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 2024-02-26 00:00:00Z and 2024-02-27 00:00:00Z

BTW @heinrich5991 that crashbug‘s solution depends on wether a pr gets merged in ddnet that is marked as depends on #7777 any chance we could merge or decline just that pr? Or somehow move forward #7777?

chillerdragon

Also btw the latest version is 0.7 not 0.6 :p

xd

01:48

what about me doing teeframe for 0.6?

01:48

pepeW

01:51

how the server send to the client that someone is chatting?

Miguilim

how the server send to the client that someone is chatting?

nvmd its on obj_character

01:52

wait, the client known every other tee health, armor, ammo count?

Ryozuki

i gotta make a website xd

can u make windows builds of edlang

02:01

bad for discovery when u only have linux build

It sends 0 values unless it’s sent to the connection that owns the tee

Replying to @Miguilim wait, the client known every other tee health, armor, ammo count?

Ewan

can u make windows builds of edlang

Nah He should Just openly brag about Linux Superiority and Hit everyone with the: try it Out with WSL2!!!!

What about it? You do Teeframe for an outdated version.

Replying to @Miguilim what about me doing teeframe for 0.6?

(joking btw)

I think

Replying to chillerdragon It sends 0 values unless it’s sent to the connection that owns the tee

02:26

Not 100% sure

02:26

I think spectators get the info too. Spectators and owners

meloƞ

Nah He should Just openly brag about Linux Superiority and Hit everyone with the: try it Out with WSL2!!!!

i think u can still build it on windows very easily

02:48

but tbh i can see him cutting off windows support in the future

02:48

which would be bad

02:49

sorry again about your api key melon

02:50

i used it a little bit but apparently it forgot or something lol

Image attachment

kek

no worries

<ChillerDragon> @Ewan need windows download? Here you go https://downloads.teeworlds.com/solidfiles.com/17358/DDRace_Windows.rar

03:10

<ChillerDragon> ddrace_windows.rar served by downloads.teeworlds.com lmao

lol thanks

03:15

but we were talking about edlang

03:15

maybe u can't see the reply

<ChillerDragon> @heinrich5991 does deprecated mean you are going to delete that directory? https://heinrich5991.de/teeworlds/maps//maps_deprecated_with_crc/ or just not update?

03:16

<ChillerDragon> @Ewan ye ik

how are u chiller dragon

<ChillerDragon> good u?

im alright

03:17

i wanted to try clear linux on my server since it's well-optimized and low maintenance

03:17

but the cdn is soooo slow

03:17

install took 90 mins or something

03:17

i didn't check uptime

03:18

seems they don't really have shittons of pkg mirrors hosted by universities and shit like most big distros have

it boots to desktop in like 15 seconds tho

https://www.clearlinux.org/clear-linux-documentation/tutorials/nvidia

04:10

this is kinda rough tho

Ewan

bad for discovery when u only have linux build

i want to but its honestly hard and no win machine

mingw is rly ez to use

07:00

wait i forgot its rust

07:00

xddd

yeah

cross compiling on rust is so ez though...

<ChillerDragon> til this works in teeworlds ./teeworlds teeworlds:127.0.0.1 as an alternative to ./teeworlds "connect 127.0.0.1" thats so weird

ye the thing is not that

testing?

i havent tried but apparently i have to use the same llvm version as rust? but not sure

07:01

also idk how the linker works

i guess you have to have the language support it too

07:02

i forgot

07:02

should still be fine tbh

https://gitlab.com/taricorp/llvm-sys.rs

Peter Marheine / llvm-sys.rs · GitLab

Rust bindings to LLVM's C API.

07:02

Windows You must use a version of Rust that uses the same compiler as you build LLVM with, either MSVC or MinGW. Fortunately, a mismatch like this will cause errors at compile-time when llvm-config provides options which are supported by only one of them, so if you're using the other it will cause the build to fail.

07:03

weird shit

bizarre

chillerdragon why couldn't u import twlib.copy_tree as copy_tree?

