozilla/uniffi-rs/

GitHub - mozilla/uniffi-rs: a multi-language bindings generator for...

a multi-language bindings generator for rust. Contribute to mozilla/uniffi-rs development by creating an account on GitHub.

14:00

I fixed that up a little while trying to get nice python bindings for libtw2-huffman

that looks nice

14:00

but llvm is made in cpp not rust

14:00

xd

I'm sure there's something similar for C++

14:01

I remember swig from the old days, but it wasn't that good IIRC

14:01

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

The Simplified Wrapper and Interface Generator (SWIG) is an open-source software tool used to connect computer programs or libraries written in C or C++ with scripting languages such as Lua, Perl, PHP, Python, R, Ruby, Tcl, and other language implementations like C#, Java, JavaScript, Go, D, OCaml, Octave, Scilab and Scheme. Output can also be i...

A common interface description language wouldn't be the worst idea. It sounds so obviously like a good idea that I wonder why there isn't one around already

there's IDL

Learath2

A common interface description language wouldn't be the worst idea. It sounds so obviously like a good idea that I wonder why there isn't one around already

yeah xd

it's used in web, and uniffi uses it as well

Learath2

A common interface description language wouldn't be the worst idea. It sounds so obviously like a good idea that I wonder why there isn't one around already

you know what is the current common interface description language? C

Ryozuki

you know what is the current common interface description language? C

that's not really an interface description language

i know it was half a joke

ah, it was called WebIDL: https://en.wikipedia.org/wiki/Web_IDL

Web IDL is an interface description language (IDL) format for describing APIs (application programming interfaces) that are intended to be implemented in web browsers. Its adoption was motivated by the desire to improve the interoperability of web programming interfaces by specifying how languages such as ECMAScript should bind these interfaces.

owo

14:03

looks like nobody rly thought much of IDL

I guess one of the biggest issues is efficiently passing stuff from language to language

but I think I heard uniffi is moving away from using an IDL :/

14:04

I liked that part about them

Learath2

I guess one of the biggest issues is efficiently passing stuff from language to language

check out uniffi, it has some good ideas

14:04

it has some primitive data types and can also handle opaque handles

14:05

https://mozilla.github.io/uniffi-rs/udl_file_spec.html

I guess what uniffi lacks is support for languages other than rust

for library languages, yes

Learath2

I guess one of the biggest issues is efficiently passing stuff from language to language

but that is something they try to do

I'll check it out, it would be nice to see how it works

limitation: https://github.com/mozilla/uniffi-rs/issues/1586

14:07

but the limitation could be fixed

Does uniffi have requirements on the types of functions that can be exposed to other languages?

yes. it has a data model that covers some types, can create more types and can pass opaque types ("handles")

14:09

it cannot pass anything that falls outside of this data model

[ddnet/ddnet] Pull request opened: #8322 Use network tokens in big-endian byte order consistently

Previously, it was handled inconsistently on big-endian machines and caused connection failures.

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 ...

14:27

[ddnet/ddnet] Pull request opened: #8323 Fix smoke particles sometimes being stuck inside solid tiles

Nudge the initial position for explosions' smoke particles slightly towards the edge of the closest non-solid tile, if it would otherwise be inside a solid tile, so the smoke particles do not get stuck inside solid tiles on explosion events happening at the edges of tiles but slightly inside them. The physical position of the explosion event is unchanged, so this does not affect physics. The explosion sprite is still rendered at the physical position of the explosion, to preserve the appreara...

some days i wanna continue edlang, some others i want to restart ryoOS

14:28

but no time kek

14:28

do programmers who retire spend their time doing their dream projects?

14:28

or they just go drink beers and forget

I wonder how well I'll be able to program when I retire (edited)

same

the ability to think decreases with age

my thought is that when i retire ill do projects that require lot of time

14:30

but maybe the brain isnt there anymore (edited)

14:30

justatest

14:30

https://stackoverflow.co/company/press/archive/openai-partnership

Stack Overflow and OpenAI Partner to Strengthen the World’s Most Po...

Founded in 2008, Stack Overflow’s public platform is used by nearly everyone who codes to learn, share their knowledge, collaborate, and build their careers.

14:31

https://stackoverflow.co/teams/ai/

Empower your organization for an AI future with Stack Overflow for ...

Build a collective, enterprise-grade knowledge base around technology transformations, including GenAI and AI/ML.

14:32

what was the previous buzz word before AI came?

14:32

i forgot

cloud

oh true

14:32

well its still here isnt it

yeah

low code

14:32

kek

14:34

another idea i have is to rewrite some c/c++ libraries to rust but only not rly big ones

14:36

https://github.com/xiph/opus/tree/main is this a big project to port

GitHub - xiph/opus: Modern audio compression for the internet.

