Teeworlds - development

[teeworlds/teeworlds] Issue opened: #2789 RFC: Official Race and its requirements

I recently talked with @Zatline about finally adding a Race gamemode. Basically there are two variants of doing this: a very simple round based gametype like proposed here #33 or something ladder-based, where you just aim for the fastest time. While the first approach fits better with the simplistic design of the game, Race in Teeworlds has been associated with a ladder-based system for over a decade now. And just to be realistic: a round-based game mode is only fun if there are enough play...

@redix Was talking about this with gfx ppl. I tried to do it. But I have no idea what I'm doing with opengl. https://github.com/teeworlds/teeworlds/issues/2157 Any tips?

Auto adaptable body markings · Issue #2157 · teeworlds/teeworlds

Right now we're using markings that are based on the default tee body shape, which on different body types might look weird and not that good. My suggestion is to make the markings auto ada...

02:56

~~Or could you do this really quick.~~

That should be possible with the stencil buffer. Never used it myself though. That probably requires some bigger changes to the rendering engine =\

~~Is stencil a boolean buffer?~~ (edited)

03:00

Nvm

03:01

Was thinking about how to preserve the ear "texture".

03:02

I'll try to learn about stencil buffer.

Problem might be that you need a mask for each body then =\

I think the body texture itself might already be a good mask

axblk

[teeworlds/teeworlds] Issue opened: #2790 RFC: Layer and Tile System

Right now we can only have a single game layer per map. This layer only has a limited amount of tiles and does not support per-tile-parameters. Due to the limited amount of tiles, mods might conflict with future vanilla tiles. The lack of per-tile-parameters (for example a teleporter number or speedup direction) requires custom layers in many mods. This leads to a lot of additional code, incompatibilities and makes it difficult to upgrade code and maps to new Teeworlds versions. One approa...

fokkonaut

[teeworlds/teeworlds] Issue closed: #2788 New tile system

brain wave collision

btw @redix what do you think are the main benefits of an official race server mod?

1vs1 would be funny xd

15:57

make it competitive xd

quickly made a shittier version of the cammo

16:14

16:20

16:21

ok not all markings look good. (edited)

16:22

toptri_fox is kinda stupid.

@Dune less work for me when there is a new version :D In the end I don't really care... several other people finally want to have official race. A round-based gamemode probably would not be that successful with today's player-count =\

=\

@redix agree yeah

The bigger benefit might be the layer system that other mods could use. Question is whether they will... Thanks to the bridges people are more concerned about ddnet compatibility.

16:28

For now it is rather my general thoughts about all that. Until next year I probably don't have enough time to work on it anyway

(edited)

16:36

what happened to the macos build

gotta check the log

i fails every time now

it does break maps though, so it'd need to be implemented with a bunch of editor features or a third party tool and lots of support

i totally support race gamemode with a ladder system @redix, i guess once sqlite is implemented it would be possible to use for vanilla too right?

personally i think it might be clunky to implement. and modders might complain more since now they have sql queries scattered around to deal with if we didn't do it "right".

16:44

might

Yeah there should be a proper conversion tool. Most user friendly would be in the editor, but that would bloat the code too much. A command line tool would be the easiest, but not too user friendly. A tiny website where you upload your old maps might work. With webassembly it could even be realized without a backend server.

i already opened an issue @Stiopa (edited)

16:49

https://github.com/teeworlds/teeworlds/issues/2786

Pipeline for MacOS is broken · Issue #2786 · teeworlds/teeworlds

See #2784

to be honest, just because we have sqlite, that does not mean that it's super easy to use for other mods. The queries need to be threaded and what I have right now is tailored for storing times

The bigger benefit might be the layer system that other mods could use. Question is whether they will... Thanks to the bridges people are more concerned about ddnet compatibility.

@redix I would, so at least half the mods in 0.8 will

I would commit time into porting ddnet to 0.7 or 0.8 if we first gave teeworlds some mod friendly features

hasn't that been happening for a year now?

Eh sure, but I mean some large extensibility changes

not enough i suppose

someone has to code those :/

uuids for datafile item types, uuids for layer types, NETMSG_EX, NETOBJ_EX, IGameController becoming an abstract base class, CGameContext being cleaned up of all gamemode dependence

17:02

Exactly, development time is scarce and most people don't want to invest it into features that might not even be accepted or would take a lot of time

that might not even be accepted

just make an issue saying what you want and wait for oy to tag it?

17:03

I don't write code for stuff that I'm not sure will be accepted either

@Dune one day when I have time I should probably create an issue for each to atleast gauge what oy thinks about it

in my experience oy is always more positive than I would imagine

Exactly, development time is scarce and most people don't want to invest it into features that might not even be accepted or would take a lot of time

@Learath2 or aren't even used when accepted ._.

Well a ddnet rewrite to 0.7/0.8 would definitely use all of the above

@Assa which one are you talking about? all your PRs were accepted afaict

Well I'm guessing one of them was accepted but not used

i mean in general, I did a lot of mods and most of them didn't get any recognition. Developing something and getting nothing out of it is really demotivating, but I think that's how it's always. Even at work ~half the Projects are done and never used

my experience for now is fairly positive overall

17:12

but i haven't done much

17:12

my experience for now is fairly positive overall

17:12

^ out of order messages

I think in teeworlds it's partly my fault, because I never finish projects when I already accomplished the hard parts xD

right, i've been keeping up with #2704 until recently.

17:16

it is getting a little bit tiring

I can imagine. I'd never embark for such a long journey!

keep it going, the updates are refreshing ^^

but we haven't get much update in master which is sorta nice but not really nice at the same time

17:22

i basically asked oy to consider merging this early in a comment.

17:22

heard nothing back yet lol

it's a lot of painful work to review something this large tbh

i know

feelstiringman

@TsFreddie for the adaptable marking, maybe a skin json file should have an option to allow adaptive markings

19:30

For the cat it also looks stupid

I would say we need a templated message queue in order for it to be easily usble. main tick method would try to retrieve data from the queue if a mutex is in the main thread and other threads would try to get data from the db and block until the message queue mutex is unlocked. then it would help

19:36

for sqlite communication without blocking the main thread

fokkonaut

[teeworlds/teeworlds] Pull request opened: #2791 Don't relay on gamestate for the tick symbol (closes #2676)