DDraceNetwork - developer

[ddnet/ddnet] Pull request opened: #9249 Add Simplify Tele popup setting

Adds a setting to revert #8961, as some mappers don't use all teleports on their maps. This change was requested on Discord. Also fixes an issue where the view instantly teleports to teleport number 1 when the teleport popup got first opened !image !image

Checklist

[x] Tested the change ingame
[ ] Provided sc...

Heinrich rejected the proposal of auto upgrade. But it would technically be possible.

Replying to @Solly can 0.7 not be auto detected on connect? or is the api so different now

ChillerDragon and you have a so-called "egg" for ddnetpp. Which is for a dedicated Pterodactyl server

?xd wat

no idea

03:38

hru chiller

It sounds like you're referring to Pterodactyl, which is indeed a popular open-source game server management panel. This software is widely used for hosting and managing game servers and applications. With Pterodactyl, you can control game servers across multiple machines and manage them through a web-based interface. It runs on Docker, which means each game server is isolated in its own container, helping with resource management and security. "Pterodactyl eggs" specifically refer to environment images or pre-configured setups for various games and services within the Pterodactyl panel. These "eggs" act as templates, containing scripts and settings that configure specific games or services so they can run smoothly on Pterodactyl. This makes deploying new servers for different games or applications much simpler. If you're setting up a server or need specific eggs, there are many available in the community, or you can create custom ones based on your needs.

03:45

https://github.com/engels74/pterodactyl-eggs

GitHub - engels74/pterodactyl-eggs: Service eggs for the pterodacty...

Service eggs for the pterodactyl panel. Contribute to engels74/pterodactyl-eggs development by creating an account on GitHub.

are you chat gpt

03:51

fork w 2 commits and 0 engagement

03:51

seems the upstream project doesn't even use the term pterodactyl

03:52

strange

03:52

https://pterodactyl.io/

Pterodactyl

Pterodactyl is an open-source game server management panel built with PHP, React, and Go. Designed with security in mind, Pterodactyl runs all game servers in isolated Docker containers while exposing a beautiful and intuitive UI to end users.

03:52

this is the thing

03:53

neat

Which file contains the Server Tick speed value?

murpii

[ddnet/ddnet] Issue opened: #9250 QoL: Add a new parameter to name_ban that ignores case sensitivity

The issue: if we ban a player named Jason, they could just rename to "jason" to avoid the ban. Using distance 1 would solve that, but it would also block players named Mason or "mason," which isn't ideal. A new parameter that ignores case sensitivity would resolve this issue. The given name is just an example. I have a couple individuals with a particularly unique names, where using the distance parameter is simply not doing the job.

lmfao

07:01

this guy does not read messages that are not from chillerdragon i see

Who? (edited)

that guy

07:44

who always talks to chiller

And who?

that guy

I'm afraid of the whipping, so I cannot tell you the name.

1

Is that me?

1

yes

08:41

That right

ChillerDragon: check my last message about your protocol documentation pwease

So I asked @heinrich5991 for a branch in order to implement the material layer without getting an answer (maybe I am too fast). The idea is, to add a layer which is painted over the gamelayer (similar to TPs or tune) which change tile behaviour according to a material. For example make a tile slippy, make it accellerate faster (booster tile, by running over it), add penalty tiles (slow down to low speeds), etc, a lot of well defined combinations of existing tunes. I already implemented it in a testbranch, but this feature went a bit over it's head and went huge. It has a lot of construction parts:

Handling new tune parameters for the materials
- Each material cold be it's own PR
Client needs a new layer in the editor, similar to tune
This comes with a physics update, how to handle standing on two different materials for example? I already put a lot of thought into this (this is implemented)
Prediction, old clients: Technically this all could be emulated similar to the tune layer, but I think this is a major break
Client animations: This one could also be done outside of it's own branch, for example use different ice sliding animations and particles on low friction ground, similar to here For a lot of these things I already have an implementation, and I would like to break it into small suggestable reviewable chunks on it's own branch, but I am not allowed to create one, what's your opinion?

which change tile behaviour according to a material. For example make a tile slippy, make it accellerate faster (booster tile, by running over it), add penalty tiles (slow down to low speeds), etc, a lot of well defined combinations of existing tunes.

i mean.. as you mentioned these are just "combinations of existing tunes" and is already possible with a few edgecases where it might not be. this sounds like a rather heavy change for small gain, espacially towards backwards compat (edited)

but wouldnt this make mapping a lot more stream line and simple especially for people who dont understand all the ins and outs of it?

10:02

honestly i couldnt know cuz i dont know the bare minimum when it comes to mapping :-]

you are still right, painting "ice" over wall would be more intuitive, than having to add 5 tunings to tune_zones in the map settings (edited)

Can I request kernel interfaces from antibot module? I want to hook up to server's IStorage

I think @meloƞ has a point, the benefit would come from the cleaner material interactions, like you are standing on ice and normal ground, and have the maximum of both friction (minimum of the values), the minimum ground control speed, maximum ground acceleration, average ground impule jump (imagine one is sticky material, one is normal, now you jump half as high), etc. This is also the part that breaks the prediction. But is it worth it? Maybe just implementing the client animations would be sufficient, since everything else can be emulated with tune layers (but not to such a detailed degree)

you know what i'd want? templates in the editor with these settings, a "community db" basically

10:27

with common settings, envelopes etc ready to use without having to create them from scratch

paid ddnet assets when

2

10:29

uh, that is also an idea, because I was thinking about adding a tuning parameter import as well, maybe we can put well defined materials into loadable configs

doesnt a new layer mean there will have to be implemented another map format version

10:30

which wont be compatible with teeworlds client and servers

you mean like the current one, which already isn't?

oh

10:31

didnt know that

default!!! teeworlds doesn't support teleports (edited)

1

depends on what kind of compatibility you mean, blockworlds is based on teeworlds, and the freeze is displayed as laser blockers in the DDNet editor for example kek

(edited)

yeah xd i dont know much about teeworlds

Solly

thats safe unless you derefence it (edited)

This is always safe, even if you dereference. Char pointers are allowed to alias any object type

@Assa Not sure if you're aware, I'm OP of #9250

https://github.com/ddnet/ddnet/issues/9250

Can someone help me find the values for deceleration and acceleration?

hi @murpi sorry if I sound harsh, I am just brutally honest

Then you don't seem to know what name_bans are for

TBH there is no point in name bans, isn't it banning by name?

At least on DDNet

Anime.pdf

Can I request kernel interfaces from antibot module? I want to hook up to server's IStorage

You'll have to change the ABI yourself. We intentionally don't really use anything from the server because it can break on accident, like the CPacker issue we had this week

why are name bans implemented, wtf, who thought of this, now implement skin bans

10:41

(╯°□°)╯︵ ┻━┻

Assa

TBH there is no point in name bans, isn't it banning by name?

They effectively take your points away without deleting ranks which we don't do

10:42

(Unless the ranks are replaybotted or use bugs we removed)

I responded to your comment, to give some examples (edited)

okay disallowing some names entirely makes sense

10:43

┬─┬ノ( º _ ºノ)

10:44

thought the idea was to get rid of the player, not to remove his records or prevent certain nicknames

Learath2

You'll have to change the ABI yourself. We intentionally don't really use anything from the server because it can break on accident, like the CPacker issue we had this week

Well, I think (thought), all game-server classes I use are compiled into .so and even if server will mess something up my AB will work fine. And because of this I compile my AB from ddnet sources, and not as a completely external project, should I worry about it?