Modern audio compression for the internet. Contribute to xiph/opus development by creating an account on GitHub.

14:39

looks quite big

14:39

also it uses goto kek

14:39

altho just to jump to error handling

definitely not small

Robyt3

thanks, I'll add that to my TODO list

maybe a skin index would still be nice, for the skin list and to not do useless requests. but maybe as a first step this would already be good

Ryozuki

https://github.com/xiph/opus/tree/main is this a big project to port

oh yeah pls

15:37

opus and vorbis please

15:37

thanks

15:37

and zlib encoding

15:37

then last c dep is ring

xd

15:54

Image attachment

Ryozuki

Click to see attachment 🖼️

justatest

author has clearly never seen a tree root system

i've never seen a tree

15:59

@Ewan where is promised stream

2 hours

16:00

i have dr appt

16:00

Image attachment

16:00

dog tho

Ewan

Click to see attachment 🖼️

be careful dog eye is reflecting u!

sad doggo

now he sweating dxd

she is chilling

16:03

dozing

Image attachment

@Ewan i can see u

Image attachment

is healthcare for dogs also expensive in america?

yes

Teero

@Ewan i can see u

omg do you also see that reptiloid?

The one in the background or the one behind the ear?

all of 'em

Teero

@Ewan i can see u

now use AI to get the dog back

Teero

@Ewan i can see u

Computer, enhance

16:09

gotta love the free tools lmao

Image attachment

so true

why do i find these memes funny

Image attachment

Learath2

Click to see attachment 🖼️

omg these american shows that made perfect 8k images out of minecraft textures

Jupstar ✪

omg these american shows that made perfect 8k images out of minecraft textures

yes yes yes i love ai lmao

Image attachment

16:12

maybe cuz of the high res pixelart?

16:12

ima try super low res xd

@Teero i also wait for ur stream

16:13

no excuses

xDD

Ryozuki

why do i find these memes funny

And this specific definition of it isn't even that bad

|x| = -x is true if x is negative?

Jupstar ✪

@Teero i also wait for ur stream

"So today ill teach you how to abuse the teeworlds master server to store files :D"

Ryozuki

|x| = -x is true if x is negative?

y

16:14

or if x = 0 (edited)

Learath2

And this specific definition of it isn't even that bad

what are ur thoughts on 0 not being on the first

Not many, I've learned to take these things as just definitions

16:16

If I were making math from scratch I'd say 0 is a counting number

16:17

Check out the construction of real numbers by Dedekind cuts, or that one construction from Z that took a computer to prove

justatest

16:18

yep

@Jupstar ✪ its fine ig

Image attachment

Image attachment

(to be fair, I never really had to think of them in any of these ways (I'm not a mathematician), I find the axiomatic way to define it more natural)

Teero

@Jupstar ✪ its fine ig

from this xd (edited)

Image attachment

R is a field under addition and multiplication
R is totally ordered
Addition and multiplication preserve order

and there was one more I can't remember

https://www.jeremykun.com/2024/05/04/fhe-overview/

A High-Level Technical Overview of Fully Homomorphic Encryption

About two years ago, I switched teams at Google to focus on fully homomorphic encryption (abbreviated FHE, or sometimes HE). Since then I’ve got to work on a lot of interesting projects, learning along the way about post-quantum cryptography, compiler design, and the ins and outs of fully homomorphic encryption. If you’ve heard about FHE and you...

Ah the last one is that every subset bounded from above has a least upper bound (edited)

Teero

from this xd (edited)

mh not bad, but would be cooler if it could extend that image instead of just upscale

16:24

reconstruct the dog

Ryozuki

|x| = -x is true if x is negative?

yeah that’s the entire point of absolute value

16:24

x > 0 ? x : -x

i know but at first it can be counter intuitive for some xd

maybe

16:24

i learned it in elementary school

Ewan

i learned it in elementary school

you learned that shit in 8th grade?

16:25

german education fails once again fr

it was a meme on twitter

16:25

and u can guess

meloƞ

you learned that shit in 8th grade?

elementary school for me was kindergarten-6th grade

meloƞ

you learned that shit in 8th grade?

i also learned that in 7th or 8th grade wtf

crythumbsup

catxplosion

Image attachment

Ryozuki

Click to see attachment 🖼️

gigachad

why im getting this memes

Image attachment

2

16:33

ill stop

16:33

i was just building the newest version from the aur and i got a bunch of warnings. is it intentional? xd

7.77 KB

gcc version?

