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 2023-08-30 00:00:00Z and 2023-08-31 00:00:00Z

did TextRender::Text signature get changed recently?

02:37

i feel like it had auto height before and not now

oh it’s the silly ptr

art0007i

hey guys I was trying to compile ddnet on windows but when compiling the client specifically (server and most other things work fine) I get a ton of these weird errors and I have no idea what they mean. can anyone help me?

how is it configured? did you look at the build instructions for Windows? at a glance this seems like a config mismatch error to me, possibly targeting an older standard version

03:00

which might happen if you didn’t use cmake

03:01

either generate for vs 2022 w/ the cmake gui or cmd line or whatever or open it as a cmake project in VS (build instructions don’t tell you to do this but it works better and lets you make multiple configurations, you have to click “continue with no code” and a cmake project is in file->open) (edited)

ReiTW

wtf is that, looks like a tuto for newbies who gets hired at valve working on cs2 (edited)

True

03:08

Tho not rly cpp

03:15

Maybe it would be worth using CUIRect in the HUD and whatnot instead of hardcoded offsets, would make finding space for new things much easier

03:16

rn with my status bar if i wanted to have it on the top i'd have to manually add the status bar height to the y of every other unrelated element that just happens to be there

03:16

but the ui could just split a shared rect and then the next whoever uses that

i didn't even look at menus_settings_assets until now. this is gonna take a while, jesus

[ddnet-maps:master] 1 new commit

2709876 A ddrace_shella2, A ddrace_shella3, A ddrace_shella4 - ddnet-maps

metamuffin

i am (somewhat) sure that if they are in the same workspace, the union of all features will be enabled.

Who are you btw. Never seen you before xd

Serde is rarely a problem cuz they are in 1.0 so all 1.0.y where y<x will be able to use 1.0.x

06:36

And there is no 2.0

good morning

0.x.y is different cuz x are not back compat to semver on 0.

06:37

Cargo tries to resolve versions to be same, but if it fails the crate will have 2 versions

Only on 0 ?

And if libs use dif rust will tell u when u use a type on the lib from another version

06:38

0.x.y x here is minor but on 0. Minors are breaking

06:38

Only patch is not

smh

Didn’t know that

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

SemVer Compatibility - The Cargo Book

Looks like a cheese

jeez i just initialized a std::array of size 6 with 5 items

06:40

it should warn for that

06:41

"why it crash"

?

06:43

C++?

06:44

his guide uses the terms “major” and “minor” assuming this relates to a “1.0.0” release or later. Initial development releases starting with “0.y.z” can treat changes in “y” as a major release, and “z” as a minor release. “0.0.z” releases are always major changes. This is because Cargo uses the convention that only changes in the left-most non-zero component are considered incompatible.

yea cpp

06:45

xd

Ewan

rn with my status bar if i wanted to have it on the top i'd have to manually add the status bar height to the y of every other unrelated element that just happens to be there

Try do keep other changes minimal if possible or move them to a separate PR. IMO the loading indicator and text should be shown as a separate list item inside the listboxes which are currently getting loaded and not in a separate status bar, so you shouldn't need to adapt CListBox in any major way. I already have a loading indicator implemented, it's just waiting for stuff that can load in the background. I can make a PR in the evening if you want to start using it.

Ewan

how is it configured? did you look at the build instructions for Windows? at a glance this seems like a config mismatch error to me, possibly targeting an older standard version

I used cmake and everything correctly just like in the readme, ended up reinstalling visual studio and it fixed it, was totally a problem with my msvc installation. (though weird that it compiled the server with no issues, but completely broke on the client)

client has more tricky deps

08:45

jopsti when add vulkan to the server?

Sounds good

08:51

More fps for server

kek

1

Server Side Rendering for faster first frames

gpu optimized 512 tee collision physics

compute shaders

Learath2

Server Side Rendering for faster first frames

take that bot-clients!

https://www.phoronix.com/news/Linux-6.6-Illicit-NVIDIA-Change

Linux 6.6 To Better Protect Against The Illicit Behavior Of NVIDIA'...

The Linux 6.6 modules infrastructure is changing to better protect against the illicit behavior of NVIDIA's proprietary kernel driver.

