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 2025-01-05 00:00 and 2025-01-06 00:00

chillerdragon: there's no way

good morning

06:49

question: If Godot would have a fully developed C++ support (using modern C++ 20 standards) and can also use Rust alongside it through Godot's API, would you guys try running Teeworlds through that Engine?

06:50

idk how much code can it be transfered to Godot, but I think the majority may be able to run the client side as you would expect it

<ChillerDragon> @MilkeeyCat well you get the shape (air or not) of one layer into another sounds like a feature to me

07:34

<ChillerDragon> jupstar does ddnet-rs feel to you as smooth as ddnet? Can you play it normally? I am getting a bit nervous something feels off

Cellegen

question: If Godot would have a fully developed C++ support (using modern C++ 20 standards) and can also use Rust alongside it through Godot's API, would you guys try running Teeworlds through that Engine?

Why would we switch to something less efficient, less stable, less straightforward, just generally less suitable? Style points I guess?

chillerdragon: But what if it's not air

Image attachment

Image attachment

risu

Why would we switch to something less efficient, less stable, less straightforward, just generally less suitable? Style points I guess?

Or is this just for the sake of argument because you're using Godot?

[ddnet-web:master] 1 new commit

106fd60 Davide funded GER1 for the rest of the year - def-

poggers2

7

Thanks @Davide !

ws-client

<ChillerDragon> jupstar does ddnet-rs feel to you as smooth as ddnet? Can you play it normally? I am getting a bit nervous something feels off

Mh? I'd say it's super smooth. The mouse movement isn't as fps dependent either. I can even play on 50 fps

09:01

The cursor is properly 360°

09:01

But do a video of both versions and show me what you mean

Chiller antiping is off by default. i also noticed the hookline is bit buggy with antiping on, lemme fix that

[ddnet/ddnet-rs] Pull request opened: #14 Move hookline adjustment where cursor is sent.

Fixes hookline with antiping on

[ddnet-web:master] 1 new commit

abdb21b Remove DDNet Turkey - def-

10:10

[ddnet/ddnet-rs] Pull request closed: #14 Move hookline adjustment where cursor is set.

10:10

[ddnet-rs:main] 2 new commits

6648fca Move hookline adjustment where cursor is set. - Jupeyy 3de6821 Merge pull request #14 from Jupeyy/pr_fix_hookline - Jupeyy

chillerdragon do you use x11 or xwayland?

10:31

i tested with ms 1 now. sdl reports both correct for xwayland and wayland but winit does smth weird with x11 xD

10:31

so maybe you are right

10:31

try native wayland for testing

chiller i reported it: https://github.com/rust-windowing/winit/issues/4077 xdd

10:57

lets see what they say

<ChillerDragon> i think i was already on xorg when i tested ddnet-rs i just switched from wayland because the ddnet client cant capture the mouse on wayland -.-

11:49

<ChillerDragon> @MilkeeyCat yes updating the brush texture would be nice

11:50

<ChillerDragon> jupstar i do not think its anything a video can catch it just feels a bit different i can retest when im back home

11:51

<ChillerDragon> can we talk about that there is a variable CCharacter::m_Pos and a function CCharacter::SetPosition which does not set it?

11:52

<ChillerDragon> https://github.com/ddnet/ddnet/blob/43524b9b20ef8a3f4a213e92b6a193236a1aba39/src/game/server/entities/character.cpp#L2445-L2448

ddnet/src/game/server/entities/character.cpp at 43524b9b20ef8a3f4a2...

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

11:52

<ChillerDragon> CCharacter is an entity so it inherits https://github.com/ddnet/ddnet/blob/43524b9b20ef8a3f4a213e92b6a193236a1aba39/src/game/server/entity.h#L51

ddnet/src/game/server/entity.h at 43524b9b20ef8a3f4a213e92b6a193236...

DDraceNetwork, a free cooperative platformer game. Contribute to ddnet/ddnet development by creating an account on GitHub.

Looks like the character m_Pos is updated every tick based on the core position

[ddnet/ddnet] Issue opened: #9481 make quad handling better: features / fixes

I will outline my frustrations: 1. The handles are too small, especially on large screens. It feels like they're a set pixel length when they should scale with screen resolution 2. You can't tell what quad each red corner corresponds to or what corners each green handle corresponds to. The red corners should look like right angles instead of squares by default - each red corner should essentially point to its two neighbor corners so if someone un-rectangles a quad you can still tell where...

PepeCross

Image attachment

Learath2

PepeCross

https://discord.com/channels/252358080522747904/293493549758939136/1322907880428408925 kekw

kekw

anyone knows whether the traffic in ddnet.org/status is displayed in bits or bytes (edited)

confused

1

12:46

Image attachment

12:46

this part

Afaik, it's straight from /proc/net/dev

12:48

(which is in bytes)

got it, thanks lerato

@eidderFsT any luck with the http register? did you get to test it?

not really

<squeeZY> yeah

13:16

<squeeZY> learath bir babadır bizim için

chillerdragon: when u home try to half the mouse sens as test pls

Jupstar ✪

https://youtu.be/BDYdIbmIN0s

i approve, watched 25 minutes in bed and just fell asleep

Learath2

PepeCross

evil ddnet soon

chillerdragon: https://github.com/rust-windowing/winit/issues/3773 i have to ping you soz hehe So nice, wayland is buggy bcs of the other bug and x11 is buggy bcs of this