1

Teero

i was just building the newest version from the aur and i got a bunch of warnings. is it intentional? xd

that sounds like lto

is C good

justatest

Image attachment

has anyone done ai generated teeworlds maps yet?

real neural net AI or proc gen

Ryozuki

why im getting this memes

yooo p-n-tunnel fires

Ewan

real neural net AI or proc gen

machine learning approaches

not afaik

17:34

cleaning up a little bit and then i will find a pr i guess

louis

has anyone done ai generated teeworlds maps yet?

@iMilchshake has done algorithmic generation but i don't think anyone has done stuff with NNs yet (edited)

17:34

https://github.com/iMilchshake/gores-mapgen-rust

GitHub - iMilchshake/gores-mapgen-rust

Contribute to iMilchshake/gores-mapgen-rust development by creating an account on GitHub.

louis

has anyone done ai generated teeworlds maps yet?

wasn't there a person that did some ddmax like maps using nn?

truly ahead of their time

or its just me, malfunctioning again xd

17:35

i did wake up btw xd

gm

gm @zhn

me too but not after 80 mins of hitting snooze

justatest

1

17:36

but tbf i set my alarm way too early

louis

has anyone done ai generated teeworlds maps yet?

i might do a prototype based on my procedual map generator as i can easily produce thousands of maps as training data o:

17:42

but then It'd just learn to replicate maps that can already be generated, so the output woulnt be that useful i guess. But maybe it learns to 'generalize' maps and can make even cooler shit o: (edited)

im thinking of downloading the ddnet map database and just feeding it through some simple vae to see what happens

17:44

also you could do some diffusion stuff since ddnet tiles can easily be represented as noise

where is the low hanging fruit

17:59

here's the stream

17:59

not quite started yet

17:59

just need a few min to get ready

louis

im thinking of downloading the ddnet map database and just feeding it through some simple vae to see what happens

yeah (V) Auto-Encoder sounds like a good idea. Please ping if u actually do it :D

looks like there aren't actually any settings i can change

18:30

to make this better

18:33

let's try this again

18:35

did those of u with connection issues try turning the quality down

Image attachment

Original message was deleted or could not be loaded.

sadge

Image attachment

what does it say

louis

has anyone done ai generated teeworlds maps yet?

https://www.youtube.com/watch?v=sJ6R6zFMhE4

Teeworlds Machine Learning Project (TMLP) - Untrained network demo

18:38

timakro is the first

18:38

but not map gen

18:38

altho nothing much came out of it but looked fun

Ewan

what does it say

just an error message saying an error appeared and i should try later

18:38

nothing

https://www.youtube.com/watch?v=8ocxnWipQDw

Teeworlds Machine Learning Project (TMLP) - Training data example

wtf

Teero

just an error message saying an error appeared and i should try later

no idea what to do

18:38

works perfectly for me watching on youtube

18:39

maybe i try twitch...

@Ewan it loaded for me

18:39

after some time xd

i think i will try twitch

@Ewan u can hide the scene lol

eh

nice american english

18:40

oh wait we talked in vc before

Ryozuki

https://www.youtube.com/watch?v=sJ6R6zFMhE4

waiting for someone to do that but with a reinforcement learner lol

what category do i put it in on twitch

software development

https://www.twitch.tv/ewancg

ewancg - Twitch

Implementing practice mode weapon commands for DDNet

18:46

here

damn white theme

Ewan

https://www.twitch.tv/ewancg

laggs for me too

i selected frankfurt server

@Ewan lmao u lure random twitch bots

yea

do you read twitch chat?

@Ewan i See what you're cooking, i Like your voice PeepoSimp

PeepoSimp

[ddnet/ddnet] Pull request closed: #8322 Use network tokens in big-endian byte order consistently

19:53

github-merge-queue[bot]

[ddnet:master] 2 new commits

36712db Use network tokens in big-endian byte order consistently - heinrich5991 910001a Merge pull request #8322 from heinrich5991/pr_ddnet_endian_token - Robyt3

Could the ddnet license be changed to prevent the distribution of open source cheats? Obviously wouldn't apply to past versions but if it's allowed by the current license it would be nice

Tater

Could the ddnet license be changed to prevent the distribution of open source cheats? Obviously wouldn't apply to past versions but if it's allowed by the current license it would be nice

how would a license like that would look like?