11:29

https://www.servethehome.com/intel-shows-8-core-528-thread-processor-with-silicon-photonics/

Patrick Kennedy

Intel Shows 8 Core 528 Thread Processor with Silicon Photonics

Intel showed an 8 core CPU with 66 hardware threads per core for 528 threads and it even included silicon photonics networking

[ddnet/ddnet] Issue opened: #7107 Some tooltips don't work anymore.

I found that the following tool tips don't seem to work anymore,

"Your movements are not taken into account when calculating the line colors" on the hook collisions page.
"Toggle to edit your dummy settings" on the player page.

how can i add multeasy map in my server

download the map

12:04

put it in your map foler

7mzom alonso

how can i add multeasy map in my server

Download the map from another server or https://github.com/ddnet/ddnet-maps

GitHub - ddnet/ddnet-maps: All released maps on DDraceNetwork, see ...

All released maps on DDraceNetwork, see also https://ddnet.org/releases/ and https://maps.ddnet.org/ - GitHub - ddnet/ddnet-maps: All released maps on DDraceNetwork, see also https://ddnet.org/rele...

use sv_map

12:04

https://github.com/ddnet/ddnet-maps/raw/master/types/novice/maps/Multeasymap.map

Guys, what's better, amd or intel?

12:26

Intel seems better

12:26

According to the first link i clicked

From my experience intel is less energy efficient with cpus

But does energy consumption really matter?

If you don’t care about your electricity bill then not really

in battery-powered devices, energy consumption matters a lot to yourself

Right

12:31

I care more about performance and features

Idk it seems like intel and amd are constantly swapping places on who’s better

12:32

Just watch and read some reviews

12:33

On the specific hardware that you want to buy

Alright, i guess it doesn't really matter

I assume you’re talking about cpus

12:35

If you were to talk about gpus then it’s a whole different story

Yeah cpus

12:35

Nvidia is the best for gpus i think

Imo nvidia is definitely the best just purely because of the software (edited)

But i hate their proprietary software

12:35

I mean i have to as a linux user lol

Idk about linux

12:36

Im proud windows user

~~me too but recently i use more linux than windows~~

12:39

I think i need to choose a cpu first

12:39

And then the rest of the components

Depends on what you’re going for, but I’d recommend starting with gpu, since it’d be the most expensive thing

I guess on steam nvida rtx 3060 is popular

Mid range cpu or high end?

12:44

For high end I'd say, u can go with both. For mid, I'd go with amd

Idk yet

12:45

I guess mid range

12:45

Because i am not rich

12:45

Or something

Amd often has better perf/$

12:46

Do u plan to buy inside poland or germany?

Poland?

Well EU.. u can buy from anywhere xd

I guess i want to have an average pc because games (software in general) probably want to work on average hardware?

Sure. I just wanted to know how much€ you pay for which CPU

12:53

And then I'll check what is better, amd or Intel. And where, shop

I wonder if pre built pcs are any good

12:58

They should technically be good in terms of components

@Jupstar ✪ do u know where is the code to render a tee

13:05

oh

13:06

i know where to find this

Robyt3

Try do keep other changes minimal if possible or move them to a separate PR. IMO the loading indicator and text should be shown as a separate list item inside the listboxes which are currently getting loaded and not in a separate status bar, so you shouldn't need to adapt CListBox in any major way. I already have a loading indicator implemented, it's just waiting for stuff that can load in the background. I can make a PR in the evening if you want to start using it.

I think we are talking about different things but that might be useful yeah

im finally finishing teerender

Nice

pilonpl

I wonder if pre built pcs are any good

I haven't seen one at a fair price since years tbh

13:09

10+ years ago aldi made ones xd

Image attachment

13:22

honestly its hard to interpret the source code xd

[ddnet/ddnet] Pull request opened: #7108 Update turkish.txt

Fixes:

kesim -> kesit
Missing colons on L1133, L1136, L1139
Typo on L1157 UPD -> UDP
Missing translation on L1175
L1241 oyuncular -> oyuncu
L1475, L1478 shortened to fit ingame