14:31

and windows cant have exclusive fullscreen: https://github.com/rust-windowing/winit/issues/3124

14:31

Tbf winit 100% has better error handling than SDL2

14:31

but man, why is window management so fcking annoying in 2024 xDD

[ddnet/ddnet-rs] Pull request opened: #15 Update winit

Remove workaround

[ddnet/ddnet-rs] Pull request closed: #15 Update winit

15:10

[ddnet-rs:main] 2 new commits

8f61ad3 Update winit - Jupeyy 7c85a2f Merge pull request #15 from Jupeyy/pr_update_winit - Jupeyy

Image attachment

15:32

filthy casuals

average gerdoe commit

kek

1

I think set position still should set both to avoid unexpected values in the first tick

Replying to @Robyt3 Looks like the character `m_Pos` is updated every tick based on the core…

16:07

Yea that should really be solved problem by now

Replying to @Jupstar ✪ but man, why is window management so fcking annoying in 2024 xDD

risu

Why would we switch to something less efficient, less stable, less straightforward, just generally less suitable? Style points I guess?

I don't think these claims are true tho, Godot is efficient enough for a 2D game like Teeworlds to run perfectly on older computers as well, since it can use OpenGL as an Engine Backend and the pipeline has also been improved by a lot since previous versions. It is actually more straight forward, than doing all of the work on Visual Studio, since the IDE has a visual interface to work on and you can test those changes in a more creative manner! I think Godot is a perfect way to let users interact with the source code better. It may be not suitable, since it's not built for Godot, but that doesn't mean it cannot be remade to be suitable. The whole reason I mentioned this, is because you guys lack ways to improve the dev experience for common folks, like me. You may say, that they could just make their own game if they don't like the engine, which I do it myself. We'll see how many people off of this playerbase want this kind of change.

Cellegen

I don't think these claims are true tho, Godot is efficient enough for a 2D game like Teeworlds to run perfectly on older computers as well, since it can use OpenGL as an Engine Backend and the pipeline has also been improved by a lot since previous versions. It is actually more straight forward, than doing all of the work on Visual Studio, since the IDE has a visual interface to work on and you can test those changes in a more creative manner! I think Godot is a perfect way to let users interact with the source code better. It may be not suitable, since it's not built for Godot, but that doesn't mean it cannot be remade to be suitable. The whole reason I mentioned this, is because you guys lack ways to improve the dev experience for common folks, like me. You may say, that they could just make their own game if they don't like the engine, which I do it myself. We'll see how many people off of this playerbase want this kind of change.

I mean straightforward on a technical level, not ease of use. The engine is doing a number of extra steps not needed for Teeworlds because it needs to be general.

16:28

I'm sure people wouldn't be getting 15k fps anymore

So you think Godot would shit itself?

Nah it would run fine, but it wouldn't be exceptionally good like Teeworlds is currently

risu

I'm sure people wouldn't be getting 15k fps anymore

Why would you need 15k FPS in the first place? You need a decent amount of FPS for rendering, the rest is handled differently

using a different ide is a pain for alot of people because all the tools you like are gone or different, especially if you are a emacs/vim bro godot is a general purpose game engine, ddnet/teeworlds has made its own. i cant say anything truthfully about speed but but it would probably be worse. the size would be much worse as well. (edited)

risu

Nah it would run fine, but it wouldn't be exceptionally good like Teeworlds is currently

If the sacrifice is performance, for the sake of better programming on it, I would take it every day

Cellegen

Why would you need 15k FPS in the first place? You need a decent amount of FPS for rendering, the rest is handled differently

Well for some people going down from 300 to 60 or something along those lines would be a significant downgrade

What is the benefit of moving from an engine purpose built for this game to one that isn't? The only thing I can imagine is attracting godot developers to bolster our dev numbers, but I don't think any would show up just because we are on godot

Solly

using a different ide is a pain for alot of people because all the tools you like are gone or different, especially if you are a emacs/vim bro godot is a general purpose game engine, ddnet/teeworlds has made its own. i cant say anything truthfully about speed but but it would probably be worse. the size would be much worse as well. (edited)

it would also be a lot of work to completley port tw to godot since ya know its a big game. starting from scratch would take less time. making teeworlds run inside a godot environment might be possible (idk about the rust bits but those can be made into cpp) would not provide much of the benefit

Also, gdscript is the best supported language in godot, but none of us write gdscript

risu

Well for some people going down from 300 to 60 or something along those lines would be a significant downgrade

even if (with vsync or otherwise) you limit yourself to 60fps its the difference between your computr blowing up, and being able to have discord open with vc and chrome open streaming music

Well I write a little bit but I wouldn't say I'm great

risu

I'm sure people wouldn't be getting 15k fps anymore

lol

16:31

11/10 best argument

ddnet-godot when?