Assa

thought the idea was to get rid of the player, not to remove his records or prevent certain nicknames

it's also used for certain.. other names it's useful given we dont have an account system and this is a proper way of stopping people to use certain names

Anime.pdf

Well, I think (thought), all game-server classes I use are compiled into .so and even if server will mess something up my AB will work fine. And because of this I compile my AB from ddnet sources, and not as a completely external project, should I worry about it?

Also main goal is to keep ABI compatible with stock ddnet

meloƞ

it's also used for certain.. other names it's useful given we dont have an account system and this is a proper way of stopping people to use certain names

okay disallowing some names entirely makes sense

Yes I understood, now honestly skin_ban would make sense

hm, why? the community and skin db's are controlled, if you use other proxies or download that skin yourself, your fault.

10:49

also people using disgusting skins can just.. not name them inappropriately

Anime.pdf

Well, I think (thought), all game-server classes I use are compiled into .so and even if server will mess something up my AB will work fine. And because of this I compile my AB from ddnet sources, and not as a completely external project, should I worry about it?

Well we compile the antibot with only a couple things from the ddnet server. That's system.h and antibot.h, nothing more iirc. Idk how you are compiling yours

is P2P skin sending a thing? When they change the skin name, others can't see them anymore or am I wrong?

Assa

is P2P skin sending a thing? When they change the skin name, others can't see them anymore or am I wrong?

No it's not. Yes they can't unless they have download skins enabled and you are using a skin that is part of their skin db provider

Learath2

Well we compile the antibot with only a couple things from the ddnet server. That's system.h and antibot.h, nothing more iirc. Idk how you are compiling yours

I include everything I need to antibot xd. base/system.h hash.h engine/storage.h message.h config.h, and much more, but only from engine and base

Base is mostly fine. Engine we sort of designed the antibot around not having to have the antibot touch the memory of the server at all

10:54

The ABI structs are all distinct from the ones the server uses internally e.g.

meloƞ

hm, why? the community and skin db's are controlled, if you use other proxies or download that skin yourself, your fault.

You could add a hook to the skins provider for banning a skin, supporting the community, but I think the benefit would be minimal as you'd need a skin that has slipped through (people do make mistakes) and been missed

Well, creating (mostly copying) all needed functionality to antibot sounds like a lot of pain and wasted time. So what was the CPacker issue? Also do you have own realization of it in AB? (edited)

Anime.pdf

Well, creating (mostly copying) all needed functionality to antibot sounds like a lot of pain and wasted time. So what was the CPacker issue? Also do you have own realization of it in AB? (edited)

We got lazy and used CPacker on the antibot, for like the first time in a decade someone made an ABI breaking change in CPacker. It started crashing because old antibot module new server

That's weird, even if CPacker changed in server binary, shouldn't old version of it be compiled into AB? ABI gives only m_pfnSend that accepts raw data, all packing is made purely on AB side(at least in my AB)

Anime.pdf

That's weird, even if CPacker changed in server binary, shouldn't old version of it be compiled into AB? ABI gives only m_pfnSend that accepts raw data, all packing is made purely on AB side(at least in my AB)

That's not how it's linked up

So it's really linked to server, I though it was simply copied to AB binary

Idk which message you mean o.0 could you resend it?

Replying to @Koll Potato ChillerDragon: check my last message about your protocol documentation p…

Anime.pdf

So it's really linked to server, I though it was simply copied to AB binary

It's really hard to know if we are both talking about the same thing, but yes it is truly actually linked if you make calls into the server

How do I specify the number of percentages of people from the total number of online users on the server who must press f3 in order for the vote to be completed

Learath2

It's really hard to know if we are both talking about the same thing, but yes it is truly actually linked if you make calls into the server

I'm talking about classes like CPacker, I thought they are compiled into AB so it uses own standalone version of it, but looks like it's referencing server's CPacker instead and is dependent of it

11:08

This makes everything much more complicated. I guess I will migrate my AB to very standalone project that will use own classes instead of referencing, I hope this will make it independent of server realization. But I'm scared to think how much it will weight

11:08

Imagine 5mb libantibot.so

11:08

how can i import the deathmatch mode to my server (edited)

ChillerDragon: so basically, in the 0.7 protocol documentation, in one place, it's written there are 501 null bytes added to the end of the token control packet and in another 508 null bytes. (I tested both cases and 508 is the correct one). (edited)

<ChillerDragon> Oh yea ofc 508 is the correct one. Thanks @Koll Potato i recently re counted it and only updated it in one place. Thanks for reporting @Koll Potato

np :D

<ChillerDragon> @hakan download this unzip it and double click on DDNet-Server.exe then login as admin in f2 (rcon console) and type in sv_gametype dm and then reload https://zillyhuhn.com/tmp/ddnet-windows-latest.zip

thanks for that, but i have linux server

12:46

does it have a linux version?

if anyone have windows installed, can you check what's the typedef for size_t in MSVC?

hakan

does it have a linux version?

Yes it does nouis

Koll Potato

Yes it does nouis

can i have a link

https://ddnet.org/downloads/

<ChillerDragon> @Koll Potato he wants death match

13:38

<ChillerDragon> @hakan do you know how to build from source?

13:38

<ChillerDragon> https://github.com/ddnet-insta/ddnet-insta/

GitHub - ddnet-insta/ddnet-insta: A teeworlds instagib (grenade/las...

A teeworlds instagib (grenade/laser capture the flag/death match/catch) mod based on DDRaceNetwork (gctf/ictf/idm/gdm/zcatch) - ddnet-insta/ddnet-insta

umm

13:39

i imported it

ws-client

<ChillerDragon> https://github.com/ddnet-insta/ddnet-insta/

this

13:39

but

13:39

the sv_gametype command didnt worked

<ChillerDragon> how so?

13:39

<ChillerDragon> did you reload

ye

<ChillerDragon> what happend?

can you come to my server?

<ChillerDragon> yes

wait

13:40

77.90.53.24

13:40

the password is 31

<ChillerDragon> can't reach it

13:43

<ChillerDragon> ah nvm im in xd

13:45

<ChillerDragon> @hakan

sudo apt install build-essential cargo cmake git glslang-tools google-mock libavcodec-extra libavdevice-dev libavfilter-dev libavformat-dev libavutil-dev libcurl4-openssl-dev libfreetype6-dev libglew-dev libnotify-dev libogg-dev libopus-dev libopusfile-dev libpng-dev libsdl2-dev libsqlite3-dev libssl-dev libvulkan-dev libwavpack-dev libx264-dev python3 rustc spirv-tools

thx

git clone --recursive https://github.com/ddnet-insta/ddnet-insta.git && mkdir ddnet-insta/build && cd ddnet-insta/build && cmake .. && make

ChillerDragon

<ChillerDragon> @Evelyn

I'm currently helping a friend Clone ddnetpp

<ChillerDragon> pro

What kind of team is there , something is wrong in the github

13:48

There is no cmake fall

<ChillerDragon> ?