@Learath2

Ryozuki

honestly its hard to interpret the source code xd

Where did you struggle. Eyes? Xd

It'd be easy to read if it were Rust

Xdd

Learath2

It'd be easy to read if it were Rust

well the tee render is mixed with python

13:27

Image attachment

13:27

body is too up right

13:27

Image attachment

13:27

in game

13:28

    // 256, 128
    let img = ImageReader::open("skins/brownbear.png")?.decode()?;
    dbg!(img.width());
    dbg!(img.height());

    let mut def = RgbaImage::new(96, 64);
    dbg!(def.width());
    dbg!(def.height());

    let filter = FilterType::Triangle;

    let back_feet_shadow = img.crop_imm(192, 64, 64, 32).resize(64, 30, filter);
    blend_region(&mut def.sub_image(8, 32, 64, 30), &back_feet_shadow);

    let body_shadow = img.crop_imm(96, 0, 96, 96).resize(64, 64, filter);
    blend_region(&mut def.sub_image(14, 0, 64, 64), &body_shadow);

    let front_feet_shadow = img.crop_imm(192, 64, 64, 32).resize(64, 30, filter);
    blend_region(&mut def.sub_image(24, 32, 64, 30), &front_feet_shadow);

    let back_feet = img.crop_imm(192, 32, 64, 32).resize(64, 30, filter);
    blend_region(&mut def.sub_image(8, 32, 64, 30), &back_feet);

    let body = img.crop_imm(0, 0, 96, 96).resize(64, 64, filter);
    blend_region(&mut def.sub_image(14, 0, 64, 64), &body);

    let front_feet = img.crop_imm(192, 32, 64, 32).resize(64, 30, filter);
    blend_region(&mut def.sub_image(24, 32, 64, 30), &front_feet);

    let left_eye = img.crop_imm(64, 96, 32, 32).resize(26, 26, filter);
    blend_region(&mut def.sub_image(36, 17, 26, 26), &left_eye);

    let right_eye = left_eye.fliph();
    blend_region(&mut def.sub_image(36 + 9, 17, 26, 26), &right_eye);

    def.save("result.png")?;

@Learath2 well rust is easy as this

13:29

honestly i just glossed over the code

13:29

its too much effort to figure the legit

13:29

way

13:29

xd

Jupstar ✪

I haven't seen one at a fair price since years tbh

Yeah, price can be unfair but the components should work together nicely, right?

13:30

Like someone chose these components

Ryozuki

its too much effort to figure the legit

Cheater

@Jupstar ✪ tell me exact coords of tee rendering and ill make u a free software service API for everyone to render skins with given colors

13:30

and also a fast rust cli tool

13:30

blazing fast

pilonpl

Like someone chose these components

Well, trust me if i say they won't work together better than if u look into the compatibility list of your motherboard manual

13:36

Prebuilt pcs tend to use cheapest hardware

Ryozuki

@Jupstar ✪ tell me exact coords of tee rendering and ill make u a free software service API for everyone to render skins with given colors

I posted a type script once that works

mine also "works"

Maybe u can find it xd

13:38

Else see patigas txt

where is it

Dunno ask him

@Patiga u got a tee skin rendering txt?

request for tee.md in libtw2/doc

13:39

or anywhere else

I have a question: Will a server with more than 128 maxplayers be masterban?

no

13:40

u will be banned if u say ur server has 128 players but they arent real

13:40

connect

13:40

connecteD*

13:40

but there is nothing wrong with 128 player slots

13:40

fokko does that

generally, it'd be nice to send correct info

13:41

there are no hard rules, but it'd be nice if your server actually supported 128 players then

Thanks, i found a way to make it.

13:45

Also, i think it can be useful for a PvE mode

ddnet is pve

13:45

its pvm

13:45

player vs map

13:45

xd

wow.

how to convert hsla to rgba

13:46

i hate colors

Me too xd

ok enough work today on teerender

13:47

i dont wanna deal with colors xd

Xddd

RemakePower

Also, i think it can be useful for a PvE mode

Imagine that the monsters on the map are all the same Tee as the players. The number of monsters and number of players add up to much greater than 64.

13:48

That's interesting.

Ryozuki

how to convert hsla to rgba

We have it in color.h

source is here btw https://github.com/edg-l/teerender

GitHub - edg-l/teerender: A blazingly fast tee skin renderer web se...

A blazingly fast tee skin renderer web service, with custom colors. - GitHub - edg-l/teerender: A blazingly fast tee skin renderer web service, with custom colors.