Image attachment

07:15

soon we will see how quickly these 3 machines can work together to compile ddnet

justatest

2

ima try to do edlang windows

Ewan

Click to see attachment 🖼️

so american

07:30

smol table

table is big

07:30

it just has topper

Ewan

Click to see attachment 🖼️

Wtf that's some interesting layered setup

Jupstar ✪

Wtf that's some interesting layered setup

Held up by clamps xddd

Yea dude

08:03

i can put my laptop and smaller keyboard behind the keyboard too

08:03

under the main platform

But you still have no mouse space xxd

08:04

Are you playing high sens?

Not really playing

1

woah i hate windows

08:13

such a dev hostile OS

Image attachment

Ryozuki

such a dev hostile OS

Replace Dev with user

replace user with (char)(8) (edited)

https://alfter.us/2023/09/29/cheatsheet-install-debloated-windows-11-on-qemu/

Cheatsheet: install debloated Windows 11 on QEMU

08:48

@Ewan im trying but i already hating myself

@Ryozuki what for?

edlang for winshit

uff

fuckyousnail

Image attachment

I'd just install windows on a different drive

08:51

Atleast that's what I do for gaming windows shit

i dont want windows

08:51

just doing this to compile

08:51

i tried making my dockerfile but i suck

Ryozuki

i dont want windows

Neither do I but if protondb says its impossible to emulate....

i dont play the game then

Ryozuki

just doing this to compile

What about GitHub actions?

Teero

What about GitHub actions?

that will be when i get this working

08:52

ci driven development is the worst

08:53

tbh if this takes me more than some hours i will pass

08:53

edlang for linux only

Ryozuki

i dont play the game then

But I really wanna play lost ark

i thought you knew better than to install visual studio

Ewan

i thought you knew better than to install visual studio

its rustup

Just install mingw xd

you're still using msvc

and no i dont know better

08:53

and i dont want to learn windows shit

lol

angry developer moment

08:54

xd

https://www.msys2.org/

Software Distribution and Building Platform for Windows

08:55

Install mingw64 and cmake and you should be good to go

no im definitly not good to go

08:56

im not doing ddnet stuff

Maybe already posted here, idk xd

Image attachment

tf?

funny winget only has llvm 15

08:57

more outdated cant be

Ryozuki

im not doing ddnet stuff

wdym

08:57

why does that matter

Compiler is compiler no matter what project it is xd. Except its not cpp

msys2 is still good setup for rust

yea

u can get rust from msys repos or add your C:\Users\ryozuki.cargo\bin to your MSYS path

08:59

either in /etc/profile or ~/.profile or ~/.config/fish/config.fish if ur using a good shell

08:59

then it's like linux ish

Ewan

why does that matter

it matters cuz it needs to match what llvm was compield with

08:59

and as i said idc i dont wanna learn winshit

you can't get a mingw build of llvm ?

idk

09:00

i definitly dont want to build llvm in a vm

weak

haha weak is using windows

[ddnet/ddnet] Pull request opened: #8018 ci: remove useless command in ci

That command didn't set python version to 3.11 anyways, and it works just fine without it, why not remove it?

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 ...

Miguilim

yes, the serverbrowser info

the connectionless stuff is only needed for lan servers nowadays. serverinfo for internet servers doesn't go through connless packets anymore

Ryozuki

Click to see attachment 🖼️

where do I find this?

https://ddstats.org/events

View DDNet / Teeworlds Statistics.

ws-client

<ChillerDragon> @heinrich5991 does deprecated mean you are going to delete that directory? https://heinrich5991.de/teeworlds/maps//maps_deprecated_with_crc/ or just not update?

ChillerDragon: currently just not update. what do you use it for?

Ryozuki

https://ddstats.org/events

cool

source?

09:22

I considered doing something like that for the client once, for live-updates of the serverlist

heinrich5991

cool

source?

check the link in the website

https://github.com/edg-l/teeobserver

GitHub - edg-l/teeobserver: This tool continuously fetches the DDNe...