13:50

Weird

<ChillerDragon> @Evelyn

sudo apt install build-essential cargo cmake git glslang-tools google-mock libavcodec-extra libavdevice-dev libavfilter-dev libavformat-dev libavutil-dev libcurl4-openssl-dev libfreetype6-dev libglew-dev libnotify-dev libogg-dev libopus-dev libopusfile-dev libpng-dev libsdl2-dev libsqlite3-dev libssl-dev libvulkan-dev libwavpack-dev libx264-dev python3 rustc spirv-tools

hehe copy paste

And then what

<ChillerDragon> @Evelyn https://zillyhuhn.com/cs/.1731678647.png here it tells you you don't have a c compiler installed

13:51

<ChillerDragon> a new system ships without the needed depdencies to build ddnet so you got to install them once using apt

13:51

<ChillerDragon> https://github.com/ddnet/ddnet?tab=readme-ov-file#dependencies-on-linux--macos

GitHub - ddnet/ddnet: DDraceNetwork, a free cooperative platformer ...

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

After he downloads it all:cd DDNetPP/build and is needed cmake .. make

<ChillerDragon> if he is still in the build folder he does not need the cd

13:54

<ChillerDragon> but yes running cmake .. and make again in the build directory

@Learath2 why most tools default to at&t syntax?

13:54

intel looks way easier and sane

<ChillerDragon> @hakan are you german? o.O

nah

13:59

turkish

14:00

can you come again

<ChillerDragon> omw

14:04

what to do

<ChillerDragon> can you show more of the error message?

14:07

<ChillerDragon> what kind of system is that?

14:07

<ChillerDragon> what does cat /etc/os-release show

14:07

<ChillerDragon> and g++ --version

Ryozuki

@Learath2 why most tools default to at&t syntax?

I honestly don't know this one. I also don't know anyone who uses att syntax

linux

<ChillerDragon> yea that much i assumed xd

14:08

<ChillerDragon> but which distro which version

Wait

14:10

<ChillerDragon> ok debian 10 not good

14:11

<ChillerDragon> too old

<ChillerDragon> get debian 12

14:12

<ChillerDragon> where did your friend get a server? is it new? can he select debian 12 instead of debian 10?

14:13

<ChillerDragon> if there is no option to run a newer debian you have to update your self for example following this https://linuxize.com/post/how-to-upgrade-debian-10-to-debian-11/

How to Upgrade Debian 10 Buster to Debian 11 Bullseye

This article shows how to upgrade your Debian 10 Buster system to Debian 11, Bullseye via command line.

He asked to install Debian 12

14:15

Just in case, ChillerDragon, write what to do next, I remember that you will go to bed soon

<ChillerDragon> if you have debian 12 it will be as easy as on your machine

14:17

<ChillerDragon> just clone and build like you already tried

ChillerDragon I actually got the same error about CMAKE_CXX_COMPILER as well, but that was right after installing both the compiler and cmake, on a system with like a year of uptime, but it compiled just fine when I manually specified the env var as g++

14:22

oof that was an ugly sentence...

14:23

main point: it's possible for that to show up even if you have a compiler

main point: you can starve even if you have food in the fridge

<ChillerDragon> @risu that sounds cursed if you have a healthy up to date system and a compiler installed with apt then cmake should find it

idk if it's very healthy, I haven't really done any active maintenance on it

bruh wtf

https://0ver.org/

ZeroVer: 0-based Versioning — zer0ver

Software's most popular versioning scheme!

i was away only for 1 second

@murpi spam (edited)

14:28

@Discord Mod

Discord takes a bit to delete all his messages

I feel like the compiler not being detected would've probably been fixed simply by opening a new shell

yes

14:29

env is not updated sometimes

ChillerDragon they can't install debian 12, will they have to upgrade from under the system?

the issue I'm more familiar with is having to open a new shell when my groups change

i have debian 12 btw

14:30

u can dual boot

14:30

with both windows and linux

<ChillerDragon> @Ryozuki common teeworlds W

I feel like this is probably not a home computer if they can't just upgrade it like that...

14:31

and if it's not a home computer I would absolutely not recommend trying anything Windows related

running a vm is another option

14:31

slow but works

<ChillerDragon> @Evelyn then you gotta do this https://linuxize.com/post/how-to-upgrade-debian-10-to-debian-11/ hf gl

How to Upgrade Debian 10 Buster to Debian 11 Bullseye

This article shows how to upgrade your Debian 10 Buster system to Debian 11, Bullseye via command line.

risu

main point: it's possible for that to show up even if you have a compiler

CMAKE_CXX_COMPILER is left empty in favor of CMAKE_CXX_COMPILER_INIT which gets its value from the env var, iirc

14:33

or maybe that’s only flags & stuff

14:34

you still want the variable set but ya it will often work

How to make more slots on the server than 64

MilkeeyCat

if anyone have windows installed, can you check what's the typedef for size_t in MSVC?

unsigned __int64 according to https://godbolt.org/z/KE9se6E48

Compiler Explorer - C++

int main() { size_t x = 42; std::cout << typeid(x).name() << std::endl; return 0; }

I didn't know printing out type's name is possible :0

15:38

and thanks

Robyt3

unsigned __int64 according to https://godbolt.org/z/KE9se6E48

does this work for structs/classes?

Teero

does this work for structs/classes?

Yes, prints class CMyClass for example

ah nice

Learath2

This is always safe, even if you dereference. Char pointers are allowed to alias any object type

((char)this + 5) = -1 Jokes on you you just overwrote the first byte of a uint32 as 0b11111111. This happened to be a size of an array. The array is then iterated over and causes a segfaults, but not before writing to a bunch of things out of bounds causing an accidental zero day which deletes the French language pack from your system, bricking it. Reading from it if you stay inside the object should be safe (or if you're in an array even further). But writing can cause unsafe stuff

It's not the writing that causes ub as long as that pointer is still within the object

15:43

The dereferencing, reading, writing, all are fine. It's the rest of your program that causes ub by respecting that new length and writing beyond the object

Learath2

The dereferencing, reading, writing, all are fine. It's the rest of your program that causes ub by respecting that new length and writing beyond the object

mmm

16:00

oke

(You are right that it's usually not a great idea and might cause unexpected stuff, but "undefined behaviour" is a very specific thing and you need to be exact about what is UB and what isn't because C/C++ is just very strict about it) (edited)

There's UB you need to be scared of and UB that just gets regularly abused for convenience

16:04

because it isn't all that undefined

16:05

I do need to teach myself to stop writing code that assumes x86... I like low power systems so I really should make my code work on ARM and the like

How to make more slots on the server than 64

Write C code until you introduce UB Challenge (IMPOSSIBLE DIFFICULTY): Survive for 5 minutes

Evelyn

How to make more slots on the server than 64

You have 3 options:

Learn to code, figure it out yourself
Learn to code, start from fokkonaut's ddnet PR and fix that up
Beg one of the 2 people that have decently well working implementations of it to implement it for you

risu

Write C code until you introduce UB Challenge (IMPOSSIBLE DIFFICULTY): Survive for 5 minutes

https://www.youtube.com/watch?v=oTEiQx88B2U

Solly

https://www.youtube.com/watch?v=oTEiQx88B2U

I remember watching this :P

Learath2

You have 3 options:

Learn to code, figure it out yourself
Learn to code, start from fokkonaut's ddnet PR and fix that up
Beg one of the 2 people that have decently well working implementations of it to implement it for you

do older clients accept more than 64 clients, but just cant show them?

16:09

fokkonauts playground laready has 128 slots and i can join fine

Some "engineer" at youtube decided to remove the confirmation before changing the video that is currently being cast to tv. So clicking that link made me lose my place in a 1h video essay

Solly

do older clients accept more than 64 clients, but just cant show them?