Or you can look at wikipedia

lerato our davinci

im not even sure my function to convert a color code to hsl works

13:48

xd

nice commit message

xd

#[allow(clippy::cast_precision_loss)]
fn color_code_base(col: u32, alpha: bool) -> HSLAColor {
    let a = if alpha {
        ((col >> 24) & 0xFF) as f32 / 255.0f32
    } else {
        1.0f32
    };
    let x = ((col >> 16) & 0xFF) as f32 / 255.0f32;
    let y = ((col >> 8) & 0xFF) as f32 / 255.0f32;
    let z = (col & 0xFF) as f32 / 255.0f32;

    HSLAColor { x, y, z, a }
}

13:50

xd

That looks fine to me

.0f32 is cursed

13:53

not as cursed as . being f32 and .f being f16 tho

omg these codes

13:54

are these java

no

13:54

these are rust (edited)

f32 suffix is unnecessary there

1

c++

XWA

c++

They are rust (edited)

i don't think c++ has 'let'

#define let auto

Then I don't think C++ returns values by this way.

Ryozuki

#[allow(clippy::cast_precision_loss)]
fn color_code_base(col: u32, alpha: bool) -> HSLAColor {
    let a = if alpha {
        ((col >> 24) & 0xFF) as f32 / 255.0f32
    } else {
        1.0f32
    };
    let x = ((col >> 16) & 0xFF) as f32 / 255.0f32;
    let y = ((col >> 8) & 0xFF) as f32 / 255.0f32;
    let z = (col & 0xFF) as f32 / 255.0f32;

    HSLAColor { x, y, z, a }
}

python?

Matodor

python?

rust.

troll

1

Ryozuki

@Patiga u got a tee skin rendering txt?

567 bytes

Patiga

Click to see attachment 🖼️

what's this?

epic

Patiga

Click to see attachment 🖼️

@heinrich5991

the values required to render a tee properly

why are eyes positions not fractions?

@Ewan "#define let auto" lmao epic

Matodor

python?

rust bro

Mr.Gh0s7

why are eyes positions not fractions?

first is the offset of the tees, the point around which they rotate. afterwards its the movement of the eyes which is calculated with the view angle

13:59

*of the eyes

ChillerDragon

@Ewan "#define let auto" lmao epic

ya it’s a feature

Ah ok ;)

Patiga

Click to see attachment 🖼️

i know we dont pin much but can someone pin this

I'll create a libtw2 pull request

Ewan

ya it’s a feature

I've seen someone define emojis to "int", "float" before

https://wiki.ddnet.org/wiki/Development#About_Tee_Skin_Rendering

1

14:02

i added it here

14:02

too

wow

RemakePower

wow

whats wowy

i didn't know the wiki has it before

it didnt have the txt from patiga, i added it just now

oh

Patiga

Click to see attachment 🖼️

can I have that under the libtw2 license (Apache2/MIT)?

sure

thanks ❤️

heinrich5991

can I have that under the libtw2 license (Apache2/MIT)?

are you adding it rn? else I'd do a pull request

Patiga

are you adding it rn? else I'd do a pull request

you can do a PR

14:22

actually

14:22

I was about to do it

14:22

what do you prefer? @Patiga

you do it, I think its fastest and its your repo :)

14:23

(-> so you know how you want it formatted etc)

Wait, what wrong with ddnet.org?

14:26

hm?

14:26

loaded now

https://github.com/ddnet/ddnet/pull/6727#discussion_r1309811240

'Mass file loader' for use in skin and asset loading by ewancg · Pu...

Summary: Currently, asset and skin loading are quite slow if you have a decent amount of them, even on a fast disk. I would like to create a generic, multi-threaded file loader system that can less...

14:33

i get it for little routines where the string goes out of scope quick but to always have 256 bytes allocated when it could be 5? idk man (edited)

@Patiga eye_blink isn't real? TIL

all lies!

the github app is so unintuitive

@Patiga is blink only vertically scaled or both ways?

only vertically

the rest both ways?

hm yes

14:50

during a blink, the eyes are probably 45% vertically, 120% horizontally (normal eyes value)

14:51

^ yes

this fine? @Patiga https://github.com/heinrich5991/libtw2/commit/044597a89faa5afaaa01e6cb2fa100d94c806738

doc: Add tee_rendering.md describing how to render a tee · heinrich...

Special thanks to Jupstar.