This tool continuously fetches the DDNet master server and observes the changes, broadcasting the events and providing a API. - edg-l/teeobserver

or read it

09:23

yes

09:26

https://doc.rust-lang.org/cargo/reference/overriding-dependencies.html

Overriding Dependencies - The Cargo Book

GitHub

Click to see attachment 🖼️

Milkeye????? Wtffff actual useful commit huh ?? You sick or smth

Ryozuki

https://doc.rust-lang.org/cargo/reference/overriding-dependencies.html

what about it?

09:31

I use it for testing the fixes in the crates I depend on

09:31

very useful for that

yeah

09:32

@Ewan no edlang for windows for now

Ryozuki

@Ewan no edlang for windows for now

what does it depend on? some LLVM shenanigans?

heinrich5991

what does it depend on? some LLVM shenanigans?

just llvm library

09:32

but its a pain

rip

09:33

distcc tricky to set up across OSes

cuz the prebuilt llvm doesnt have llvm-config which is a tool it tells u the llvm version and library paths

09:33

and llvm-sys depends on it

09:33

and i dont want to waste more time on a propietary thing

09:33

it makes me angry xD

09:33

so to get llvm-config i would need to build llvm on windows myself

09:33

= lose 2 hours

Go See FISHYYYYYYYYYYYYS If angry ryo - they Help!

Ryozuki

it makes me angry xD

^^

Lmao my Phone autocorrected that

09:34

I Just Typed fishies

macos support is easy tho

09:34

its just windows is so special

Ryozuki

macos support is easy tho

do you even need support for anything other than ryoos?

for now ill only support FOSS operating systems

09:35

nobody will use this anyway

xd

gotta check openbsd

u never know

well then my goal will be to make such a good programming language it will entice ppl to come to linux

09:36

isnt that what windows does with games?

09:36

owo

meloƞ

Milkeye????? Wtffff actual useful commit huh ?? You sick or smth

https://discord.com/channels/252358080522747904/745926398140612678/1211381450587246672

KEKW

gigachad

github-merge-queue[bot]

[ddnet:master] 2 new commits

b6c75e1 ci: remove useless command in ci - MilkeeyCat ae5bf90 Merge pull request #8018 from MilkeeyCat/pr_remove_useless_command_in_ci - def-

09:40

[ddnet/ddnet] Pull request closed: #8018 ci: remove useless command in ci

Did my pr just got merged in main repo? santatrollet

santatrollet

Image attachment

I thought u just edit to "contribute code" xd

09:58

But having dev role feels good

it's a lot harder for windows-exclusive game devs to target Linux than it is for you to target windows lol

10:00

though i understand the frustration of having to compile stupid shit

10:01

rn im compiling llvm-mingw on my old 6 core i5 because it doesn't have a build for my distro

10:03

also congrats milkyboy

I didn't code anything ;-;

MilkeeyCat

But having dev role feels good

trollet

MilkeeyCat

I didn't code anything ;-;

if you feel bad about that, just change it?

Jupstar ✪

if you feel bad about that, just change it?

Yea.. i should

MilkeeyCat

Yea.. i should

try to find out how to save teams on map change

10:12

good feature but naive implementation sucks

MilkeeyCat

Did my pr just got merged in main repo? santatrollet

santatrollet

am I a dev then aswell justatest

justatest

MilkeeyCat

Yea.. i should

wait wait wait. you now have dev role?!

yippee

im still chilling as normal user

10:54

.....

That didn't age well

Jupstar ✪

That didn't age well

xd

someone is feeling generous

Jupstar ✪

That didn't age well

my ages since 2020 i suppose

[ddnet/ddnet] Pull request opened: #8019 Rename all variables for strict camel-casing of abbreviations

This is the strict version, ID → Id, UI → Ui, except DDNet which stays DDNet. This would fix #7750. Done using a naive rename script: ```fish sed -i \ -e 's/([a-z]?)ID/\1Id/g' \ -e 's//Ui/g' \ -e 's//Id/g' \ -e 's/([a-z]?)GPU/\1Gpu/g' \ -e 's/([a-z]?)IP/\1Ip/g' \ -e 's/([a-z]?)CID/\1Cid/g' \ -e 's/MySQL/Mysql/g' \ -e 's/([a-z]_?)SQL/\1Sql/g' \ -e 's/DPMode/DpMode/g' \ -e 's...

