DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.tw/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 2021-08-18 00:00:00Z and 2021-08-19 00:00:00Z

cool thing we could do is create a form that lets you enter like 5 GitHub issues you would like to see fixed and then people can review what issues are the most wanted

2

2

for those who saw my box2d video and are interested http://github.com/headshot2017/ddnet-box2d

GitHub - headshot2017/ddnet-box2d: Box2D implementation on DDNet se...

Box2D implementation on DDNet server. Contribute to headshot2017/ddnet-box2d development by creating an account on GitHub.

[ddnet-maps:master] 1 new commit

a89a2ca A Orange 2 - ddnet-maps

or u can learn cpp and fix it monkalaugh

monkalaugh

Ryozuki

or u can learn cpp and fix it monkalaugh

monkalaugh

will anyone mind if i'll replace NoHit* bools in CCharacterCore with int m_Hit?

i prefer using bools

07:18

u will only know if u make a pr and ppl discussi t

Headshot

for those who saw my box2d video and are interested http://github.com/headshot2017/ddnet-box2d

pro

gerdoe

will anyone mind if i'll replace NoHit* bools in CCharacterCore with int m_Hit?

1: Why? 2: The character core is shared across the network, have you considered compatibility?

Learath2

1: Why? 2: The character core is shared across the network, have you considered compatibility?

1 - possibility to iterate DISABLE_HIT* 2 - i don't want to change net objects, only game part

@Deleted User can we use https://github.com/google/angle ?

GitHub - google/angle: A conformant OpenGL ES implementation for Wi...

A conformant OpenGL ES implementation for Windows, Mac, Linux, iOS and Android. - GitHub - google/angle: A conformant OpenGL ES implementation for Windows, Mac, Linux, iOS and Android.

08:32

for bad graphics opengl driver problems

mh, only GLES 2.0 on dx9, that would default all users to our GL 1.x backend probably And also depends how good the dx9 backend is written But we could theoretically try

08:59

Most ppl that currently have bugs manually change the graphic settings the fallback works

@Learath2 if you are on your mac, could you do me a favour and do gfx_opengl_major 4; gfx_opengl_minor 5 restart and send me the renderer string in the F1 console

09:37

i think except on linux we can also just prevent GL 3.3 if you don't have like OpenGL 4.3 or even 4.5

Deleted User

@Learath2 if you are on your mac, could you do me a favour and do gfx_opengl_major 4; gfx_opengl_minor 5 restart and send me the renderer string in the F1 console

on master?

doesnt matter

09:38

latest version is fine too

https://paste.pr0.tips/di

ok thanks

I can create a 4.1 context btw

09:42

that's the max supported on atleast my version of macOS with my hw configuration (10.15 with an amd gpu)

09:43

s/amd/ati/

ok

well I guess radeon 555 is amd, apple just didn't rename the driver

then i'll atleast do it on windows for now

Why btw?

09:47

just a heuristic you noticed that drivers that can't do 4.3, 4.5 don't do 3.3 correctly?

yeah seems like there are drivers that are old but say they already support 3.3

09:48

probably mostly AMD

09:48

AMD 7800 or smth

09:48

or 6000 series

09:50

https://de.wikipedia.org/wiki/ATI-Radeon-HD-4000-Serie they for sure atleast

ATI-Radeon-HD-4000-Serie

Die Radeon-HD-4000-Serie ist eine Serie von Desktop-Grafikchips des Unternehmens AMD und der Nachfolger der Radeon-HD-3000-Serie. Alle Grafikprozessoren dieser Serie unterstützen das Shadermodell 4.1 (SM 4.1) nach DirectX 10.1. In dieser Serie wird zum ersten Mal GDDR5-Speicher verwendet. Der Nachfolger ist die Radeon-HD-5000-Serie.

09:50

rest isnt clear, if thats only on windows 10 or also on win7

Deleted User

yeah seems like there are drivers that are old but say they already support 3.3

But they don't support it fully?

they have bugs

09:51

i cant imagine another reason ppl change to GL 3.3 and have texture glitches

09:51

my father also has an 6000 series i think

09:51

with win7

Are you 100% sure that we are within spec of 3.3? We were using functions from the future in the 2.x backend atleast

yeah but 3.3 has renderdoc

09:53

2.x was blindly writing

How strict is renderdoc when it comes to the spec? Or the software renderer you tested on, how strict is that?

renderdoc atleast shows when you use invalid API, i tested it

09:54

ofc not all possible stuff

Invalid as in the driver says its invalid or does renderdoc do its own checking

09:55

I've encountered drivers that just let you use higher version functionality or silently fail

renderdoc has an own software renderer

I see, so barring a logic mistake in our code, the 3.3 renderer is atleast correct as far as renderdoc is correct

well my android phone works atleast xxD

09:58

if a shitty 5 years old 5 watt system can render it correctly, then the chance is atleast not too small

If only I could get my hands on such a broken system it'd be pretty fun to debug

[ddnet/ddnet] Pull request opened: #4093 Make fallback stronger and safer

Save renderer string, so the fallback isn't portable between driver updates or system changes
Only allow modern GL on modern systems(windows)
Don't show modern GL in settings, if the driver is on the blocklist(this prevents Intel users from doing it, they can still force it over F1 console)

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...

let me find that old amd card xd

ok HD 5670 and it even renders on GL 4.5 correctly under win10 the driver is from 2015

10:29

either under win7 it was worse and they fixed the drivers for win10 or i had a different problem last time i tested xd

10:30

but the wikipedia article is wrong then

10:30

it says 5000 series only support GL 4.4

MINIMUM:
OS: Microsoft® Windows® XP or newer

Someone with steam access should update this XP isnt supported and probably won't anymore

HTTPS and discord api

for xp we can use old serverbrowser and disable discord api

18:21

not sure what the current state is and if it starts on xp at all. maybe no one cares anymore

it doesnt start

ok, probably not worth making a build without discord api

image.png

Image attachment

yeah probably not ^^

lol

then we can remove the ifdefs in code for xp too

yeah sounds good

@ChillerDragon find -delete or sth lile that

rm test_1* did the job

18:23

:D

Someone recommends some c++ gui (other than GTK+) for linux?

im just suprised never saw that

18:23

qt5

or ls -print0|xargs -0 rm etc.

ah ye

18:25

i just crashed my client by loading 100k skins :D

chillerdragon

qt5

Ty, I'll check that

chillerdragon: really? but memory limit?

i had cases where i had to increase the stack size so i can pass more files to a program, was multiple MB of params

wtf

llvm-cov wanted to have every path and our source code at work is huge

impressive

18:29

few mb of paths as code

monkaS

must be more than 50k files

also historically grown long parhs

ah ok

18:30

i calculated with 7mb stack and 128 bytes per path

only some part of stack is reserved for params i think

ah makes sense

[ddnet/ddnet] Issue closed: #4092 Map not showing up as already finished in server browser

22:48

[ddnet/ddnet] Issue closed: #1689 News tab reads uninitialized values in the start

[ddnet/ddnet] Pull request opened: #4094 Add prediction for switch tiles

Prediction for freeze and other tiles in switch layer.

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 if it works standalone, system.c especially
[ ] 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](htt...

duplicate

Exported 96 message(s)