Idk

wouldnt that already require to define what a cheat is

Sure

sounds hard lol

lol

Anything that modifies a players input

also let's be honest.. ppl that share cheats don't care about licenses lmao

But open source cheats will be subject to github rules

and github will completely ignore us

dmca exists and is free xd

Tater

dmca exists and is free xd

then they will move to gitlab or their own vcs server

Significantly more restrictive licenses exist so I think it's definitely technically possible

its very hard to even sue them

You don't need to sue just dmca

I dont get it tho xD

20:02

They are probably not gonna go open source

20:02

because they want to put in their malware stuff

i personally dislike making ppl criminals, be it on accident or on purpose. so i prefer the most free license. that cheats are spreading sucks, but probably is inevitable

true but atleast it prevents the spread of the cheat source

I doubt

20:03

It will just be a lot of work for ddnet contributors (edited)

Even if you don't act on it atleast it's a legal deterrent like gpl licenses

mhmm, they mostly dont care about legal stuff

But they're not doing anything illegal as it is

on long term it will be much more work than usage

Tater

Even if you don't act on it atleast it's a legal deterrent like gpl licenses

there is a reason i kinda dislike gpl too

Avolicious

on long term it will be much more work than usage

It's only as much work as you want to do, you can just put the clause in the license and do nothing if you want

@Tater but if u know other games doing it. i'd like to see such a license that you imagine. i wonder if it legally (at least theoretically) works

[ddnet/ddnet] Pull request opened: #8324 Add chat commands for toggling weapons in practice mode

Some people wanted the ability to give themselves weapons within practice mode without having to teleport to the location of the weapon on the map. This allows them to do so by typing something like /grenade in the game chat (as long as they're in a team with practice mode enabled). Note that with the way things are currently being done, there's no clean way for me to "wrap" the callbacks used in the rcon commands cleanly. Currently there has to be a duplicate function for every comman...

1

bluekitty

1

brownbear

1

heartw

1

2

Jupstar ✪

@Tater but if u know other games doing it. i'd like to see such a license that you imagine. i wonder if it legally (at least theoretically) works

Well its mostly against ToS, there are just a few open source games available

Image attachment

sweet

(joke)

Tater

Could the ddnet license be changed to prevent the distribution of open source cheats? Obviously wouldn't apply to past versions but if it's allowed by the current license it would be nice

such a license would not be an open-source license anymore

️

heinrich5991

such a license would not be an open-source license anymore

Why take an ideological open source stand point when we already prevent the discussion of cheats in the discord and in game.

20:20

If we truly did not care about what people did with the source code then all discussion of it would be allowed

well but we don't have to actively give ppl a platform for discussing it

Tater

Why take an ideological open source stand point when we already prevent the discussion of cheats in the discord and in game.

that was mostly a statement of fact

20:21

but you're right that I also want ddnet to stay an open-source game

https://blog.rust-lang.org/2024/05/06/check-cfg.html

Automatic checking of cfgs at compile-time | Rust Blog

Empowering everyone to build reliable and efficient software.

You don't need it to be obstructive of legitimate uses, cheats which are unable to join official servers could be allowed by the license

20:25

Actually nvm our license forbids this

20:25

Lol

what does our license forbid?

The license can never be meaningfully changed because point 3 said "this" instead of "the above"

I think that's compatible with GPL and a lot of others

Hmm actually I might have been wrong

20:40

I guess both terms are used in foss licenses

tater should be purple

true /team 48

Ryozuki

(joke)

If only that was real

Tater

Could the ddnet license be changed to prevent the distribution of open source cheats? Obviously wouldn't apply to past versions but if it's allowed by the current license it would be nice

ddos is illegal too, but that doesn't stop anyone

21:10

it would not make any difference

Peter0x44

ddos is illegal too, but that doesn't stop anyone

ddos is more anonymous than cheat distributing. but you're right, people do illegal things when they won't be caught

Tater

ddos is more anonymous than cheat distributing. but you're right, people do illegal things when they won't be caught

but... it was known who was doing some of the ddos and it made no difference

I think deen did not want to pursue criminal charges or did not have the time and resources

law enforcement doesn't care about that, let alone whatever this license would be

Tater

I think deen did not want to pursue criminal charges or did not have the time and resources

afaik the german police simply ignored his request

the idea is that law enforcement doesn't need to care because you can use dmca which github and gitlab would have to respect, but I think the idea is dead anyway

Jupstar ✪

afaik the german police simply ignored his request

idk it's not easy for them to gather proof so I can see why they don't care

21:16

I've never heard of regular police investigating ddos before

Jupstar ✪

true /team 48

i think 48 isn't purple anymore

lol someone from intel (with a intel.com domain) contacted me

Image attachment

Ireland?

idk why ireland but i guess intel has bases in places xd

21:41

im more of a amd guy tho NekoEvil

NekoEvil

(edited)

your github profile does look very relevant for what they're asking for

owo

r u gonna apply

i think ryo is not so versed in cpp tho?

Ewan

tater should be purple

what's their github username?

Exported 580 message(s)