Teero

wait wait wait. you now have dev role?!

Cool I have a Dev role now but I don't feel worthy yet. Guess I'll have to fix some things

GitHub

Click to see attachment 🖼️

@meloƞ

GitHub

Click to see attachment 🖼️

Write a tool for my ide that can auto fix it. Else its just style bitching

fix_style.py should be enough

11:33

No ide integration needed

Conventions that are enforced theoretically are useful for programmers

chillerdragon

No ide integration needed

I disagree

11:33

Without ide support it still results in force pushes which slow down the working process

11:35

https://doc.rust-lang.org/std/collections/struct.BTreeMap.html

BTreeMap in std::collections - Rust

An ordered map based on a B-Tree.

11:36

Not even rust follows its own rules

11:36

Btree should be one word in their logic

Send pic with tw open and I can include it in ma new YouTube video c:

Replying to @Ewan IMG_4952.jpg

The module is also called like that

Jupstar ✪

https://doc.rust-lang.org/std/collections/struct.BTreeMap.html

@heinrich5991 statement

For writing tests and documentation for the 0.6.4 protocol. I found it easier to work with crc maps for that.

Replying to @heinrich5991 ChillerDragon: currently just not update. what do you use it for?

if you want permalinks, use the sha256

I have a compilation of pcap files and python tests and html files etc where i would like to reference that this is downloading map xyz with crc so and so and it can be downloaded from this heinrich5991.de url

11:39

but 0.6.4 uses crc

then mention the sha256 of the maps that were downloaded

11:39

to get a permalink

<ChillerDragon> its more complicated to verify then https://chillerdragon.github.io/teeworlds-protocol/06/system_messages.html#NETMSG_MAP_DATA

11:40

<ChillerDragon> here it sends a crc

11:40

<ChillerDragon> crc is shorter to type creates shorter file names in my pcap collection

11:40

<ChillerDragon> and i do not need to translate the sha to crc

all the wrong reasons. if it motivates you, I can delete the crc endpoint

<ChillerDragon> nah just keep one forever :p

11:41

<ChillerDragon> its annoying to use sha when the whole thing is about crc

11:41

<ChillerDragon> just feels wrong

sha256 is just another file checksum

11:42

(that is actually secure)

<ChillerDragon> yes i understand

11:42

<ChillerDragon> but its still additional complexity where i really do not want it

11:42

<ChillerDragon> for no gains

the gains are that people can actually verify they got the correct map

<ChillerDragon> its fine i can just build my own crc mirror then i guess

11:43

<ChillerDragon> crc is good enough for that in a sample map

it's usual, even without any protocol shenanigans, to provide sha256 hashes of files

<ChillerDragon> yes i know

see e.g. https://ddnet.org/downloads/sha256sums.txt

<ChillerDragon> but crc is baked into the protocol

11:43

<ChillerDragon> so thats the interesting value

and you can additionally specify the sha256 (edited)

<ChillerDragon> when i am documenting the protocl

11:44

<ChillerDragon> then the url is still non obivious

11:44

<ChillerDragon> because it does not match

then build your own mirror, please

<ChillerDragon> ye k

I'm still in trouble about if I prefer blake2 or 3 vs sha3

11:44

Chiller opinion?

what do you want to optimize for?

11:45

I heard blake3 is good for speed

Yes

11:45

But not standard

<ChillerDragon> it didnt use it yet for anything was just thinking about doing so @heinrich5991 so from my side you can delete the crc folder but i always liked using that for testing etc. Espeically since many clients still only print crc in the logs

But sha3 adaption sucks too

11:46

So speed might be more convincing xd

<ChillerDragon> i dont even know blake

11:46

<ChillerDragon> i use sha1 mostly trololol

11:46

<ChillerDragon> because it is installed on most systems

you could use sha256 if you want to have a standard thing

<ChillerDragon> sha1sum

But I heard sha3 can be extremely fast too on fpga

heinrich5991

you could use sha256 if you want to have a standard thing

Old is always bad

<ChillerDragon> sha256sum is always trouble on figuring out where to install it from and then you have to think about portability for other ppl

11:47

<ChillerDragon> as someone who uses shell as their main language xd

Only the newest of the newest is good enough

soo, we should move to teeworlds 0.7 protocol?

