Voxel 2022-03-30 01:09:09Z

def-

[ddnet-web:pr-16.0] 1 new commit

723627e Version 16.0 - def-

07:41

def-

[ddnet/ddnet] Pull request closed: #4863 Version 16.0

bors[bot]

[ddnet:master] 2 new commits

3bcb49f Vk only allow asserts after initialization - Jupeyy af0b979 Merge #4907 - bors[bot]

07:56

bors[bot]

[ddnet/ddnet] Issue closed: #4891 Switching to vulkan and restarting will crash

07:56

bors[bot]

[ddnet/ddnet] Pull request closed: #4907 Vk only allow asserts after initialization

@deen steam release candidate branch is empty again

08:29

was the upload killed again?

def-

[ddnet/ddnet-web] Pull request closed: #181 Version 16.0

08:33

def-

[ddnet-web:master] 2 new commits

723627e Version 16.0 - def- cc7b350 Merge pull request #181 from ddnet/pr-16.0 - def-

@Jupstar ✪ does vulkan need any libraries installed on linux?

08:47

im gonna update aur

vulkan loader

i mean to compile

08:48

not run

just see readme

08:48

vulkan headers and loader i think

08:49

sudo pacman -S --needed base-devel cmake curl freetype2 git glew gmock libnotify opusfile python sdl2 sqlite wavpack x264 ffmpeg vulkan-icd-loader vulkan-headers glslang spirv-tools

08:49

xd

vulkan-icd-loader vulkan-headers

08:49

true

08:49

xd

glslang spirv-tools are also important

08:49

even for compiling

@Jupstar ✪ and do u need those 4 after compiling?

the loader yes

08:52

rest no

ok

08:55

https://aur.archlinux.org/packages/ddnet

08:55

fast af

arch the fastest in the west

def-

[ddnet-scripts:master] 1 new commit

736a5a1 Turn off videorecorder on win32 - def-

@Ryozuki doesn't build for me on arch:

/home/deen/.cache/yay/ddnet/src/DDNet-16.0/src/engine/client/video.cpp:745:85: error: ‘AV_CH_LAYOUT_STEREO’ was not declared in this scope

09:47

I guess because I enabled video recorder by default

true

09:49

but there is already a dep on ffmpeg

09:49

hmm

Or maybe just an include missing?

it works for me

09:50

xD

I have ffmpeg4.4 and ffmpeg installed on this machine

09:50

maybe that's related

09:50

and I just cleaned my cache dir, let's see if it reproduces from scratch

if it doesnt work, try installing ffms2

09:51

09:51

maybe its that

Full output

message.txt

8.77 KB

idk why it happens

09:57

did installing ffms2 fix it?

def-

[ddnet/ddnet] Pull request opened: #4914 Fix video compilation on Arch

/home/deen/git/ddnet/src/engine/client/video.cpp:842:37: error: ‘AV_CH_LAYOUT_STEREO’ was not declared in this scope

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
...

Didn't try, ffms2 looks unrelated. Including the header works

:o

09:58

its weird cuz i dont have that error

09:58

c++ funny stuff

maybe its simply a header not included

I'm guessing channel_layout.h gets included implicitly on your system somehow, but not on mine

09:59

we could run the preprocessor over the file and compare outputs

how to do that

introduce a compilation error in the file, take the build command, add a -E at the end, then look into the -o file

10:00

(and be shocked that we include 100k lines of headers :D)

10:00

u need to fix the style

10:00

on ur pr

let's try to understand the reason first

10:01

video.cpp.preprocessed

2.07 MB

where do i put the error

anywhere in src/engine/client/video.cpp

10:05

this? xD

10:05

im lost

yes

10:06

but change the path of the -o

video.cpp.preprocessed

135.9 KB

10:09

hope i got it right

10:10

ok i guess smth is wrong since urs is 2mb

I think I told you the wrong file

10:10

src/engine/client/video.cpp, my bad

oh i used shared

10:11

video.cpp.preprocessed

2.07 MB

10:13

the diff is big cuz all paths are different

replace ryozuki with deen xd

you have an old file state

10:13

git pull first

true

This error and the AV_CH_LAYOUT_STEREO came in from vulkan change

i think it was there before

Well, in Ryozuki's file there is no mention of AV_CH_LAYOUT_STEREO at all

AllocAudioFrame(AV_SAMPLE_FMT_S16,

747 "/home/ryozuki/Documents/ddnet/src/engine/client/video.cpp" 3 4

(0x00000001|0x00000002)