16:34

also ddnet-ue5, ddnet-unity

risu

Well for some people going down from 300 to 60 or something along those lines would be a significant downgrade

that just means it needs optimizing, nothing new... for old developers

Solly

it would also be a lot of work to completley port tw to godot since ya know its a big game. starting from scratch would take less time. making teeworlds run inside a godot environment might be possible (idk about the rust bits but those can be made into cpp) would not provide much of the benefit

How good is Rust's C++ interop? It's all native modules after all

risu

How good is Rust's C++ interop? It's all native modules after all

to me it looks like an ugly mess and i label it as "it works".tldr dont ask me :)

16:37

you should be able to easily convert rust -> cpp pretty easily (minus crates)

Cellegen

that just means it needs optimizing, nothing new... for old developers

But if we have to start modifying Godot ourselves and maintaining our own fork or at least low level add-ons then I doubt the effort is worth it

Solly

you should be able to easily convert rust -> cpp pretty easily (minus crates)

True! Simply rewrite in not Rust!

an effort to make a bigger game, than what we have? From not getting limits on an old ass engine?

16:39

You guys just cannot take a risk don't you?

@Cellegen just do your teegalaxy. if it's done i bet many ppl will like it

16:40

because of the more advanced features godot can do

the limits that there are is more from lazy programming and can be refactored to remove them for less work the risk is 100s of hours of work, and no one wants to do that even if it is worth it

if godot rewrite bad why is a rust rewrite a thing? (edited)

Cellegen

an effort to make a bigger game, than what we have? From not getting limits on an old ass engine?

How much are you thinking of adding? If we keep it a platformer with funny circular player models then I feel the current engine is plenty good

Jupstar ✪

@Cellegen just do your teegalaxy. if it's done i bet many ppl will like it

I only hope I can finish the basics, so people have some good ideas on how to improve / build on it

16:41

main reason why it takes so long, as is, alongside real life issues

hey, could any devs please look at and consider adding #deleted-channel and #deleted-channel would be really helpful to mappers and I don't see why they would break anything

Unless you're seeking modern graphics effects instead of flat 2D

even for 2d godot is defs better than the ddnet engine

16:41

if your goal is to have more complex effects at least

risu

How much are you thinking of adding? If we keep it a platformer with funny circular player models then I feel the current engine is plenty good

How much? I cannot list all in my head, but I think https://discord.com/channels/252358080522747904/1291415577142755399 has your answer (TeeGalaxy post)

https://cdn.discordapp.com/attachments/957377564387733595/1324640139707682900/trails.png?ex=677b85c3&is=677a3443&hm=547fbe9c85f53474bcc4a5fa56475e93988577cc1232ad384e5ff1cd820b4d0b& hows this for effects?

Embedded image

risu

Unless you're seeking modern graphics effects instead of flat 2D

I'm seeking a better playerbase

16:43

plain and simple

16:43

not in a way of "better", but "more"

16:43

more people = good

Cellegen

I'm seeking a better playerbase

Better as in more mature? More dedicated (doubt you'll find that)? More creative?

Cellegen

more people = good

Ok I agree with this but I think it's a marketing thing because the game itself is peak

Better as in more active, less troublesome and more fun to play with them

16:45

Ask people how they enjoy the game, and they'll say less positive things, than negative things

16:46

I want to flip it around, so we'll have more positive things and near non-existent negative things

risu

Ok I agree with this but I think it's a marketing thing because the game itself is peak

if you say a few thousand (from russian and eastern regions) is peak, then I hope you don't find my work to be too much of a hassle when I say, that I target at least 10k

Free game that runs on potato will attract kids, can't avoid that, so the troublesome and socially inactive portion is just a natural consequence

risu

Free game that runs on potato will attract kids, can't avoid that, so the troublesome and socially inactive portion is just a natural consequence

kids?

Cellegen

if you say a few thousand (from russian and eastern regions) is peak, then I hope you don't find my work to be too much of a hassle when I say, that I target at least 10k

No I mean that the gameplay is already great

Are you guys attracting kids with the game?

Cellegen

kids?

Well russian kids are a big part of the problem

Cellegen

Are you guys attracting kids with the game?

arent games made for kids?

Russians. You know I never got a surprise bigger, than having russians and eastern people suddenly know the game

Koll Potato

arent games made for kids?

supposedly, until adults are playing them too

16:49

the majority of the veteran part of the community is already in adulthood

16:50

including me, who was around 9-10 years old when I started playing this game

16:50

I'm 23 currently

16:50

If this doesn't tell you, that the game is stuck in an endless loop of nostalgia, then idk

16:53

Nowadays, people can buy the cheapest, most acceptable computer, for around 200-250€. Not a PC, maybe a laptop or a notebook, but that can also run Godot the same way. Problem with Godot is mostly draw calls, physics engine and cycles in code.

16:54

It can reach a maximum of 1000-3000 fps for 2D, which is more than enough. You can build around it, and whenever it drops significantly, that's where you need to optimize

Cellegen

You guys just cannot take a risk don't you?