11:47

no, newest is not always best

<ChillerDragon> now

How is that new?

<ChillerDragon> im ready

It's like 10 years old

<ChillerDragon> ipv6

afaik NIST did not create a new hash function because the old one was obsolete

But in case of it getting unstable

<ChillerDragon> arbeitsbeschaffungsmaßnahme

11:48

<ChillerDragon> some1 was bored and built new hash

As backup

yes. just like sha2 might be there if sha3 gets broken

So Blake it shall be u say?

<ChillerDragon> why do you need standard in your case?

blake is also parallelizable

11:49

so you can use all your cores to compute it

ws-client

<ChillerDragon> why do you need standard in your case?

I don't need it

<ChillerDragon> if its fastest and u have a good lib in rust for it that you can ship to client and server it seems like the best fit for ddpg2

It's just an pro argument

11:50

Not a decision making argument

ws-client

<ChillerDragon> if its fastest and u have a good lib in rust for it that you can ship to client and server it seems like the best fit for ddpg2

OK boss

<ChillerDragon> im so curious if u ever reach playable state

11:51

<ChillerDragon> that would be crazy

11:51

<ChillerDragon> ryo released a full programming language already axaxax

11:51

<ChillerDragon> and there is no alpha rls of jupjapjapsti-net2

Why should I not

11:52

But only vanilla as first step

<ChillerDragon> idk its big project

But the engine is almost done

<ChillerDragon> pog

Vanilla itself is actually not lot of code

11:53

I just hate it xd

<ChillerDragon> rude

But thanks to zwelf I have a new door open

<ChillerDragon> wot

11:53

<ChillerDragon> u use his lib?

11:53

<ChillerDragon> isnt there license issues?

I can break some tw "physics" without breaking ddnet physics

<ChillerDragon> woah

11:54

<ChillerDragon> 0.5 wall hammer

And I explicitly don't mean core physics

11:54

Yeah the way projectiles are handled and stuff

11:54

It frustrats me a bit

ws-client

<ChillerDragon> u use his lib?

Not yet

11:55

License is a kinda problem

<ChillerDragon> lawyer jopsti moment

I think it would be nice if ppl can still base mods based on ddnet physics

11:55

But generally physics are completely separated

11:55

So u could say it will simply not live in the main repo

<ChillerDragon> this all sounds very annoying already

I agree

<ChillerDragon> BSD masterrace

Easiest would be to not use copyleft

<ChillerDragon> yes

11:56

<ChillerDragon> using copyleft is supporting peophelia

Xd

<ChillerDragon> omg i typoed that word hard

11:57

<ChillerDragon> probably for the better

It's just so complicated to understand

<ChillerDragon> ye

Ofc proprietary is not better xd

11:57

Or even worse

<ChillerDragon> iirc BSD was literally created to not deal with any legal questions

I'd like to have a citation for that

<ChillerDragon> for what=

[iirc] BSD was literally created to not deal with any legal questions

<ChillerDragon> "iirc"

yes. do you remember correctly?

<ChillerDragon> idk where i heard that

I'm interested in that

11:58

ok

<ChillerDragon> usually not

12:01

<ChillerDragon> I did a quick search didn't find anything. I think i heard that in some tech talk about one of the BSD operating systems like OpenBSD or FreeBSD

12:02

<ChillerDragon> Something about the devs wanting to be let alone and just publish their code

U could probably argue that using BSD license means you also care less about the legal consequences

12:05

I doubt smone with BSD would try to sue u. Or is less motivated than a GPL guy

afaik linux hasn't sued anyone

Epyc

couldn't find anything quickly using "linux epyc" or "linux epyc lawsuit"

12:07

do you have a link?