14:53

including using you as the author

looks good!

ok, pushed

heartw

1

Nice

15:12

Do we want to add a flag to io_read_all to differentiate between files of 0 bytes and files >2GiB

epic

Ewan

Do we want to add a flag to io_read_all to differentiate between files of 0 bytes and files >2GiB

Or just always return -1 for sz or w/e if over 2GB (edited)

Ewan

Do we want to add a flag to io_read_all to differentiate between files of 0 bytes and files >2GiB

files with >2GiB should probably not be loaded into memory as a whole. if you still want to fx that function, make it return uint64_t I guess

15:17

I'm fine with w/e for these large files

Seems stupid to remove safer code just because it’s not done elsewhere

then put it into io_read_all

maybe io_length or w/e

You'd probably have to change all of the io offset functions if you want to support largers files since we use fseek

I wouldn't call code that loads arbitrary amounts of data into RAM safe, but it's not worse than the current code

heinrich5991

files with >2GiB should probably not be loaded into memory as a whole. if you still want to fx that function, make it return uint64_t I guess

Why uint64_t over size_t?

size_t is the wrong type, it differs between 32 bit and 64 bit systmes

15:21

32 bit systems can have files larger than 4 GiB

But they can't load them into memory

15:21

Which is what the function does

15:21

But I get your point for the other functions

streams?

loading a 4 GiB file into memory is also a stupid idea on 64 bit

heinrich5991

loading a 4 GiB file into memory is also a stupid idea on 64 bit

depends

15:22

i like loading a sqlite dbb into memory

15:22

xd

Ryozuki

depends

give me an example in the context of ddnet

ok but u didnt mention ddnet

15:23

u just said 64bit

15:23

i dont know any rn

15:25

@heinrich5991 btw, if a file is in a tmpfs and u read it, is it 2 times in memory? not ddnet context

15:25

i guess a tmpfs is more good at avoiding writes

Ryozuki

ok but u didnt mention ddnet

check discord name

we're accidentally talking about ddnet development

xd

Ryozuki

@heinrich5991 btw, if a file is in a tmpfs and u read it, is it 2 times in memory? not ddnet context

twice in memory. you can mmap it so it might only be there once

in the context of ddnet 32 bit is also stupid

it's fine to play ddnet on old computers, although I'm not sure there are still 32 bit computers around

it's bad, bcs it blocks progress for the silent majority

IMO it's rather bad to stop supporting old hardware for software that doesn't need new hardware

15:49

contributes to waste

old hardware usually uses more energy for the same effective result

15:50

also software would be less bloated without old hardware

I'd bet it's better to use old hardware with more energy requiremenets than produce new hardware

=> less work to do

heinrich5991

I'd bet it's better to use old hardware with more energy requiremenets than produce new hardware

i wouldn't count on that

15:50

old cars from europe are now driven in africa

random article: https://www.networkworld.com/article/2229029/computer-factories-eat-way-more-energy-than-running-the-devices-they-build.html

Computer factories eat way more energy than running the devices the...

Study finds 70% of the energy a typical laptop will consume during its life span is used in manufacturing the computer

15:51

70% sounds like using an old computer is always better than buying a new one

15:52

in terms of energy requirements

15:52

(not even talking about resource usage, where the waste goes)

well this clearly does not include the result (less performance)

15:54

more software bloat

15:54

more maintaince

Jupstar ✪

well this clearly does not include the result (less performance)

are you still talking about energy usage? or is this a different topic?

[ddnet/ddnet] Pull request closed: #7103 Fix some CLocConstStrings being localized twice

heinrich5991

are you still talking about energy usage? or is this a different topic?

yes i am

I'd guess it's mostly the "modern software" that's more wasteful, not backcompat

15:56

old software runs much faster (more energy efficient?) on new hardware AFAICT

and what does that have to do with the age of the software?

15:57

new hardware is usually faster

15:57

that python sucks is crystal clear but vulkan is clearly more energy efficient than OpenGL

15:57

so targeting vulkan makes sense

got a citation on that btw?

15:58

the vulkan being more energy efficient?

it's designed to use less CPU

15:58

benchmarks back that up

can you link one that shows that it uses less power?

https://www.youtube.com/watch?v=rvCD9FaTKCA is a video from ARM good enough for you?

First comparison of Vulkan API vs OpenGL ES API on ARM

sounds good

16:00

what's opengl es vs opengl?