No they don't. All implementations of 128p use an id mapping, they map the 128 players to 64, using heuristic measures like closest/most likely to interact with players (edited)

bw historically goes up to 256p and works the same, tho I never use it lmao

Any MSVC enjoyers? I gen an error when I try make a typedef for size_t: 'size_t': redefinition; different basic types(C btw). For some reason in MSVC size_t can be used without including a header file, is there a compiler flag which can change this behavior or maybe a macro which indicates that the type is defined :\

(multiple maps, single server instance)

MilkeeyCat

Any MSVC enjoyers? I gen an error when I try make a typedef for size_t: 'size_t': redefinition; different basic types(C btw). For some reason in MSVC size_t can be used without including a header file, is there a compiler flag which can change this behavior or maybe a macro which indicates that the type is defined :\

Why are you typedefing size_t?

Learath2

No they don't. All implementations of 128p use an id mapping, they map the 128 players to 64, using heuristic measures like closest/most likely to interact with players (edited)

oh no it's dirty hacks all the way down

Learath2

No they don't. All implementations of 128p use an id mapping, they map the 128 players to 64, using heuristic measures like closest/most likely to interact with players (edited)

can clients support 128 clients and not do this jank?

risu

oh no it's dirty hacks all the way down

It's somewhere in the roadmap for ddnet to get native 128p support or n player support

Solly

can clients support 128 clients and not do this jank?

ddnet client already does this I think?

Solly

can clients support 128 clients and not do this jank?

Soon(tm) as in hopefully within the next 6 months

wait no I'm dum but like, doesn't the player list support that?

<ChillerDragon> isnt that just setting MAX_CLIENTS to 128 and call it done?

ok but then won't it break for older clients?

ws-client

<ChillerDragon> isnt that just setting MAX_CLIENTS to 128 and call it done?

I think we had snap size issues :/

when /show-everyone-which-i-can-seee-when-zoomed-out-instead-of-actually-everyone-cuz-thats-dumb

<ChillerDragon> oh yea that makes sense but whats the solution for that? increase snap size?

If you want to up it to 128 and test it out it would be appreciated

16:15

Also there are 64 wide masks that need to be widened to std::bitset<MAX_CLIENTS>

16:15

This change to 128? (edited)

<ChillerDragon> I am too tired now. But I have the feeling there has to be some low hangung fruit that could already be released now on the client side to later not have to use player mappings anymore

Learath2

Also there are 64 wide masks that need to be widened to std::bitset<MAX_CLIENTS>

gcc __int128 anyone? (edited)

<ChillerDragon> @Evelyn not it requires quite a bit of writing actual code to get 128 players to work. Remind me tomorrow and I can have a look.

Solly

gcc __int128 anyone? (edited)

But then we'll talk about this again in 2034 when upping it to 256p

Learath2

But then we'll talk about this again in 2034 when upping it to 256p

yep

Ok

at that point __int256 will be a thing xd

16:17

supported by only 128 bit architectures

16:18

this is al a /j btw

ChillerDragon maybe I'll remind you about slow motion tomorrow? It's not like making it. And choose the actions that need to be done in tune to slow down or accelerate

I wonder if we'll ever get 128b architectures. The size of the address/data bus seems to be enough for pretty much everything we can imagine for now

Learath2

Also there are 64 wide masks that need to be widened to std::bitset<MAX_CLIENTS>

I think that was fixed or are there some remaining? We have a typedef std::bitset<MAX_CLIENTS> CClientMask;

Robyt3

I think that was fixed or are there some remaining? We have a typedef std::bitset<MAX_CLIENTS> CClientMask;

Oh, cool

Learath2

I wonder if we'll ever get 128b architectures. The size of the address/data bus seems to be enough for pretty much everything we can imagine for now

ugh yeah but what if im in the year 292,277,026,596.927,714,913,658???

1

<ChillerDragon> @Learath2 client side masks? For predictions?

Solly

at that point __int256 will be a thing xd

Well we do already have 512 bit registers on some CPUs...

risu

Well we do already have 512 bit registers on some CPUs...

avx registers for storing bitsets?

hehe

16:23

I mean, pretty sure you can do bitwise operations on them, wouldn't make sense to not have the basics in there

16:23

probably slower than just doing it 64 bits at a time lol

yeap

As for performance, how many tees can we handle before it becomes an issue?

16:25

I guess network would be the biggest bottleneck

risu

As for performance, how many tees can we handle before it becomes an issue?

bros are running ddnet at 1000s of frames per second

16:25

so yes network

since actually processing a few thousand tees would be nothing

also alot of people play with /showall which is a huuuuuge waste

16:27

/showall needs to be nuked (edited)

risu

As for performance, how many tees can we handle before it becomes an issue?

I think 512 was breaking down with the current code

We need a distributed Teeworlds server, make an MMO, call it World of Tees

4

Learath2

I think 512 was breaking down with the current code

jupstar can handle more i think (edited)

Teero

jupstar can handle more i think (edited)

That's why I said current code, without touching much. His rewrite in rust can indeed do much more

Learath2

Why are you typedefing size_t?

because there's no existing typedef for it

Learath2

That's why I said current code, without touching much. His rewrite in rust can indeed do much more

what rewrite in rust o-o?

Solly

what rewrite in rust o-o?

i assume they are talking about this https://github.com/Jupeyy/dd-pg

GitHub - Jupeyy/dd-pg: don't use or read this code

don't use or read this code. Contribute to Jupeyy/dd-pg development by creating an account on GitHub.

Solly

what rewrite in rust o-o?

Jupsti rewrote DDnet in rust

16:52

Hey respect him saying to not read it

16:52

Smh

i didnt

16:52

and i definitely havent cloned the repo

16:52

and didnt run it

MilkeeyCat

because there's no existing typedef for it

for now I just check for _MSC_VER and if it's set don't make the typedef

meloƞ

Jupsti rewrote DDnet in rust

error: failed to get `hashlink` as a dependency of package `base v0.1.0 (/home/solly/Code/dd-pg/lib/base)`

Caused by:
  failed to load source for dependency `hashlink`

16:59

hmm

you need nightly toolchain (edited)

nightly rust?

17:01

  failed to authenticate when downloading repository: git@github.com:Jupeyy/hashlink/

  * attempted ssh-agent authentication, but no usernames succeeded: `git`

MilkeeyCat

because there's no existing typedef for it

You are supposed to include stddef.h for it. Why don't you just do that? Then on every compiler a typedef will exist

Solly

nightly rust?

ye

1

Learath2

You are supposed to include stddef.h for it. Why don't you just do that? Then on every compiler a typedef will exist

because stddef.h doesn't work for kernel module :p

MilkeeyCat

because stddef.h doesn't work for kernel module :p

? Why are you compiling a linux kernel module with msvc or a windows kernel driver with gcc?

MilkeeyCat

because stddef.h doesn't work for kernel module :p

Also stddef.h is part of the compiler, I see no reason why you shouldn't be allowed to use it

Learath2

Also stddef.h is part of the compiler, I see no reason why you shouldn't be allowed to use it

configure: nasm/yasm not found or too old. Use --disable-x86asm for a crippled build. hehe crippled

17:06

WHY IS IT CLONING FFMPEG??

why not

Solly

configure: nasm/yasm not found or too old. Use --disable-x86asm for a crippled build. hehe crippled

sudo apt install nasm

for rendering ig

theres a reason i dont like rust

huh

17:07

rust hater??