Epyc future human moment

12:07

Try that

Ddnet has a bsd build?

lmao

Replying to @Jupstar ✪ Epyc future human moment

12:27

I am not a lawyer. But I’d say ddnets license is close to bsd

Replying to @ReiTW Ddnet has a bsd build?

Zlib license

Thanks lawyer jopsti

12:29

Whatever the diff between zlib and bsd is

I think that's hard to say. Zlib is not very common. So I guess you'll not find many answers on that

[ddnet/ddnet] Pull request closed: #8007 Respond to pings sent as vital messages with vital pongs

12:39

github-merge-queue[bot]

[ddnet:master] 2 new commits

e69ea46 Respond to pings sent as vital messages with vital pongs - heinrich5991 2258cbc Merge pull request #8007 from heinrich5991/pr_ddnet_vital_ping - Robyt3

Chiller when are you gonna release chillerdragores16?

@Ewan got new speakers

13:13

these have better lows im sure

13:13

and they like 2x bigger xD

awesome

13:13

which speakers

im sure u will say they bad

13:13

@Ewan i didnt know these speakers used cable directly

13:13

to connect each other

13:14

u know like bare cable

yea

13:14

i hope that isn't a problem for u

Image attachment

13:14

nope it works

oh you got actual monitors

13:14

awesome

what does that mean

When go DM

i also bought a table xD

13:14

wanted to dabble a bit

table looks cool

no

13:15

i mean dj one

13:15

not the wood table

13:15

xd

uhh

13:15

like a turntable?

mixing table

13:15

i thin its the name

WTF

u know

13:15

idk bro

13:15

its 100 €

13:15

nothing fancy

fotos?

this one

Image attachment

13:16

it works with mixxx

oh like a cdj

13:17

nice

13:17

do you do any pro audio stuff already

what do u consider pro audio stuff

13:17

i mixed a bit before (or tried)

13:17

with software

13:17

mixxx

i see

@Ewan looked before and made sure this table was full usb midi compliant or smth

cool man

so it works with linux

nice

13:18

ive never used mixxx before so idk

i think its nice

13:19

its open source

i never mixed as like a standalone thing

13:19

always other parts involved

Image attachment

i see

13:20

i just use fl studio mixer

13:20

it’s goated…

isnt it more to produce music?

13:21

idk but fl sutdio is propietary

13:21

im a free software dj

13:21

gigachad

13:22

https://github.com/LMMS

free cross-platform software which allows you to produce music with your computer - LMMS

vile piece of software

trait Trait {
    fn abs(self) -> Self;
}

impl Trait for i64 {
    fn abs(self) -> Self {
        2 * self
    }
}

fn main() {
    let x = 42;
    println!("{}", x.abs());
    println!("{}", x.abs());
    println!("{}", x.abs());
}

New Rust riddle

with non-suggestive print calls: What will this program output?

13:34

ah forget the second and third print for a sec

42 * 2 ?

13:35

default int is i64 iirc

13:35

or was it i32

default is i32

i32 ye

13:35

so its 42

13:35

xd

and now ofc: not all print statements print the same xd

or maybe print makes some cast

13:36

ra says its i64

well casts would need to be explicit in Rust

im sure the print makes it change

13:36

the infered type

13:37

#![feature(prelude_import)]
#[prelude_import]
use std::prelude::rust_2021::*;
#[macro_use]
extern crate std;
trait Trait {
    fn abs(self) -> Self;
}
impl Trait for i64 {
    fn abs(self) -> Self {
        2 * self
    }
}
fn main() {
    let x = 42;
    {
        ::std::io::_print(format_args!("{0}\n", x.abs()));
    };
    {
        ::std::io::_print(format_args!("{0}\n", x.abs()));
    };
    {
        ::std::io::_print(format_args!("{0}\n", x.abs()));
    };
}

13:37

i think format args is compiler builtin

Answer: 84 42 42

13:39

Relevant Rust issue (contains spoiler): https://github.com/rust-lang/rust/issues/121453

damn

awesome

subscribed

Patiga

Answer: 84 42 42

Oh, all 42 would probably still make sense

Jupstar ✪

Oh, all 42 would probably still make sense

ye

13:44

its a inference bug

13:45

"bug"

13:45

idk if they consider it a bug

13:45

I guess the type infer cause first x.abs() call search abs method and Trait::abs is found. Once the type of x is inferred to be i64, then it will use core::num::abs.