Risks are only worth taking if the upsides justify it This would be investing significant dev effort for almost no gain

Well this is what a risk is really. You never know if it turns out to be a net positive or net negative. It's a gamble, willing to take it

I dunno what the problem is. If your base game has a good concept and ppl like it they will try it out

17:04

I don't see it as risk

17:05

The risk would be if all developers at once would stop the current version and only do the new version

Jupstar ✪

I don't see it as risk

Not what you made, I'm talking about him wanting a godot rewrite

Learath2

Not what you made, I'm talking about him wanting a godot rewrite

Yeah, I mean, he isn't really contributing to ddnet. If he makes a really good godot base, that has potential. Why not

Cellegen

If this doesn't tell you, that the game is stuck in an endless loop of nostalgia, then idk

It's not nostalgia, it just caters to an audience of players with patience who like tons of practice. People's tastes have shifted, but there are still plenty of younger people who this sort of game appeals to.

He made it sound like there even is a risk

I'm a bit older than you and I just recently started playing, definitely not nostalgia for me

Jupstar ✪

He made it sound like there even is a risk

Well if he is just doing it on his own then it really is none of our concern, the only risk is he wastes his time. I'm just saying it's not an investment that's worth it for us

Jupstar ✪

He made it sound like there even is a risk

Only risk is the opportunity cost for the developers

risu

It's not nostalgia, it just caters to an audience of players with patience who like tons of practice. People's tastes have shifted, but there are still plenty of younger people who this sort of game appeals to.

naturally, I don't doubt that it is the case

Learath2

Well if he is just doing it on his own then it really is none of our concern, the only risk is he wastes his time. I'm just saying it's not an investment that's worth it for us

I asked it to see what you guys think of it, not that I order you all to do that xd

17:09

a conversation is not all about getting you to a different mindset

17:10

but rather let me understand how you guys think generally

risu

I'm a bit older than you and I just recently started playing, definitely not nostalgia for me

u can be my grandpa poggers2

poggers2

ddnet code is so bad that there are currently 3 known rewrites going on right now

XD

An anecdote: after seeing the game's website and playing for a bit, a younger friend of mine was genuinely surprised that this game still gets updates frequently

Hecta

ddnet code is so bad that there are currently 3 known rewrites going on right now

no matter how good or bad a codebase is, it will at least have a rust rewrite

Hecta

ddnet code is so bad that there are currently 3 known rewrites going on right now

wait are there really that many

17:11

ddnet-rs is not the only one?

Cellegen

I asked it to see what you guys think of it, not that I order you all to do that xd

And I told you what I personally think, that's all

I mean, a CPP to Rust is probably a lot easier for them, than to GDScript or native langauge compatibility on Godot

Learath2

And I told you what I personally think, that's all

Good, nice to see we can talk about this stuff ^^

DDNet is an old game and it feels like an old game. I think that pushes away some people

it also looks like an old game as well, design haven't changed that much from the vanilla, except from the main menu

That's why I'm saying it's a marketing thing

17:15

Some visual redesign, better support for UI customization on the side, less reliance on the website, and redesign that too

does not look that hard

Accounts, accounts would be huge

Some parts of the website are really outdated, like https://ddnet.org/client/ doesn't describe any recent features

like, im literally right here spitting out new ideas but no one cares because it breaks backwards compatability

i care

like. no one uses any old version anymore and we are literally on our knees to maintain a version only 10 people or so use

1

real

17:27

its not 2007 anymore, updating the client takes 3 seconds, i dont get it why old versions are supported

because people are stubborn and hate it when something changes

but literally nothing changes, just because it needs to be backwards compatible kek

kek

the ui changes and thats enough for them

remove uuids from the protocol

or, OMG THE TEES ARE SITTING or, OMG THEY REMOVED ONE BUTTON (that was useless to begin with)

save bandwidth

Hecta

or, OMG THE TEES ARE SITTING or, OMG THEY REMOVED ONE BUTTON (that was useless to begin with)

if thats real then its probably like 2 people

17:31

making the life harder for 10k ppl playing atm (edited)

exactly!

in many cases breaking compatability with 0.6 also means breaking compatibility with every other ddnet version which alot of people use eg if you were to increase the switch limit by making it a big number, oh no now you have to be on the latest version for this to work or you get corrupted maps

why do they use old versions

kebs

why do they use old versions

dont ask me

17:38

well do ask me, before i compiled it myself i didnt use steam so i only updated once every so often

17:38

and i bet there are people who havent updated in ages

Solly

dont ask me

who do i ask then

17:38

these are imaginary people

@everyone what version do you use

17:39

doesnt ddnet say what version people are on when they join

fng does

17:39

everyone was on latest whenever i played

19.0?

The noob experience is also quite ass imo

technically new features that do not change physics can be brought to the game without compatibility issues

17:41

and physics do not change that often xd

the old clients would have to atleast not freak out when new physics happen

It's Tutorial (great btw), Linear, Multeasy, and the list of novice maps does not help noobs pick suitable maps

1

ddnet still supports vanilla clients

17:41

imagine

17:41

freeze is treated like ninja

Like, recommended novices are so old and crusty

1

ninja has the sword, freeze doesnt

The tutorial also only covers single player mechanics

Koll Potato

ddnet still supports vanilla clients

supports, but its laggy and kinda unplayable

17:42

fall into freeze and hook starts glitching etc

kebs

supports, but its laggy and kinda unplayable

yes, but why should it still exist today

risu

It's Tutorial (great btw), Linear, Multeasy, and the list of novice maps does not help noobs pick suitable maps