16:00

do you think it performs similarly?

yes

ok

it's a subset of OpenGL 4.3 (OpenGL ES 3.2)

[ddnet/ddnet] Pull request opened: #7109 Remove CLocConstString

This makes it more obvious where text is localized. This class was also broken for localized strings with context, since the member variable m_ContextHash was uninitialized.

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 ar...

16:08

[ddnet/ddnet] Pull request opened: #7110 Remove unused enum literals

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-addresss...

heinrich5991

what's opengl es vs opengl?

opengl es is what used to run on mobiles iirc

16:13

since mobiles are arm too

[ddnet/ddnet] Issue opened: #7111 Sometimes zoom changes quickly when first joining a server

Like it starts zoomed out too far and zooms in very quickly when joining a server after first launching the client. Doesn't happen consistently though. I'm using default zoom settings. Could be related to #7016.

cargo clippy --workspace -- -D warnings -D clippy::nursery -A clippy::collection_is_never_read -A clippy::string_lit_as_bytes -A clippy::option_if_let_else -A clippy::use_self

i cooked this up for smth at work

16:15

easy pr

16:16

ohh

16:17

i found a false positive

16:17

i guess clippy cant detect if a destructor is non const and it suggest to make a function const

Ryozuki

cargo clippy --workspace -- -D warnings -D clippy::nursery -A clippy::collection_is_never_read -A clippy::string_lit_as_bytes -A clippy::option_if_let_else -A clippy::use_self

i cooked this up for smth at work

clippy's words are no longer god's words? ^^

Ryozuki

i guess clippy cant detect if a destructor is non const and it suggest to make a function const

it should be able to, maybe make a bug report

it is, most of the time xd

16:21

unless compiler says no

16:22

justatest

heinrich5991

clippy's words are no longer god's words? ^^

i put allow* cuz i know ppl at work arent madmans like me (edited)

16:22

also the stirng_list_as_bytes comes from a macro

16:22

string*

16:23

(and no we dont run nursery anywhere, its just a thing i like to do)

16:23

i actually found out a lot of possible const fns

Ryozuki

also the stirng_list_as_bytes comes from a macro

report a bug ^^

Ryozuki

i actually found out a lot of possible const fns

this is public API in library, got to be careful there

oh true

16:24

well its unlikely to change

16:25

the consts are mostly "new" style functions

16:27

@heinrich5991 whatu think about https://rust-lang.github.io/rust-clippy/master/index.html#/equatable_if_let

A collection of lints to catch common mistakes and improve your Rust code.

meh

16:35

not worth changing code over

odd i cant make a min example

16:48

for clippy

[ddnet/ddnet] Pull request closed: #7110 Remove unused enum literals

[ddnet/ddnet] Issue closed: #7101 Crash in AntibotOnDirectInput

github-merge-queue[bot]

[ddnet:master] 2 new commits

0a290bc Remove unused enum literals - Robyt3 ddb3c2d Merge pull request #7110 from Robyt3/Serverbrowser-Enum-Cleanup - heinrich5991

Image attachment

17:58

Image attachment

Ryozuki

Click to see attachment 🖼️

the funny thing is, the third happens quite often and mostly that comment then doesn't help xD

Ryozuki

Click to see attachment 🖼️

you are 6th xd

i knewe u would say that

18:03

xd

[ddnet/ddnet] Pull request closed: #7109 Remove CLocConstString

[ddnet/ddnet] Pull request closed: #7108 Update turkish.txt

royowo the tab girl uwu

Replying to @Jupstar ✪ you are 6th xd

github-merge-queue[bot]

[ddnet:master] 2 new commits

bc82a45 Remove CLocConstString - Robyt3 7d6ec23 Merge pull request #7109 from Robyt3/CLocConstString-Removal - def-

I identify as the 1st

Replying to @Ryozuki F4uKl0RXYAEX_Ac.jpg

18:29

Jospsiti is the question answerer

Replying to @Ryozuki F4uKl0RXYAEX_Ac.jpg

chillerdragon: about your pr for login censorship with the streamer mode. maybe just make aliases for specific phrases. so you could type /login MY_PASS in the chat and MY_PASS just gets replaced before sending the msg.

18:34

so ig it would be some kind of chat alias mod

So put your password in a Config and only censor your password?