10:15

it was replaced

10:15

for some reason

oooh, right, preprocessor ran

video.cpp.preprocessed

2.07 MB

10:17

im on af0b9790d

aaaah, I see why

:o explain

swresample.h includes channel_layout.h

10:18

But I seem to have some old self-build ffmpeg version in /usr/local/include: 1 "/usr/local/include/libswresample/swresample.h"

aaah

which doesn't include it, so my fault and PR probably not necessary

nice mr detective

1

def-

[ddnet/ddnet] Pull request closed: #4914 Fix video compilation on Arch

I have started our last Steam Visibility Round

:O

what is your 100 character description xd

nice

10:21

make more clickbait ddnet now renders with 1 trillion fps on any NASA computer, and all for free xd

3

we could try out https://github.com/include-what-you-use/include-what-you-use

GitHub - include-what-you-use/include-what-you-use: A tool for use ...

A tool for use with clang to analyze #includes in C and C++ source files - GitHub - include-what-you-use/include-what-you-use: A tool for use with clang to analyze #includes in C and C++ source files

I have some experience with this tool from work

i think i used it on ddnet, but was to lazy to actually test it

There is a more modern solution, but haven't used it yet: https://clang.llvm.org/extra/clang-include-fixer.html

10:24

ah no, that was the one for adding, wait a minute

Some bullet points in changelog are missing [author]. Are they [Deen] implicit?

yes

11:39

@deen do u know if many players joined tutorial server

11:39

currently 28 players

11:39

on them

Tutorial is on special server? Not offline?

Yep, just search for Tutorial in DDNet tab

Jupstar ✪

we could try out https://github.com/include-what-you-use/include-what-you-use

tried it out: 243 files changed, 1742 insertions(+), 656 deletions(-) xd, would need to benchmark if its worth it

Jupstar ✪

mh weird stuff

xD I just installed the 16.0-1 version from the AUR

there it happens again

what was the fix last time?

12:30

just rebuild?

12:30

or switch to master branch?

12:30

did we not include some commit or what

12:31

@Ryozuki can u update aur ddnet-git too?

12:31

then we can compare xd

ddnet-git doesnt need updating

12:32

well maybe the dependencies

12:32

but it automatically uses the last commit

oh ok

12:32

then its just the title "ddnet-git 15.5.1" XD

one moment I will test ddnet-git

def-

[ddnet/ddnet] Pull request opened: #4915 Upgrade zlib to 1.2.12

"Fix a deflate bug when using the Z_FIXED strategy that can result in out-of-bound accesses." sounds security relevant.

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...

@Jupstar ✪ ddnet from aur is also build by urself

12:33

but just from the tagger version

yeah

tagged

when do we get an upgrade

12:33

i bet ddnet is one if not the most played open source game rn

it can be done

upgrade to arch pages

if ppl vote in the ddnet aur package

packages*

i wonder who would maintain it tho

12:34

it probs wouldnt get updated this fast

12:34

By the way I find the message in the terminal a bit alarming " [2022-03-30 14:32:56][storage]: failed to remove: /home/daniel/.teeworlds/" xD

Ryozuki

@deen do u know if many players joined tutorial server

it's still one of the most finished maps

c0d3d3v

By the way I find the message in the terminal a bit alarming " [2022-03-30 14:32:56][storage]: failed to remove: /home/daniel/.teeworlds/" xD

yeah seen it too

c0d3d3v

By the way I find the message in the terminal a bit alarming " [2022-03-30 14:32:56][storage]: failed to remove: /home/daniel/.teeworlds/" xD

that's harmless i think, had that for a while

Jupstar ✪

i bet ddnet is one if not the most played open source game rn

really?

not the most

well what games do you know that have 2k active players

12:36

0ad is huge too

there is a gpl game

12:36

in steam

I think I stated this once before and heinrich had 2 examples

i dont remember the name

12:36

about tower defense

but most games arent really activly played

There's a list here, someone could check steamdb with https://github.com/Poussinou/FLOSS-Games-on-Steam

GitHub - Poussinou/FLOSS-Games-on-Steam: A list of FLOSS games avai...

A list of FLOSS games available on Steam. Contribute to Poussinou/FLOSS-Games-on-Steam development by creating an account on GitHub.

they might have a bigger total count

https://store.steampowered.com/app/1127400/Mindustry/

Mindustry

Gameplay Create elaborate supply chains of conveyor belts to feed ammo into your turrets and produce materials to use for building Use production blocks to create a wide variety of advanced materials Defend your structures from waves of enemies Play with your friends in cross-platform multiplayer co-op games, or challenge them in team-based PvP ...

Price

$5.99

Recommendations

9613

12:36

this game is GPL

Some factory logistics thing and maybe openarena?

open source

mindustry is open source?