The Tutorial map is great but it's already too long IMO, we should have multiple smaller Tutorial maps teaching individual concepts

Robyt3

The Tutorial map is great but it's already too long IMO, we should have multiple smaller Tutorial maps teaching individual concepts

Yep yep yep

17:43

A tutorial taking over ten minutes your first go is unusual

idk it was kinda fun

17:44

took me like 12 or 15 minutes

Koll Potato

idk it was kinda fun

It is, but the number of people I see getting stuck on the grenade part justatest

justatest

grenade is unintuitive for new players

17:45

and why the tutorial only has grenade

17:45

no laser no shotgun

Yeah I only knew how to do it because I've played Quake

17:49

Rocket jump even working is a quirk of how games tend to add velocities in a physically inaccurate way

if we had ✨ accounts ✨ we could dish out achievements for doing slightly shorter than tutorial but more comprehensive tutorials for each mechanic and that would motivate atleast me (i like achievements)

17:50

achievements could also be purely client side / steam but ✨ accounts ✨are cool

or just smaller tutorial maps with a popup to continue to next map when you finish

17:51

and when you finish all of them it should teach you how to use server browser

(edited)

when i was novice (still am kek) and i found someone decent to play with it was a real gamble to whether or not they could follow me xd

there should be a feature to follow someone

2

Also a way for people to return to the tutorials if they stop playing them halfway through

(or autosaves for all solo maps)

1

In-client notifications for when a friend joins a server

17:54

osu!lazer has a whole notifications sidebar

17:55

Better difficulty ratings would be nice, the discrepancy between new and old maps is huge

1

17:56

Some systematic way of updating them...

imagine ddnet!lazer

17:56

osu!lazer was a rewrite but osu was previously made by only 1 person so it worked

Solly

imagine ddnet!lazer

I keep thinking about osu! in general because you can draw so many parallels between it and DDNet

1

18:00

Both are community based games completely funded by donations where you grind skill for hours on end

osu is more mainstream and easier to get into tho

Yeah and that's kinda what I would want to happen to DDNet

we need better ddnet advertizing

I could see it happening

Hecta

we need better ddnet advertizing

I've said this like thrice now :P

oh then i agree

Hecta

we need better ddnet advertizing

I think we should address some more of the issues before we attract many more users

18:11

We scale well, but it's all extremely manual and when we do scale we end up on servers with no ddos protection. If this game gets too popular too quick we'll have a lot more ddos issues and ddnet is simply unplayable with ddos

18:12

Ideally steam datagram relay support and a some other relay solution for standalone clients should come before any massive increase in playercount.

18:12

We only get one chance to make a good impression

Yeah and advertising is secondary to making the game itself feel fresh and fluid

18:12

We have modern UX standards to live up to

ddnet logo redesign when

osu! gets away easy-ish, they just do everything over http and partially IRC

its too big imo

I did do a couple UI reworks on my free time a couple months back, but I always felt like by modernizing I lost that unique teeish feel

Though they certainly pay Cloudflare more than DDNet's entire budget

Learath2

I did do a couple UI reworks on my free time a couple months back, but I always felt like by modernizing I lost that unique teeish feel

are prs on improving the website design welcome?

I think some work can be done there to help new players ease into the game. Show them how to use the serverbrowser and vote for maps and stuff

Koll Potato

are prs on improving the website design welcome?

Sure, but a nasty surprise awaits you in the website source. It's quite unlike anything a modern web dev is used to

Learath2

I did do a couple UI reworks on my free time a couple months back, but I always felt like by modernizing I lost that unique teeish feel

Now this is a statement dripping with a concerning amount of nostalgia

Learath2

Sure, but a nasty surprise awaits you in the website source. It's quite unlike anything a modern web dev is used to

well its not like it is html 4 with jquery right

justatest

1

18:17

nothing against jquery

18:17

jquery is fine

Almost everything on that website is statically generated

I would prefer a comprehensive customization system with the old design as a preset, since I too am scared of angering the old timers

Learath2

Almost everything on that website is statically generated

does loading the player page take so long because of querying finished maps? (edited)

But I know UI is pain and complex UI is prone to becoming a nightmare

Koll Potato

does loading the player page take so long because of querying finished maps? (edited)

Player pages aren't that bad but some map pages just kill everything

We should not write another UI framework, but that's also kinda what we'd want for a nice user customizable system

18:22

osu! players love their waifu skins but I'd like to go further than just retexturing the same layout

Honestly I'm not the biggest fan of customization for this. It's a lot of work to make things like this customizable and I don't see much reward in the entire ui being customizable

18:23

I see HUD customization being very valuable

Some people love it

Backgrounds are also nice and very cheap devtime and maintainance wise

DDNet already suits the kind of obsessive person who spends their entire life in the client

18:24

I'm telling you people would love their waifu skins :p

We should also look into making a UI tutorial of some sort and make the tutorial map integrate with the client somehow

On the topic of scalability, I was terrified when I learned that the game loads hundreds of skins into VRAM on load

18:26

And submission through Discord feels anything but elegant

risu

On the topic of scalability, I was terrified when I learned that the game loads hundreds of skins into VRAM on load

You need to be careful with this. Not loading them at start is a good idea but you need to load them on a different thread at runtime so as not to cause lag