18:39

As Heinrich says that would encourage password reuse

18:40

And still leak the restore safe words ddnet generates

18:40

And the password your team mate chooses

use tokens instead of passes

??

then make an alias to the the token

What are you talking about

18:40

Im not following

18:41

What token

18:41

And who should do that?

u know how the alias cmd on linux wokrs right?

Yes

18:42

So whats your idea?

18:42

Only replace all passwords the user blacklisted?

no

Then what

you type /login PasswordForFokko the server gets /login youractualpassword

Don’t get it

18:44

Okay let’s ignore that I do not understand it. How would it be better than my approach?

simple aliases so streamers dont leak anything and if you have tokens instead of passwrds then you dont have to bind them to remember

18:44

-> Okay let’s ignore that I do not understand it. How would it be better than my approach? more usecases

You mean text replace

18:45

For specifc keywords

exactly

That’s more variables than aliases on Linux

18:45

But yeah I get it now

18:45

Still would leak more than my pr

how so?

Ddnet sends credentials in chat

18:45

They have to be hidden

18:46

And one can forget to use the Alia’s

what if you have a token and dont want to bind it?

If it censors all parameters to confidential commands you can’t accidentally leak it

18:46

Then just bind it?

18:46

It’s a different feature imo

dont have any keys left to bind /login randomfkingtopken (edited)

Then bind exec calls

18:47

Still would encourage Password reuse

chillerdragon

If it censors all parameters to confidential commands you can’t accidentally leak it

If you type /loginhunter2 or /lgoin hunter2 then you still accidentally leak it though

Just use chillerbot-ux password manager then :p

18:48

No

18:48

Ah with mine?

18:48

Yes

what bout 7LOGIN CHILLERHUNTER

But that’s not related to streaming

18:48

Streaming is about someone watching your screen

18:48

If you publish a pw it’s Public yes

18:49

I don’t understand why we are discussing this

Well, you could make it simpler. Add a key bind to toggle the entire chat input being censored

Having password fields not show letters is standard

18:50

Including your input?

Replying to @Robyt3 Well, you could make it simpler. Add a key bind to toggle the entire cha…

I mean only the inputs, yes

Then you don’t see what command you type

18:50

Still leaks the ddnet response in chat

18:50

Or are you only talking about input for now?

Yeah, the response is a separate issue

18:51

Though, yeah, also not ideal that you can't see if you typed the command correctly

Much uglier ux

18:51

To not see the command

18:51

Caused was more 7login cases

18:51

And is just weird

18:51

And has to be activated

18:51

vs mine being automatic for the full stream

Then server commands should also be communicated to the client like on upstream

No brain farts

So you can select server commands from a menu

Why?

Replying to @Robyt3 Then server commands should also be communicated to the client like on u…

And the command can advice the client that its arguments should be censored

Ok sure would be cool

18:52

But we also hacked the auto complete

chillerdragon: the client somehow has to know which commands should be censored and which should not be. maybe some gamemode has a diffrent command for login or other specific things that should not be seen.

But yea command system would be nice

18:52

But should not block that pr

18:52

Which?

Replying to @Teero chillerdragon: the client somehow has to know which commands should be c…

Would be a lot less hacky with a command system instead of having a hard-coded list of commands that need censoring

login and the ddnet commands is covering pretty much all isn’t it

18:53

Yes

18:53

So?

Teero

chillerdragon: the client somehow has to know which commands should be censored and which should not be. maybe some gamemode has a diffrent command for login or other specific things that should not be seen.

Yeah, the server could tell the client exactly that

18:53

Like the server sends rcon commands to the client it also can send chat commands on 0.7

Auto complete is also hacky and my pr is already useful

[ddnet/ddnet] Pull request closed: #7094 Remove gfx_finish config variable, remove dbg_hitch config variable, improve config variable descriptions, add missing descriptions

Im not Building a cmd system now I’m busy with 0.7

18:55

Yea I get your point and agree it would be nicer. But it shouldn’t be a reason to not help streamers right now at all.

18:55

My code is working and at the same hack level as chat command auto complete. And it covers all mods I saw in the last decade.

18:56

Sure it doesn’t cover edge commands such as /pin or /changepassword and a command system would be better. But this pr is good enough to be merged.