17:07

i dont mind the language, the ecosystem sucks tho

MilkeeyCat

Click to see attachment 🖼️

huh

Solly

theres a reason i dont like rust

its not the reason, you will have plenty others when you will code rust for real

Solly

i dont mind the language, the ecosystem sucks tho

imo its the best one there is

Solly

i dont mind the language, the ecosystem sucks tho

lmfao

_kernel_size_t perhaps?

im coming from javascript sooo

17:07

its better

zhn

for rendering ig

ffmpeg doesnt do rendering

17:07

and i already have ffmpeg installed

Solly

ffmpeg doesnt do rendering

for rendering videos i mean

17:08

like demo to mp4 encoding

There is plenty of uses of size_t in the kernel, idk what your issue is tbh

zhn

like demo to mp4 encoding

o i guess that is fair, but you can hook ffmpeg to your window

17:08

it doesnt need to be a compile time dependency

languages that ship with package managers are dangerous because they end up with stupid dependency trees for basic apps

ask @Jupstar ✪

and if anything should be installed for it, it should be a c -> rust compat layer

Solly

and if anything should be installed for it, it should be a c -> rust compat layer

but it is?

zhn

ask @Jupstar ✪

its probably a dependency of a dependency cuz rust

the difficulty of handling dependencies for C makes the software better lol

risu

languages that ship with package managers are dangerous because they end up with stupid dependency trees for basic apps

https://www.npmjs.com/package/is-even

is-even

Return true if the given number is even.. Latest version: 1.0.0, last published: 7 years ago. Start using is-even in your project by running npm i is-even. There are 49 other projects in the npm registry using is-even.

Learath2

There is plenty of uses of size_t in the kernel, idk what your issue is tbh

I'm making a linux kernel module and I can use size_t and everything works, but I'm also making a library which can be used by the kernel module

risu

languages that ship with package managers are dangerous because they end up with stupid dependency trees for basic apps

yeah, cpp ends up with manual bootstrapping of such xddd

also takes many years

17:10

omg ai

and a good internet connection

you can download it once and work as you wish

I don't want to include any kernel related headers so I did this https://github.com/MilkeeyCat/ddnet_protocol/blob/master/src/common.h

offline i mean

Solly

WHY IS IT CLONING FFMPEG??

how else do you expect audio to work

meloƞ

how else do you expect audio to work

audio?? xd

meloƞ

how else do you expect audio to work

ffmpeg shouldnt handle audio playing

17:11

and that would be ffplay anyway

ofc audio, noone would be crazy enough to use it for videos!

xddd

xd

stop troll

Solly

o i guess that is fair, but you can hook ffmpeg to your window

ddnet doesn't just shell out to ffmpeg, it uses libavcodec/libavformat (the libraries making up ffmpeg) to encode directly, which is much more robust

1

MilkeeyCat

I don't want to include any kernel related headers so I did this https://github.com/MilkeeyCat/ddnet_protocol/blob/master/src/common.h

Why can't your library include stddef? I don't get that part. Is your compiler misconfigured or sth? (edited)

Solly

o i guess that is fair, but you can hook ffmpeg to your window

just grep ffmpeg in sources and look what it is used for

17:13

you'd need to grep for av or avcodec

Learath2

Why can't your library include stddef? I don't get that part. Is your compiler misconfigured or sth? (edited)

I can include it in normal projects but it doesn't work with kernel module :\

arch1t3cht

you'd need to grep for av or avcodec

this is not about ddnet

arch1t3cht

you'd need to grep for av or avcodec

but we are talking about ddnet-pg, not ddnet

17:14

ddnet ofc doesn't use full ffmpeg bundle, but seems ddnet-pg does

no more issues allowed

17:15

oh god...

ah I didn't read enough backlog it seems

MilkeeyCat

I can include it in normal projects but it doesn't work with kernel module :\

The kernel itself includes stddef all over the place, so it's not like a rule or sth. Anyway, idk

zhn

just grep ffmpeg in sources and look what it is used for

yeah but it means it wouldnt be a compiletime dependency but runtime optional dependency

17:16

which is yipee

17:16

also it feels like teeworlds

17:16

very ... rough in every way but it works :D

reminder that ddnet is just a teeworlds mod, so it makes sense kek

Solly

very ... rough in every way but it works :D

crashed after a minute of gameplay for me

17:16

but it worked

17:16

so epic

Koll Potato

crashed after a minute of gameplay for me

i had to host a local server

skill issue, i played hours of gores on it, and i also beat @zhn in a few rounds of ctf (edited)

meloƞ

skill issue, i played hours of gores on it, and i also beat @zhn in a few rounds of ctf (edited)

can it join normal servers nowadays?

meloƞ

skill issue, i played hours of gores on it, and i also beat @zhn in a few rounds of ctf (edited)

in (old) ddrace servers

17:17

?

Koll Potato

crashed after a minute of gameplay for me

skill issue, played insta for hours on it

Learath2

can it join normal servers nowadays?

iirc jupsti doesnt intend for it to be compatible (edited)

arch1t3cht

ddnet doesn't just shell out to ffmpeg, it uses libavcodec/libavformat (the libraries making up ffmpeg) to encode directly, which is much more robust

this is still true though

I thought Jupstar was intentionally never going to support normal ddnet/teeworlds protocol

meloƞ

skill issue, i played hours of gores on it, and i also beat @zhn in a few rounds of ctf (edited)

yeah, mister 20 ping

1

ye

zhn

yeah, mister 20 ping

skill issue

zhn

skill issue, played insta for hours on it

skill issue, i touched grass for hours

meloƞ

skill issue

Learath2

I thought Jupstar was intentionally never going to support normal ddnet/teeworlds protocol

yeah, it doesn't justatest

dd-pg feels like ddnet with antiping on by default, it felt like cheating

1

i had 60-9999 ping, still playable enough

dd-pg is better in most ways imo

It's such a shame that's the direction he took with it

is he going to make a brand new protocol or what?

he mentioned he will allow modifications when he releases it

17:19

so we might see that at some point

It'll cause unimaginable amounts of migration pain and community splitting for absolutely 0 reason

meloƞ

he mentioned he will allow modifications when he releases it

my insta is ready for tweaking brownbear

zhn

my insta is ready for tweaking brownbear

my zombies can't wait to be implemented boom

Learath2

It'll cause unimaginable amounts of migration pain and community splitting for absolutely 0 reason

for 0 reason?

Ok -1 reason

Learath2

It'll cause unimaginable amounts of migration pain and community splitting for absolutely 0 reason

it is a different game

he'll sell dd-pg, "for only 29,99$"

he literally remade teeworlds in rust with accounts in mind

Teero

it is a different game

A different game that looks and behaves almost exactly the same way as another

i wouldn't say he had 0 reasons to do it

zhn

i wouldn't say he had 0 reasons to do it

Many reasons to give ddnet/teeworlds a proper rewrite. 0 reasons not to properly support the current set of mods

he was tired of needing to do backwards compatible things and getting nowhere

we can do that together when he releases it

17:21

wasm ftw

17:22

ok i stop trolling

remaking the mods is easier than maintaining backwards compatibility for the next 10 years

17:22

also mods in dd-pg are much more powerful

When he releases it properly one project is definitely going to die off. We barely have the people to maintain this project let alone give it a complete rewrite in wasm to support dd-pg

Learath2

Many reasons to give ddnet/teeworlds a proper rewrite. 0 reasons not to properly support the current set of mods

since ddnet 2.0 was a thing as an idea of throwing out old ddnet/teeworlds i wouldn't say it either

twgame exists already with most stuff implemented

17:23

dd-pg uses that currently as a physics module and it works quite well

didn't heinrich make his own proper protocol based on quic?

even tho many things are not done yet

Just remember me when the game dies, that's all I ask

17:24

You can tell people there was this old 80 year old dude that warned us that breaking 10 years worth of mods wasn't the best idea

1

bruh, i heard that almost on every major release

17:24

better thing will live, thats all

zhn

didn't heinrich make his own proper protocol based on quic?

quic is just wrapping the already existing protocol 1 to 1, it's a different transport, not a different protocol

0.5 died, 0.6 lived

17:25

0.7 bappened, 0.7 died

zhn

0.5 died, 0.6 lived

We had the time, the people and the talent to port ddnet to 0.6. We have none of the above nowadays

Learath2

quic is just wrapping the already existing protocol 1 to 1, it's a different transport, not a different protocol

wait so there was no work done on making a better protocol?