Learath2

We should also look into making a UI tutorial of some sort and make the tutorial map integrate with the client somehow

yeah, we should have a UI that guides the user through multiple different Tutorial maps and shows the completion state; or just a UI for all DDNet maps in general

Robyt3

yeah, we should have a UI that guides the user through multiple different Tutorial maps and shows the completion state; or just a UI for all DDNet maps in general

I really really want to implement a map vote thing with proper filtering and map thumbnails

Learath2

You need to be careful with this. Not loading them at start is a good idea but you need to load them on a different thread at runtime so as not to cause lag

Yeah you absolutely want to pre-allocate the vram for min(number of installed skins, max players on server) skins

I've however been debating whether I want to make server-pushed-ui first and build it on top of that

Learath2

I really really want to implement a map vote thing with proper filtering and map thumbnails

Thumbnails poggers2

poggers2

risu

On the topic of scalability, I was terrified when I learned that the game loads hundreds of skins into VRAM on load

It's on my list to dynamically unload downloaded skins again so the memory usage doesn't keep increasing without bound

I at least appreciate that the discord is being archived

18:33

I think discord as documentation/forums is an awful idea

It is but the forum is a dead concept to youngsters. They don't like to register for different things

I hate that I have to use discord in general. Feels almost like Facebook back in the day when you couldn't do things without an account because everything was there

18:36

Luckily I am young enough to never have needed an account with them

Learath2

It is but the forum is a dead concept to youngsters. They don't like to register for different things

Well if we happened to have DDNet accounts... greenthing

greenthing

Learath2

It is but the forum is a dead concept to youngsters. They don't like to register for different things

lets have ddnet forums but put subway surfers on the side

Learath2

Almost everything on that website is statically generated

holy damn

Image attachment

18:38

over 50 folders

I had a wacky idea a while back... Different hit sounds for how you hit your hammer

18:40

Like make it sound different if the other tee was not colliding with you

18:40

Unfair advantage go brrr

18:41

Mainly for teaching your brain how to play, not for execution

@Learath2 where can i find the source of the player page?

risu

Like make it sound different if the other tee was not colliding with you

you mean like a distance thing? its different for if a tee is closer to you or not?

Yeah, and a distinct difference between in your hitbox and outside, so you can easily notice even grazing hits

f3 menu like in minecraft

Koll Potato

f3 menu like in minecraft

Koll Potato

f3 menu like in minecraft

Ctrl+D or do you also have to press Shift?

risu

Ctrl+D or do you also have to press Shift?

its bad

18:44

does not show anything useful

18:44

position and velocity hud can be part of it

For coordinates and velocity and fps and angle you already have dedicated displays

well yeah

18:45

it can be a single bind

You can bind them to a key if you find them distracting

which toggles all of that

Writing binds is piss easy but intimidating to people who don't use text based interfaces in general

18:47

And managing complex ones requires file management skills and terrifyingly enough those are no longer common

a scripting api would have fixed that but

18:47

hax

Koll Potato

@Learath2 where can i find the source of the player page?

Hm, it's probably in ddnet-scripts rather than ddnet-web

Learath2

Hm, it's probably in ddnet-scripts rather than ddnet-web

such a mess xD

Sadly a menu good enough to produce most common advanced binds would be so complex that nobody would learn that either

18:50

Maybe a feature for displaying custom icons could be cute for replacing echo Deepfly ON and the like

18:50

but ugh bloat

i had the same idea

18:51

but its not bloat

Koll Potato

such a mess xD

Any changes I make to it takes many minutes of investigation

u can take down whole ddnet if u dont put a semicolon lmao

Koll Potato

but its not bloat

Yeah the lack of this sort of thing is what makes the game feel old

solution: rewrite it in rust

18:53

no wonder player page takes so long to load

18:53

its written in python

It's endearing and pleasant for the technical crowd but weird for everyone else

python 2

justatest

1

18:53

xD

Koll Potato

no wonder player page takes so long to load

It has nothing to do with that. It's that queries have become very slow

It's fiiiine

Learath2

It has nothing to do with that. It's that queries have become very slow

alright alright

Yeah the DB scares me a bit

There are millions of ranks and the db structure is absolutely godawful for the queries we are doing

why not use a template engine?

Image attachment

I've never had to scale up you know :)

I wish we had a db wizard around that can help us a little

Learath2

I wish we had a db wizard around that can help us a little

we all want a db wizard

18:55

dbs are hard

Koll Potato

why not use a template engine?