Should be tested with unicode input, text selection and IME in the censored part though

I build a fakin std function for you which also less hacky than the existing rcon password censor. And now you are saying it needs a whole command system to be merged?

xDDD

18:57

imagine having unicode symbols in your pw XDD

Asians

like yea i got that wonky smiley face as a pw

It's okay for now yeah

¯\_(ツ)_/¯

Yea the android player might

18:58

It’s surprisingly secure pw for one letter to use a poop emoji

github-merge-queue[bot]

[ddnet:master] 5 new commits

31a80c9 Remove gfx_finish config variable - Robyt3 1028ced Remove dbg_hitch config variable - Robyt3 943f98d Improve config variable descriptions, add missing descriptions - Robyt3 98b3fe4 Fix typo in internal name of cl_assets_entities config variable - Robyt3 16115c2 Merge pull request #7094 from Robyt3/Variable-Desc-Fixes - Jupeyy

chillerdragon

It’s surprisingly secure pw for one letter to use a poop emoji

true

I like this character: 〠

18:59

"Postal Mark Face"

Robsti pw leaked

〠〠〠

19:04

@Robyt3 this is the pot

19:04

from elden ring

had to zoom 300% to even read the symbol

Search puzzle: find the Postal Mark Face in the font atlas picture from release 17.2

chillerdragon: damn, my fav subscriber

Ikr

19:06

Idk how I missed your channel

19:07

Just saw it on fddrace discord

I just got a whole push notification cuz u subbed

get pingerd

19:09

Faak soon you have more subs than me

Replying to @Anime.pdf I just got a whole push notification cuz u subbed

Would be sweet

Jupstar ✪

had to zoom 300% to even read the symbol

looks like some kinda monk tbh

Image attachment

Teero

looks like some kinda monk tbh

for me it looks like this xD

Image attachment

justatest

u need to zoom in first before making a pic, else it's too bad resolution xd

chillerdragon

It’s surprisingly secure pw for one letter to use a poop emoji

no, it's not

github-merge-queue[bot]

[ddnet:master] 3 new commits

ca306b9 kesim -> kesit & improvements - h-kaan cf3f90f add the modification date - h-kaan 693fec2 Merge pull request #7108 from h-kaan/update-turkish - def-

[ddnet/ddnet] Pull request opened: #7112 Fix some menus tooltips not working anymore

For some tooltips, the associated UI element ID was not being set as hot item, which is required for tooltips. The tooltip for the "Dummy settings" checkbox was only present on the tee settings page but not on the player settings page. Closes #7107.

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 t...

https://lichess.org/p6nJeGxR

Blitz (5+3) casual Chess • Challenge from Ryozuki (1001)

Join the challenge or watch the game here.

20:03

i challenge thee

uff ryo nub

20:08

u drunk? xd

20:09

uff white

20:09

u drunk?

1

lol

20:09

xd

20:09

im bad

20:09

been long i havent played

20:09

but i want to cum back

gg ryo xd

he missed it xD

i think white could have won a few times

ye :(

r u white? xd

I am a ghost!

xd

bro

20:16

how can u not know

20:16

im black

? xD

20:17

nobody talks about u

[white]yo Ryozuki who u [white]I am a ghost! [white]xDD [white]xDD

20:17

ah ok

20:17

xd

20:17

nvm

xDDD

xDD

i cant read your ingame chat btw xd

a xd

Chat room

    [white]yo
    Ryozuki who u
    [white]I am a ghost!
    [white]xDD
    [white]xDD
    [white]whoopsie
    [white]gg
    [white]Good game
    [white]Well played
    Ryozuki I've got to go
    Ryozuki Thank you
    [white]bb

@heinrich5991 wtf is cactus client lmao

[ddnet/ddnet] Issue opened: #7113 #7016 breaks zooming.

Their change (#7016) sets the zoom back to default every tick while you aren't in spectator meaning you wouldn't be able to change the zoom while playing.

Wot cactus

Replying to @ReiTW heinrich5991 wtf is cactus client lmao

ReiTW

@heinrich5991 wtf is cactus client lmao

where did you see "cactus client"?

Iza

where did you see "cactus client"?

was a joke

should i split up the file loader for the assets into each type of asset (allowing you to pick & choose which type of assets to load if that ever became a necessity) or do it all at once

menus_settings_assets not being its own component really is kind of a bummer because i could totally use that stuff

22:42

and it would be nice to not have to add to the already gargantuan class that is cmenus

22:42

what is the limitation making menus_settings_assets a part of cmenus as a whole?

Exported 518 message(s)