Learath2

We had the time, the people and the talent to port ddnet to 0.6. We have none of the above nowadays

because 0.5 and 0.6 were pretty similar, aren't they

zhn

wait so there was no work done on making a better protocol?

There is no need for a "better" protocol as was mentioned to you many times now. The protocol is just fine except for the fact that it contains some crud

code wise, protocol wise

zhn

because 0.5 and 0.6 were pretty similar, aren't they

Almost exactly the same, the bulk of the work was that a lot of code was moved around by ddrace and by teeworlds, so when 0.6 came out merging the two was almost impossible

17:27

Hours and hours of resolving merge conflicts and testing, but as I said back then we had much more time on hand with more people

17:28

btd was still around, greyfox was here

Learath2

There is no need for a "better" protocol as was mentioned to you many times now. The protocol is just fine except for the fact that it contains some crud

so still we will have no server side assets, no dynamic map chunk patching, etc

zhn

so still we will have no server side assets, no dynamic map chunk patching, etc

Those have nothing to do with quic, correct

and it doesn't solely depend on protocol, map format should be reconsidered too as it was discussed many times before

Learath2

Those have nothing to do with quic, correct

i don't argue about quic being a new protocol, i was thinking that heinrich tries to improve existing one and implement quic base for it

17:29

nothing else

zhn

so still we will have no server side assets, no dynamic map chunk patching, etc

"the protocol" isn't stopping you from implementing these, it's perfectly capable of expressing these things

Learath2

"the protocol" isn't stopping you from implementing these, it's perfectly capable of expressing these things

yeah, tell me

zhn

i don't argue about quic being a new protocol, i was thinking that heinrich tries to improve existing one and implement quic base for it

quic is mostly just replacing udp 1 to 1, and giving us transport security, the second one is needed for accounts to be secure

backcompat is part of ddnet and it was the only reason why jupstar did it

zhn

so still we will have no server side assets, no dynamic map chunk patching, etc

holdon i got an xkcd for that

if ddnet eould not support older clients we would not have to see this reimplementation

17:30

jupstar would just work on that in current codebase

And when he does finish up and release his work you can all have fun with your shiny new clients playing vanilla

yeah, i will praise it, because its better and more flexible

Solly

holdon i got an xkcd for that

https://en.wikipedia.org/wiki/Greenspun%27s_tenth_rule

Greenspun's tenth rule

Greenspun's tenth rule of programming is an aphorism in computer programming and especially programming language circles that states: Any sufficiently complicated C or Fortran program contains an ad hoc, informally-specified, bug-ridden, slow implementation of half of Common Lisp.

zhn

yeah, i will praise it, because its better and more flexible

certainly, at the cost of throwing away a decade of modding

17:32

thousands and thousands of human hours spent on developing all the mods that have to be completely re-done

Learath2

certainly, at the cost of throwing away a decade of modding

we will be able to make more stuff with it anyway

i think u want roblox (that is a game engine which has inherit multiplayer functionality which can do anything)

and since wasm modules are reusable, these will be easily transferred

Solly

i think u want roblox (that is a game engine which has inherit multiplayer functionality which can do anything)

a framework for ddnet servers

Solly

i think u want roblox (that is a game engine which has inherit multiplayer functionality which can do anything)

its called game engine smh

can you.. read the thing in brackets

and yes modders always considered teeworlds as game engine with its own constraints

zhn

we will be able to make more stuff with it anyway

you'll slowly recreate the mods we currently have, while slowly leeching off players and devs, which will kill both because both will end up too small to be sustainable

not all game engines have inherit multiplayer functionality, and those which do often require a lot of code to do so or are easy to cheat on

zhn

and yes modders always considered teeworlds as game engine with its own constraints

although oy would always say opposite

many an open source project died this exact death, fork -> dev split -> not enough dev time for either -> dead project

Learath2

you'll slowly recreate the mods we currently have, while slowly leeching off players and devs, which will kill both because both will end up too small to be sustainable

does anyone force all the devs to move to dd pg?

17:35

i don't see it

17:35

also we don't have so much active developers to be actually split

zhn

does anyone force all the devs to move to dd pg?

No, but dd-pg needs dev time and there is a finite amount of it unless you think new developers are going to show up to help recreate the library of mods

17:38

Anyway, think whatever you like. I'm just making a prediction based on what I've seen happen to other projects and what I've seen happen within this project. You obviously disagree, that's fine

if this is an election im siding with the "dont fork or split anything, jupeey can do whatever they want, if it attracts people then let it"

Learath2

No, but dd-pg needs dev time and there is a finite amount of it unless you think new developers are going to show up to help recreate the library of mods

Well the rust cult is pretty good at doing just that...

17:38

They love rewriting things :)

wd40 keep me safe

Hopefully someone will implement a compatibility layer to solve it, but given the amount of activity we have nowadays that's even doubtful

the only developers we actually have to contribute to ddnet upstream are... robyt3, because no one else wants to waste their time on thing they implemented in their own fork to be closed just like any other major prs we had. i don't say that there's no reason to close them but that's the thing

17:39

the only great thing that was added from another fork is fokkos 128 id mapping, thats all i can recall for now

zhn

the only developers we actually have to contribute to ddnet upstream are... robyt3, because no one else wants to waste their time on thing they implemented in their own fork to be closed just like any other major prs we had. i don't say that there's no reason to close them but that's the thing

You and I obviously have very different ideas on how this project is supposed to run, so it's quite normal we disagree

Solly

Click to see attachment 🖼️

so what?

zhn

so what?

das alot of people

Learath2

You and I obviously have very different ideas on how this project is supposed to run, so it's quite normal we disagree

ye, but i still don't see any reason for cpp developers to split over rust project

zhn

ye, but i still don't see any reason for cpp developers to split over rust project

There is a limited amount of developers in general that are interested in teeworlds/ddnet, most of us know both C++ and Rust. There won't be random Rustaceans showing up to write mods (edited)

Solly

das alot of people

It's understandable people are annoyed that not everything is getting merged and getting things merged is the only way to mod this game properly

Learath2

There is a limited amount of developers in general that are interested in teeworlds/ddnet, most of us know both C++ and Rust. There won't be random Rustaceans showing up to write mods (edited)

i was about to say that we have less rustaceans than cpp enjoyers there

17:43

at least most active ones, i.e chiller and robyt3

Learath2

It's understandable people are annoyed that not everything is getting merged and getting things merged is the only way to mod this game properly

u can do alot of server side only. the problem is alot of mods are unmaintaed, not open source and hard to find

Learath2

It's understandable people are annoyed that not everything is getting merged and getting things merged is the only way to mod this game properly

This is also exactly why ddrace moved away from teeworlds client, providing a smooth experience requires client support

Learath2

This is also exactly why ddrace moved away from teeworlds client, providing a smooth experience requires client support

history repeats itself, eh

zhn

history repeats itself, eh

Not really, we were all extremely careful not to break backwards compatibility so as to ensure a unified playerbase and not splitting the community

1

Learath2

Not really, we were all extremely careful not to break backwards compatibility so as to ensure a unified playerbase and not splitting the community

i just played on vanilla 0.6, still works fine

17:45

:D

Learath2

Not really, we were all extremely careful not to break backwards compatibility so as to ensure a unified playerbase and not splitting the community

all other clients died just because they needed to handle with ddrace support, didn't they?