(edited)

imo this is clearly a bug ^^

Well I don't even understand why it's casted to i64

u dont specify a type

13:46

its never casted

13:46

its infered

I'd have expected i32::abs to be called

it infers it to be i64 because it sees a possible trait method i think

13:47

and then not and it uses i32 abs

13:47

weird stuff

13:47

idk

it uses i64 abs after

It seems only the first call in a block will be resolved to Trait::abs and all the calls after in the same block to the latter.

I still find that behavior weird

yea, once iti s infered as i64, it uses i64::abs

13:48

*its

13:48

yea def a bug

Before changing to i64 I expected it to search all i32 traits

13:48

And then i32::abs would be found

or, alternatively, the trait method all the way

@Patiga thanks for posting always interesting this kind of things

13:49

and i quizzed my coworkers

13:49

kek

nice ^^

@Ewan do u know any good dac/amp with rca out

14:29

since this speakers have separate left right rca

14:29

the fioo i have doesnt have rca

14:29

what u think of Fosi Audio K5 Pro

14:32

iFi ZEN DAC V2 looks good too

ok got some already

[ddnet/ddnet] Issue opened: #8020 segafult when create a bezier curve and right click on it

1. Go to editor 2. Open envelope tab 3. Click on "Color +" 4. Set curve to bezier 5. Rick click on point

@MilkeeyCat ur time to shine

15:52

‍♂️

Jupstar ✪

@MilkeeyCat ur time to shine

I spent 3 hours in editor code and an ready to kms

Yesterday you said tomorrow

15:53

Stop giving up

Nah, tomorrow xd

Jupstar ✪

Stop giving up

U wanna help me? owo

owo

15:55

I know what the problem is, u just need to fix it

GitHub

Click to see attachment 🖼️

MilkeeyCat

I spent 3 hours in editor code and an ready to kms

please don't touch editor code too much

archimede67

please don't touch editor code too much

I wrote more dbg_msg's in that bich than in my entire life (edited)

3

Is the convex hull of our bezier curves always inside the time frame between two animation points? (edited)

Ryozuki

this one

poggers2

[ddnet-web:master] 1 new commit

6992730 Add 28 € donation by Hamelin - def-

1

what is this black magic.

Image attachment

Image attachment

ub?

Maybe the vector already has capacity so it's not technically out-of-bounds of the underlying array, or it doesn't crash immediately if you only try to print from an invalid address

if you working in scoreworker.cpp if(w == Write::NORMAL){} I always see it but never cant interpret it 100%. Is this something like Post trigger? like NORMAL_SUCCEEDED and NORMAL_FAILED are self explained

MilkeeyCat

I wrote more dbg_msg's in that bich than in my entire life (edited)

this is cpp not js

17:30

but to be fair im doing to too like that xD

Robyt3

Maybe the vector already has capacity so it's not technically out-of-bounds of the underlying array, or it doesn't crash immediately if you only try to print from an invalid address

i can write to it as well

17:31

ill call it a feature and go back to writing hello world in c++

zhn

@meloƞ

Damn - should we bother ?