12:36

ah yeah

12:36

right

ye

he just takes 5€ on steam

u pay but its GPL

but still

12:37

400ppl ingame

12:37

much less than ddnet

true

see community hub

xd

I get on both DDNet and DDNet-git the following error, if I have vulkan active and only start DDNet and close it (takes also long to close): /usr/include/c++/11.2.0/bits/stl_vector.h:1045: std::vector<_Tp, _Alloc>::reference std::vector<_Tp, _Alloc>::operator[](std::vector<_Tp, _Alloc>::size_type) [with _Tp = std::vector<CCommandProcessorFragment_Vulkan::SRenderCommandExecuteBuffer>; _Alloc = std::allocator<std::vector<CCommandProcessorFragment_Vulkan::SRenderCommandExecuteBuffer> >; std::vector<_Tp, _Alloc>::reference = std::vector<CCommandProcessorFragment_Vulkan::SRenderCommandExecuteBuffer>&; std::vector<_Tp, _Alloc>::size_type = long unsigned int]: Assertion '__n < this->size()' failed. [1] 258581 IOT instruction (core dumped) DDNet I can try to debug it

a stacktrace would also already be nice

How can I create it?

Load the core in gdb with the executable and do bt

i can never find my cores

12:40

i guess i never configured it

12:40

xd

dmesg should tell you where

12:40

systemd puts them in some special dir maybe

12:41

but I have the same problem on macOS with arm. Apparently Apple doesn't allow coredumps for some applications and I can't figure it out

i create my cores in place xD

https://wiki.archlinux.org/title/Core_dump

Core dump

@Jupstar ✪ why does the text size thing happen on gl1 btw?

bcs it doesnt use a texture overlay

12:48

but the actual text renderer

Jupeyy

[ddnet/ddnet] Pull request opened: #4916 Try default opengl settings first, when vulkan fails

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](https://github.com/ddnet/ddnet/#using-addresssanitizer--u...

12:59

def-

[ddnet:master] 1 new commit

3d67ccf Version 16.0 - def-

def-

[ddnet/ddnet] Pull request opened: #4917 Use std::size instead of sizeof(a) / sizeof(a[0])

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 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](https://github.com/ddnet/ddnet/#using-addresssanitizer--u...

13:21

def-

[ddnet-web:master] 1 new commit

69573f6 Update settings & commands - def-

bors[bot]

[ddnet:master] 2 new commits

0d8338f Try default opengl settings first, when vulkan fails - Jupeyy 8cd78a8 Merge #4916 - bors[bot]

13:37

bors[bot]

[ddnet/ddnet] Pull request closed: #4916 Try default opengl settings first, when vulkan fails

def-

[ddnet/ddnet] Issue opened: #4918 Can't control client after exiting demo renderer

@Jupeyy might be caused by your recent change? Mouse is not moving anymore, after I alt tab away and back it seems to work again.

@Jupstar ✪ the problem does not occur in any case when I build myself master, with the following debug options: CC=clang CXX=clang++ CXXFLAGS="-fsanitize=address,undefined -fsanitize-recover=address,undefined -fno-omit-frame-pointer" CFLAGS="-fsanitize=address,undefined -fsanitize-recover=address,undefined -fno-omit-frame-pointer" cmake -DCMAKE_BUILD_TYPE=Debug .. I'll try to get the stacktrace from the version in the AUR.

weird stuff going on

13:58

you said it happens on alt f4?

13:59

does it also happen if you lcikc the menu exit button?

yes

ok

can i just give you some core dumps? i get the stack traces out but they don't look useful,

i'd need your DDNet executable too then

the first one is with alt f4 the second with exit button

DDNet

3.58 MB

core.DDNet.1001.a285a523ac49418194a2d2d8a806a006.275000.1648648551000000.zst

32.85 MB

core.DDNet.1001.a285a523ac49418194a2d2d8a806a006.276011.1648648805000000.zst

33.01 MB

deen

https://wiki.archlinux.org/title/Core_dump

xD coredumpctl is pretty nice, I did not know that yet

14:05

ah but the DDNet executable has no debbuging symbols

does it happen if you build in RelWithDebInfo from master

@deen @Jupstar ✪ what happened in this update again?!

14:10

16.0 zooming out laggs ass

14:11

compare the version before to 16.0

14:11

i noticed when zooming out on the tutorial map

14:11

but u can also see the difference on my server cuz thats where i know it didnt lag this hard xD

why u tag me, i never change network code

its not network? xd

what else

rendering? xd

havent changed anything rendering related

then anything else gl related

14:13

or sdl

14:13

idk

i especially didnt change the GL backend

ur usually the guy who is connected with these changes

i only added the vulkan backend

fokkonaut

@deen @Jupstar ✪ what happened in this update again?!