Probably just how deen felt like doing it. I would probably do it in go with go templates if it was me doing it now

Time to download the entire db and use it as learning material hehe

Learath2

Probably just how deen felt like doing it. I would probably do it in go with go templates if it was me doing it now

so it is going to stay like that forever? (edited)

An optimization we pinpointed was to start keeping another table with only top ranks that is kept up to date with triggers. That should improve query times immensely. I just didn't have time to do it yet

Koll Potato

so it is going to stay like that forever? (edited)

Not really. But you'd need to put significant effort to re do all of it

risu

Time to download the entire db and use it as learning material hehe

ddnet ai :(

Learath2

Not really. But you'd need to put significant effort to re do all of it

yeah i know

Hecta

ddnet ai :(

Nah I'm a natural intelligence

Learath2

An optimization we pinpointed was to start keeping another table with only top ranks that is kept up to date with triggers. That should improve query times immensely. I just didn't have time to do it yet

and perhaps use redis for that db

18:57

nvm

18:58

now u will have to manage 2 dbs

No need for any other db imo

Koll Potato

python 2

based

ok unironically if someone were to train an ai off of ddnet skins id be fine with it. not like im getting anything from my skins anyways

I also wanted to move to postgres sometime. I'm more familiar with postgres

I meant to get a big blob of representative data that I can use for testing to see if I break things with changes

We have terabytes of teehistorian runs that might lead to veeery interesting ai projects in the future

18:59

It's raw input from real humans

Ooh, anti-cheat material

19:00

Don't release it to the bad guys

the only problem is money

19:00

where will u train the ai

Koll Potato

where will u train the ai

This is not NLP or even image generation, it's probably way easier to get useful output with a reasonable amount of training

Money is something we can figure out np. I have a good friend with like a dozen 4090s I can probably even use for free

1

19:03

I've just never been much of an AI guy, so never really had any projects

The data is probably most comparable to audio? considering all the mouse aim with a very human frequency range

Learath2

I've just never been much of an AI guy, so never really had any projects

I think it's really fun as long as you don't start going for massive scale before elegant solutions

Koll Potato

the only problem is money

~~monetize ddnet~~

risu

The data is probably most comparable to audio? considering all the mouse aim with a very human frequency range

My next ddnet project is fourier analysis of teehistorian data to see if I can get some antibot features out of that

Learath2

My next ddnet project is fourier analysis of teehistorian data to see if I can get some antibot features out of that

Now this is the right kind of thinking brownbear

brownbear

I have a feeling the input data from a bot looks pretty different to humans in frequency domain

19:09

I might first write a client that displays live windowed fft of different values, to see if my hypothesis is right before writing code to parse the teehistorian files

[ddnet/ddnet] Pull request opened: #9482 Support reconnecting dummy by redirecting to same port

Previously, the redirect message NETMSG_REDIRECT would always cause the main connection to be redirected and the dummy to be disconnected no matter on which connection it is received. Now, if the redirect message is received on the dummy connection with the destination port being idential to the current port then the dummy only will be reconnected. Redirecting the dummy connection to a different port is not allowed, as the dummy should not be connected to a different server than the main, a...

Learath2

It is but the forum is a dead concept to youngsters. They don't like to register for different things

forum crawling is a thing ive done but rarely. its not any more fun than searching thru discord. often theres an attempt to organize information is an almost wikilike fassion (sprawling faqs) but i dont think its a thing which should be brought back

[ddnet/ddnet] Pull request opened: #9483 Update spanish.txt for 18.9

Checklist

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

Is there toggle on/off for new "zoom copy" when you are spectating someone? its quite annoying to change zoom all the time i start to spec someone and i cant find the cmd. DDNet 18.9 (edited)

@eidderFsT

Learath2

We have terabytes of teehistorian runs that might lead to veeery interesting ai projects in the future

is that compliant with RGPD ?

Chairn

is that compliant with RGPD ?

I don't think it contains any PII

but you did not get consent from users

Chairn

but you did not get consent from users

Do you even need consent to collect non-PII?

20:50

PII is any information about an individual maintained by an agency, including (1) any information that can be used to distinguish or trace an individual‘s identity, such as name, social security number, date and place of birth, mother‘s maiden name, or biometric records; and (2) any other information that is linked or linkable to an individual, such as medical, educational, financial, and employment information.

hmm, definitely not an rgpd expert, i thought consent was always mandatory

Chairn

hmm, definitely not an rgpd expert, i thought consent was always mandatory

Whenever personal data is involved, yes

weirdly enough, im pretty sure you can recognize people through their keystroke, so that could still be considered personal data

Honestly it is all very complicated. We probably would need a lawyer to parse through it and he would probably advise to take the conservative approach and just request consent even if not needed

Chairn

weirdly enough, im pretty sure you can recognize people through their keystroke, so that could still be considered personal data

But you can't really connect that to their real person without a database that does contain PII. Hm

don't you store the pseudonym and maybe the ip as well ?

Chairn

don't you store the pseudonym and maybe the ip as well ?

Pseudonyms yes we store it in records. Ip we store for 48h and delete. Those probably do constitute personal data

20:57

We probably should have a discussion sometime about going through all the data we have and determining if we need to add a consent prompt somewhere

20:58

But we don't even have accounts to store that consent anywhere

On disk is good enough

Is it really? What if they sue us after wiping their disk saying they never consented?

Cookie banners store your consent in a cookie

I guess you can show that you can't get past the prompt without consenting?

21:00

Idk, all very complicated

Yeah

This is very anti small-business/community-project

21:01

With what money are we supposed to hire a lawyer to help us navigate this?

we're not supposed being sued

If keystroke data can be considered PII it would be very hard to fulfill data deletion requests

Learath2

I guess you can show that you can't get past the prompt without consenting?