Teero

Cool I have a Dev role now but I don't feel worthy yet. Guess I'll have to fix some things

lol gg wp

@Ewan https://www.youtube.com/watch?v=5QHVvJsn2Ug

And One - Topic

Love You to the End

18:01

listen to this

[ddnet-discordbot:master] 1 new commit

08af8b8 map-testing/auto_archive: Continue if MapState == RC - murpii

Did you know protobuf does not have fixed size array support?

github-merge-queue[bot]

[ddnet:master] 3 new commits

268b79c handle invalid filename for save_replay - dobrykafe 91349e4 add error handling for CDemoEdit job - dobrykafe 7c4d51f Merge pull request #8016 from dobrykafe/pr-save-replay-improve - Robyt3

19:07

[ddnet/ddnet] Pull request closed: #8016 Improvements regarding save_replay

Why would anyone make a binary packing format and not have a way of including a fixed size array?

[ddnet/ddnet] Issue closed: #8012 add error handling when renaming fails using save_replay

With the release of electron, I think I just ended up on the wrong side of a multiverse split

19:17

There is no u8 either, I simply do not get modern software at all nor do I get the mindset in which it is designed

zhn

try to find out how to save teams on map change

oh is that the: redirect players on map change thingie?

chillerdragon

fix_style.py should be enough

i wish someone would finally update the script so it works with clang-format 19, clang-format 10, its not even supported by most package manager's anymore OHNOOOO

OHNOOOO

meloƞ

i wish someone would finally update the script so it works with clang-format 19, clang-format 10, its not even supported by most package manager's anymore OHNOOOO

OHNOOOO

We are trying to avoid large formatting commits and I think current clangformat also misformats some of our code

what about just shipping ddnet with a standalone version of clang-format 10 that it refers to?, or will that interfere with IDE's ?

https://github.com/muttleyxd/clang-tools-static-binaries this is where I refer people to and there is an AUR package that installs these into /opt/

GitHub - muttleyxd/clang-tools-static-binaries: static-linked versi...

static-linked versions of clang tools for Linux, Mac OS X and Windows, available in AUR as clang-format-static-bin - muttleyxd/clang-tools-static-binaries

idk if i should refer to clang-format 10 as a stand-alone tool or a lib

Learath2

https://github.com/muttleyxd/clang-tools-static-binaries this is where I refer people to and there is an AUR package that installs these into /opt/

oh this is.. nice - thanks :D

https://capnproto.org/ I found an alternative built by a person that is actually sane

19:34

Not at all surprised that they work at cloudflare, seems to be the last bastion of good engineering these days

[ddnet/ddnet] Pull request opened: #8021 Various fixes for envelope and bezier evaluation

See commit messages. Closes #8005. Screenshots:

Example 1:

Before:

!screenshot_2024-02-26_20-44-03

After:

!screenshot_2024-02-26_20-46-26

Example 2 (you can see the missing first line segment here):

Before:

![screenshot_2024-02-26_20-44-18](https://github.com/ddnet/ddnet/assets/23437060...

Learath2

https://capnproto.org/ I found an alternative built by a person that is actually sane

we talked about this several times in this channel xd

20:23

@Learath2 the author made or used to do protobuf at google

20:23

but was fired iirc

20:23

idk

He used to work in the protobuf team indeed

20:24

He was probably too sensible for google

Image attachment

youtube having hard time

21:25

kek

github-merge-queue[bot]

[ddnet:master] 5 new commits

5aab996 Fix first envelope line segment not being rendered in editor - Robyt3 cc1d43e Fix division by zero when envelope point times overlap - Robyt3 47a8156 Fix inaccurate evaluation of bezier curve envelopes - Robyt3 ecfc18d Minor refactoring of bezier curve evaluation - Robyt3 758664c Merge pull request #8021 from Robyt3/Envelope-Render-Fixes - archimede67

22:09

[ddnet/ddnet] Pull request closed: #8021 Various fixes for envelope and bezier evaluation

22:09

[ddnet/ddnet] Issue closed: #8005 Incorrect evaluation of bezier curves in rare cases

Ryozuki

Click to see attachment 🖼️

who is this memer

lmao why? I forgot which one that was do you even know it? I think it got rejected. Only maps I’m working on these days are ChillerDragores2 and ChillerDragores6

Replying to @Teero Chiller when are you gonna release chillerdragores16?

22:59

lmao

Replying to @Ryozuki GHQQCArWkAADRb_.png

Exported 543 message(s)