16.0 vs 16.0-rc2 or what are you comparing?

i compare last relase

14:13

15.x or so

14:14

it got waaaaaay worse

14:14

opengl 3

maybe your pc is just lagging xD

no u

can i reproduce it without going on ur server?

what opengl version is being shown in graphics settings?

wtf, with Vulkan i get almost 3k fps btw compared to a bit over 1k with opengl

yeah, that's expected, say thanks to Jupstar

ok switched back to opengl

I can't reproduce the lags with opengl 3, but I get them with opengl 1 (as always), so maybe you somehow switched to that?

3 or 3.3

Jupstar ✪

does it happen if you build in RelWithDebInfo from master

yes it also happens. one moment I send it

i double checked gfx_gl_major was 3 and minor patch was 0

14:17

i can try selecting again with the menu

14:17

which one was it before? the good one? 3.3

3.0 is worse than 3.3 in zooming out yes

14:17

bcs it creates the border rendering on the CPU, while 3.3 leaves it up to the driver

ok ye 3.0.0 sucks ass

so might run on the GPU

trying 3.3 now

14:18

yeee

14:18

thats how it was before nice, but why did it change upon update

DDNet

3.28 MB

core.DDNet.1001.a285a523ac49418194a2d2d8a806a006.292716.1648649821000000.zst

32.49 MB

i renamed the config

ah

probs resetted your config, but i wanted to remove opengl from all config names

just use vulkan

14:19

is there any difference in 3.3 and vulkan?

vulkan is more modern

like, why is it so much better

14:19

ah

and will probs be the most maintained one

vulkan is more low level

so more control

fokkonaut

i compare last relase

so when i have both versions on side by side it feels much smoother in the new one. But could be subjective and I don't have a gaming pc just a slow intel cpu.

better multithreading too

14:19

iirc

14:19

xd

c0d3d3v

so when i have both versions on side by side it feels much smoother in the new one. But could be subjective and I don't have a gaming pc just a slow intel cpu.

problem solved already

it has multi threading support

14:20

thats the difference xD

14:20

opengl can have it over extensions

14:20

but its not builtin

@Jupstar ✪ when I start the game it defaults to OpenGL 1.5.0 for me, is that expected?

on macos?

yes, I guess because OpenGL 3.0 is not supported

i think you had that before

14:21

yes

14:21

macos is special case

hm, slightly annoying, could we try 3.3 instead first?

ok nvm

well i have nothing against it, on Windows it will only allow 3.3 if opengl 4.5 is supported too

in vulkan the hookline looks a bit odd

14:21

4k fps tho feelsamazingman

14:22

zooming out BlmapChill with all objects being sent etc leaves me with 250 fps on vulkan

14:22

260

deen

hm, slightly annoying, could we try 3.3 instead first?

macos was the reason we didnt default to it in first place tho, bcs learath had a bug there later the windows bugs joined xd

oh right

fokkonaut

zooming out BlmapChill with all objects being sent etc leaves me with 250 fps on vulkan

is that good? xd

But I usually use 3.3 on macOS, the only alternative is 1.5

Jupstar ✪

is that good? xd

with gl3.3 i have 130 when zooming out entirely on it, and around 1300 in normal zoom state (edited)

c0d3d3v

Click to see attachment 🖼️

looks stripped

can it be that debugging symbols are removed after installation

possible dunno

14:25

but it is still in your build dir right?

14:25

just send that one

14:25

stripping doesnt change assembly code

14:25

normally xd

help someone finished tutorial map for me

14:26

delete rank

i finished fly out of hell for you

14:26

now its out

bors[bot]

[ddnet:master] 2 new commits

c648793 Upgrade zlib to 1.2.12 - def- ee5652c Merge #4915 - bors[bot]

14:26

bors[bot]

[ddnet/ddnet] Pull request closed: #4915 Upgrade zlib to 1.2.12

14:28

def-

[ddnet/ddnet] Issue closed: #4918 Can't control client after exiting demo renderer

Jupstar ✪

but it is still in your build dir right?

xD nope yay deleted that after ziping it wait I build it myself without yay using the modified PKGBUILD

@Ryozuki in aur/ddnet-git how exactly do versions work? I noticed that it never really autoupdates when running yay

thats cuz u didnt configure yay for it

14:31

u need to update git db or smth

pkgver=15.5.1.r4.g32523b1ea 15.5.1 is the ddnet version i assume and 2523b1ea

14:31

is a git commit

14:31

but what does r4.g stand for?