All other clients died just because the ddrace client was just technically superior and a superset to all of them

and all major cool stuff just was merged in ddrace client, just to not waste time duplicating it

Learath2

All other clients died just because the ddrace client was just technically superior and a superset to all of them

not vanilla ones

ddnet is just a continuation of ddrace, so now that's the most useful client

gamer and tdtw are still better in ui ux way for vanilla

zhn

gamer and tdtw are still better in ui ux way for vanilla

Vanilla died in general, how many people play it now, 20-30?

17:47

If there were more players there would be popular maintained vanilla clients

Learath2

Vanilla died in general, how many people play it now, 20-30?

we have at least 30 russian vanilla player rn, i don't know numbers from eu and us

i wana play some ctf ):

17:48

the non i/g kind

Learath2

If there were more players there would be popular maintained vanilla clients

newbies enter vanilla mods and ask us "new mod?"

the kind where its 8 per side on ctf5 and its 100 100 (both sides captured the flag 100 times)

theh also tend to play on ddnet vanilla servers, they just don't see any other

Anyway, the lack of actively maintained vanilla clients is not because vanilla clients didn't handle ddrace well. It's because there are no active developers in the vanilla sphere

17:49

ddrace based clients are just easier to maintain because you can piggyback off of our maintainance of it

imagine trying to handle ddrace updates with client you code solely

17:50

ofc they dropped it

They dropped it because these developers aren't involved much at all with the game anymore. Gamer client is by Dune, he had 6 contributions in the last 5 years

it's one of the reasons, not the only one

17:53

psychogod was active enough but still dropped tdtw (edited)

What does ddpg offer to players over ddnet? I find hard to envision it replacing ddnet unless it's an official switch

If there was a demand for a vanilla client, there would be a maintained vanilla client based on ddnet at the very least, if you think ddrace support is so important that it'd kill the client

Learath2

If there was a demand for a vanilla client, there would be a maintained vanilla client based on ddnet at the very least, if you think ddrace support is so important that it'd kill the client

every good vanilla player has their own tweaked ddnet client with few adjustments :)

17:54

and that's thr way how small things like clan in spectators come into ddnet

Tater

What does ddpg offer to players over ddnet? I find hard to envision it replacing ddnet unless it's an official switch

proper modding api, an actual modding

Tater

What does ddpg offer to players over ddnet? I find hard to envision it replacing ddnet unless it's an official switch

It can only be an official switch if we somehow find hundreds and hundreds of hours to reimplement ddrace as a wasm module with the 3 Rust developers we have around that are all so busy that they either don't contribute nowadays or only do maintainance

17:56

Also all the old ranks need to play back perfectly on the newly reimplemented ddrace, and we have to support both dd-pg based clients and old clients at least for a while because we can't just drop support for it completely overnight

zhn

proper modding api, an actual modding

This isn't a direct offer to players, it's indirect

client mods xd

Tater

What does ddpg offer to players over ddnet? I find hard to envision it replacing ddnet unless it's an official switch

Way more fps, probably more stable once it's released because Rust, and in the future when there are people around to make mods for it it's more flexible so stuff like the old nodes mod

Way more?

didn't you test it?

Surely 5000 -> 20000 is not important?

Tater

Surely 5000 -> 20000 is not important?

we have players that owe computers handling ddnet at less thab 100 fps

Tater

Surely 5000 -> 20000 is not important?

It's not imo but I've given up on that discussion long ago. I now pretend it matters

18:00

It's always better to have more frametime available anyway, you can do more

I guess but you need to do physics which is kinda fundamental limited

zhn

we have players that owe computers handling ddnet at less thab 100 fps

Surely these ancient machines can't run dd-pg anyway, it needs vulkan

Tater

I guess but you need to do physics which is kinda fundamental limited

Our physics are tied to ticks, not fps, so I guess frametime doesn't have much to do with it

Learath2

Surely these ancient machines can't run dd-pg anyway, it needs vulkan

i mean the ones with vulkan, even

zhn

i mean the ones with vulkan, even

What toaster that is new enough to support vulkan properly runs ddnet at less than 100 fps? My intel atom netbook with igpu still ran it at like 150 last year when I checked

18:05

Anyway, all of this is beyond the point. I've delivered my grand prediction. I'll go make some food

Learath2

It can only be an official switch if we somehow find hundreds and hundreds of hours to reimplement ddrace as a wasm module with the 3 Rust developers we have around that are all so busy that they either don't contribute nowadays or only do maintainance

can we remove rust from ddnet, its just annoying x-x

Solly

can we remove rust from ddnet, its just annoying x-x

It's about to get an actual use soon(tm)

(removing all the stubs from client removes 1mb of the executable and probably speeds up compile time) (edited)

https://github.com/ddnet/ddnet/pull/6961

QUIC PoC by heinrich5991 · Pull Request #6961 · ddnet/ddnet

Still missing: timeout protection works on computers other than my own? works on windows? works for cross-compiling review protocol for segmenting messages, is there some sort of standard may...

couldnt be done in cpp?

18:08

too late now not telling heinrich to redo their work xd

could be, but it isn't

sigh

Learath2

Our physics are tied to ticks, not fps, so I guess frametime doesn't have much to do with it

My point was vague sorry, I meant if you need to spend 1/1000th of a second to do physics the extra frames don't increase smoothness.

18:11

Unless you have a separate thread

Tater

My point was vague sorry, I meant if you need to spend 1/1000th of a second to do physics the extra frames don't increase smoothness.

Hm, it is technically quantized but we interpolate, so it might look smoother

Tater

This isn't a direct offer to players, it's indirect

well i'd guess the indirect benefits would be the reason people switch if they do

but my point with frametime was that you can do other stuff, you just need to update physics at the correct times

is there any way to fix this? the half of mine server is this "bots"

uncheck show connecting players

18:16

if htis is ur server then uhhhh idk

Solly

uncheck show connecting players

how

in the filter settings

18:16

turn on "filter connecting players"

Learath2

Anyway, all of this is beyond the point. I've delivered my grand prediction. I'll go make some food

bon appetite, big boy

1

Solly

turn on "filter connecting players"

huh

18:18

idk ¯\_(ツ)_/¯

Solly

idk ¯\_(ツ)_/¯

this "players" is not letting other join game

montikboom

this "players" is not letting other join game

this is called a slow loris attack

18:19

theres probably some setting somewhere to turn timeout to really low

Solly

theres probably some setting somewhere to turn timeout to really low

where?

finding

18:20

conn_timeout

18:20

but youre probably better off banning the ips

18:20

because they can just reconnect afterwards (edited)

Solly

conn_timeout

this in f2?

yeah

18:20

set it to like 10 or something

18:20

idk if its in ticks or seconds

Solly

set it to like 10 or something

i put 5

18:21

ty, this worked

cool

18:22

if its automated the bots r probably still connecting over an over again

18:22

so u should probably ban by ip

Solly

idk if its in ticks or seconds

seconds

1

why is it 100 by default then

18:23

thats maaaasive

18:23

is it for map download?

i was connecting to one brazil server for almost 2 minutes, so its gut xd

.-.