A forced consent message before loading the list should do the trick, if you reject the ToS the client will close itself, after your first consent the preference will be saved on disk, the fact that the consent is mandatory (code side) is already legally enough to certify that a player accepted the ToS to play (edited)

We also need some sort of privacy policy anyway if ever want to release on the Google Play Store

It can be done at the same way I guess

i cant wait for half the current fanbase to just quit LOL

Oh no my keystrokes

Learath2

I have a feeling the input data from a bot looks pretty different to humans in frequency domain

oh that is an interesting idea :D

22:36

Can you ping me when you have initial results? would love to see that

[ddnet/ddnet] Pull request closed: #9482 Support reconnecting dummy by redirecting to same port

22:37

github-merge-queue[bot]

[ddnet:master] 2 new commits

2cba565 Support reconnecting dummy by redirecting to same port - Robyt3 5ba3b1d Merge pull request #9482 from Robyt3/Client-Dummy-Redirect-Reconnect - heinrich5991

22:37

[ddnet/ddnet] Issue closed: #9480 TOKEN ignored in the redirect packet

22:40

[ddnet/ddnet-rs] Pull request opened: #16 Properly encode path like urls.

Merges the logic from the http server and the download components.

.''.

22:42

.''. '' :/: .

22:42

:/: . .:./ : /\ : .'.:.'.

22:42

.''.: /\ : (/ ':'* /\ * : '..'. -=:o⁠:=-

22:42

В :/:'.:::.В /)*''* .|.*В '.'/.'(/'.':'.'

22:42

: /\ : ::::: '/ | | -= o =- /)\ ' *

22:42

'..' ':::' * /\ * |'| .'/.'. '._

22:42

__.. | | : |. |' .---"|

22:42

* .-' '-. | | .--'| || | | |

22:42

.-'| .| | || '-_ | | | || |

22:42

|' | |. | | | | | |

22:42

| '-' ' "" '-' '-.' '` |_

22:42

ɪʀᴄ.ꜱᴜᴘᴇʀɴᴇᴛꜱ.ᴏʀɢ #ꜱᴜᴘᴇʀʙᴏᴡʟ

22:42

"i⁠s⁠n't it be⁠auti⁠ful?"

22:42

faatv fstd scrumplex_ deen ws-client bridge ChillerDragon jtbx b3z ochristi minus Learath2 Q heinrich5991 Tobii eeeee

[ddnet/ddnet-rs] Pull request opened: #17 Fix map vote index in lists

i am still waiting for my merge queue :c

lmfao

[ddnet/ddnet-rs] Pull request closed: #16 Properly encode path like urls.

23:05

[ddnet-rs:main] 2 new commits

cffef27 Properly encode path like urls. - Jupeyy e1a71f9 Merge pull request #16 from Jupeyy/pr_path_to_url - Jupeyy

23:05

[ddnet/ddnet-rs] Pull request closed: #17 Fix map vote index in lists

23:05

[ddnet-rs:main] 2 new commits

04448b3 Fix map vote index in lists - Jupeyy 78d77d4 Merge pull request #17 from Jupeyy/pr_fix_map_vote_index - Jupeyy

wait sry what is ddnet-rs justatest

justatest

23:06

a different ddnet rust project

23:06

or renamed & moved

it's ddpg

[ddnet:18.9-rc3] 12 new commits

032cf5c Update Chinese translations for 18.9 - Pioooooo f0182eb properly clear autospeczooming state - TsFreddie cf04137 Revert "Scale angles using MousePos with zoom" - TsFreddie fec6cf7 Update brazilian_portuguese.txt - rffontenelle d21b834 remove spec zoom value and use user zoom always - TsFreddie 5fc9221 Update Polish translations for 18.9 - Shkyyl efce1c1 fix demo cursor - TsFreddie 356e243 update community info - TsFreddie fae2544 Update Ukrainian l10n (18.9) - EGYT5453 289e502 Update Swedish translations for 18.9 - furo321 20f590f Update spanish.txt for 18.9 - n0Ketchp 22dbf66 Support reconnecting dummy by redirecting to same port - Robyt3

cat_woah

1

23:12

[ddnet-web:master] 1 new commit

3d60474 Update translation stats - def-

[ddnet/ddnet] Pull request opened: #9484 Add a message to tell the client to reconnect to the current server

This is useful for implementing the semantics that @LW-Davide needs, without needing to hack it on top of the REDIRECT net message. CC #9480 CC #9382

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
[ ] Ch...

github-merge-queue[bot]

[ddnet:master] 2 new commits

c5805e7 Update spanish.txt for 18.9 - n0Ketchp e55cf29 Merge pull request #9483 from n0Ketchp/patch-24 - def-

23:28

[ddnet/ddnet] Pull request closed: #9483 Update spanish.txt for 18.9

That sounds like you are shooting at deen the full time database engineer

Replying to @Learath2 I wish we had a db wizard around that can help us a little

23:40

There are code bases that people complain about. And there are code bases nobody uses.

Replying to @Hecta ddnet code is so bad that there are currently 3 known rewrites going on …

[ddnet/ddnet-rs] Pull request opened: #18 Disable cursor in emote selection.

Not needed & desired anymore, because it requires a custom cursor anyway

i wish i wasnt so stupid so i could actually make the updates i want

you can only make updates you dont want?

kek

1

Exported 435 message(s)

Timezone: UTC+0