How do I iterate over the game layer using twmap and place a specific tile from a layer based on the game tile?

you mean copy game layout into tiles layout?

19:15

design layer i mean

19:16

@fokkonaut

yes

look up imilsch's exporter from gores generator

19:17

lemme send link

19:18

https://github.com/iMilchshake/gores-mapgen/blob/main/src/twmap_export.rs

gores-mapgen/src/twmap_export.rs at main · iMilchshake/gores-mapgen

Procedual random map generator for the gores gamemode in teeworlds/ddnet - iMilchshake/gores-mapgen

Thanks

19:20

How does it work for bash?

19:20

chillerdragon: maybe

19:20

m.groups[16].layers[0].color = (234, 234, 234, 255) This is how to modify color of a tile layer

19:21

cant find a doc about twmap to get and set a tile from that layer

are u using the python version?

yes

i think u have to use help(twmap.Map) and find the layers attribute

Oh, okay lemme see

if you want to modify a design layer then u need to find the index of the layer it corresponds to

19:31

if u want to extract game tile layers:

with open(map_path, 'r') as file:
        map_data = twmap.Map(map_path)
game_layer = map_data.game_layer().tiles
front_layer = map_data.front_layer().tiles
...

Learath2

Way more fps, probably more stable once it's released because Rust, and in the future when there are people around to make mods for it it's more flexible so stuff like the old nodes mod

people will have 5000 instead of 4000 fps at the small cost of full rewrite and no new updates

19:35

worth it

i agree

19:43

totally worth it

I'm the right potato

for some reason it doesnt seem to load the map into python

20:02

m = twmap.Map(sys.argv[1])

20:02

this does crash silently

20:03

map is in the same path + admin cmd

huh

l-ouis

[ddnet/ddnet] Pull request opened: #9251 Add sv_practice_in_team_0 command

Might be useful for trashmap servers or servers where owners want to allow /invincible, /telecursor, in team 0 by default instead of making players go into a team first. Not sure if it would be used beyond trashmap servers, that can be something to discuss.

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 cover...

hi again, how to get position and timer value by player id?

20:25

in server side

pls help

idk

hi again, how to get position and timer value by player id?

int CurrTime = (pSelf->Server()->Tick() - pChr->m_StartTime) / pSelf->Server()->TickSpeed(); (edited)

20:37

position is in CCharacterCore

thank you

kebs

int CurrTime = (pSelf->Server()->Tick() - pChr->m_StartTime) / pSelf->Server()->TickSpeed(); (edited)

i write in server.cpp. how to get pChr by id?(for(int i = 0; i < MAX_CLIENTS; i++))

You can't use the character data in the engine (server.cpp), check gamecontext.cpp and gamecontroller.cpp for other loops with i < MAX_CLIENTS condition

Robyt3

[ddnet/ddnet] Pull request opened: #9252 Remove unused parameters of CMapLayers functions

Use width and height of STileLayerVisuals instead of passing CMapItemLayerTilemap * separately. Remove unused CMapItemGroup * parameters.

Checklist

[X] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Written a unit test (especially base/) or added coverage to integration test
[ ] Considered possible null pointers and out of bounds array indexing
[ ] Changed n...

@Assa https://github.com/ddnet/ddnet-discordbot/issues/87 Could you clarify what the command is supposed to do? You mentioned twmap-check, but I assume you meant twmap-check-ddnet?

just have the ddnet bot check it the same way it checks skins etc - (e.g in dms) ?

21:18

or is that to heavy on the bot

murpi

@Assa https://github.com/ddnet/ddnet-discordbot/issues/87 Could you clarify what the command is supposed to do? You mentioned twmap-check, but I assume you meant twmap-check-ddnet?

yes exactly, I know that you can Download compile and run this tool, but thats not really accessible for the majority

21:41

also I did a PR for patiga in order to check tunings

21:45

fixed the description for discordbot issue (edited)

heinrich5991

[ddnet/ddnet] Pull request opened: #9253 Document that using integers in boolean contexts is not recommended

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-addresssan...

is there any kind of rule when to use references vs pointers? :p

22:00

except when null pointer can be passed (edited)

rly

22:04

do reports even do anything

22:04

same msg as yday

https://factorio.com/blog/post/fff-437

Friday Facts #437 - Cargo Pod Deep Dive | Factorio

Howdy, this week we will be diving deep into the workings of cargo pods and their seamless flight animations. Most of you playing the expansion have probably already seen them in game. The positive reception makes us very happy. Ascending to platform in orbit above Nauvis.

github-merge-queue[bot]

[ddnet:master] 2 new commits

557a594 Remove unused parameters of CMapLayers functions - Robyt3 b83bad7 Merge pull request #9252 from Robyt3/Client-CMapLayers-Parameter-Cleanup - heinrich5991

22:11

heinrich5991

[ddnet/ddnet] Pull request closed: #9252 Remove unused parameters of CMapLayers functions

22:11

github-merge-queue[bot]

[ddnet:master] 2 new commits

1cb3a6b Substantially reduce stack memory usage of CDataFileWriter - Robyt3 495e262 Merge pull request #9246 from Robyt3/Datafile-Writer-Stack-Usage - heinrich5991

22:11

heinrich5991

[ddnet/ddnet] Pull request closed: #9246 Substantially reduce stack memory usage of CDataFileWriter

kebs

same msg as yday

almost as if something you can't see must be going on with those "same" messages

def-

[ddnet/ddnet] Pull request closed: #9253 Document that using integers in boolean contexts is not recommended

22:35

github-merge-queue[bot]

[ddnet:master] 3 new commits

666d9f0 Replace GitHub-specific emoji names with emojis directly - heinrich5991 9ffde94 Document that using integers in boolean contexts is not recommended - heinrich5991 aa408ef Merge pull request #9253 from heinrich5991/pr_ddnet_contributing_if_int - def-

Learath2

almost as if something you can't see must be going on with those "same" messages

if "client.com" in message: return (edited)

22:36

3

<ChillerDragon> @fokkonaut you can iterate the game layer like this for (y, x, flags), tile in numpy.ndenumerate(m.game_layer().tiles):

i'm abusing client rn

<ChillerDragon> for setting tiles you have to reassign the tiles attribute m.groups[collision_group].layers[collision_layer].tiles = edited_collision

23:17

<ChillerDragon> full context here https://github.com/ChillerDragon/teeworlds-reverse-gametiles/blob/master/reverse-gametiles.py

teeworlds-reverse-gametiles/reverse-gametiles.py at master · Chille...

Tool for mappers that like to map with gametiles first. It generates tiles based on gametiles so the opposite of what "Game Tiles" button does in the editor. - ChillerDragon/teewo...

@rob the 5px padding is the same as on the friends list (edited)

<ChillerDragon> @fokkonaut what are you working on?

23:25

Also for keeping the background color the same thruought why not just move the background to the draw() call done ages ago

23:25

instead of relying on backgrounds drawn on backgrounds