yay -Y --gendb

14:31

do this once

14:31

to register ur -git packages

14:31

now everytime u do yay

i see

it will check

14:31

Use yay -Y --gendb to generate a development package database for *-git packages that were installed without yay. This command should only be run once.

i actually like that it does not rebuild everytime

it only rebuilds on new commits

14:32

and if u accept

14:32

lol

yea but decline is effort and new commits are there everyday

14:32

but ill keep in mind if i want to

wait

14:32

maybe that doesnt work

im currntly trying to understand what version i would set for my -git package i try to craft

and u need to do yay -Syu --devel

14:32

u can save this by doing yay -Syu --devel --save

14:33

to make it the default

14:33

https://github.com/Jguer/yay

14:33

just read this nobo

so what does the g stand for in your pkgver ?

14:33

pkgver=15.5.1.r4.g32523b1ea

commit

Jupstar ✪

we could try out https://github.com/include-what-you-use/include-what-you-use

saves 1 second in single threaded build, totally not worth it xD tho i have a m2 ssd and files might have been prewarmed, maybe worth more on slower computers

but u can ignore that

14:34

on git packages

14:34

the pkgver is generated

14:34

u can check the function

14:34

in the pkgbuild

who generates it?

14:35

aaa

14:35

i see

check the fukin pkgbuild

yes yes

so there is a pkgver var and a pkgver func

14:35

weird

all these new icons aaaaaaaa

fokkonaut

help someone finished tutorial map for me

i just finished faster troll

ez

haha is 16.0 the first rls with icons?

no

14:36

i mean

it just updated them

ryo designer

editor finishes news verified tick etc#

i did nothing

was me

14:36

next replace all remaining icons xd

jupjapjaster

14:37

should be fokkonaut {verified} server

use vulkan bro

14:37

literally all advantages

u

chillerdragon {verified influencer}

Ryozuki

literally all advantages

hookline is ugly

xaxaxax

enable ffa

where

fsaa

ffs

no

it causes delay

14:38

real pro

fpga

it causes no delay

14:38

if u have 4k fps

it can tho

see

nah

then i play with 1300 fps

14:38

is also fine

depends if his hardware can do it inside the GPU

14:38

else copy back to cpu do there, copy to monitor

@Jupstar ✪ hookline is invisible

14:38

if u do

14:38

-----------

xd

14:39

xd

for me it doesnt work on vulkan, only with the software renderer xd

14:39

RADV too new for my GPU

lol nobo

14:39

i use radv

i dont use it anyway

6600 xt gang

next year i might buy a new GPU

Ryozuki

Click to see attachment 🖼️

cancel vulkan

14:40

@Jupstar ✪ make a hookline shader

u can increase hookline size now

add raytracing

someone added it

true

14:40

add lights

cl_hook_coll_size

path tracing could be cool

14:40

but would also require map file format changes

i had it at 0 lol

so its clear where the light has to bounce

cl_hook_coll_size was 0

maybe its cuz that

0 means normal mode

0 is default, yes

well with 1 when ---

14:41

is like when 0

14:41

im sure its a bug

xd

on vulkan if u do it angled a bit it looks like the hook coll is made out of multiple shorter lines which arent connected properly

is it because i have fsaa?

yes

14:42

it will interpolate next pixels

14:42

and if there is 1 pixel and sourounded by no pixels

14:42

it will kill that 1 pixel xD

xd

maybe its visible very vaguely (edited)

14:43

i dunno, probs depends too much on the impl of fsaa

https://twitter.com/icculus/status/1509161793242996746

Ryan C. Gordon (@icculus)

Would it work if someone wrote a parser that turned C code into English and then people used that English description to manually write clean-room implementations of the original code?

14:44

interesting take from a copyright perspective xd

14:45

I assume you asked Twitter for the not-a-lawyer guesses, but that sounds too much like it's still defining the copyrightable actual implementation, not the requirements specification.

>

Otherwise it'd be like license-shifting things by ROT13ing them and back, just with more toil.

14:45

(thats the sdl dev btw)

lol you have to set it in /etc/makepkg.conf in OPTIONS that you do not want to strip the symbols -.-

cant u like just build it from master branch xd

14:47

and not use install

I did

14:48

I told you

14:48

There it does not happen

ah thought it just didnt happen in debug mode

14:48

can you just use git check 16.0 maybe its a missing commit

14:49

https://ddnet.tw/downloads/DDNet-16.0.tar.xz

14:50

the aur package uses this

14:50

looks better

is that debug mode?

no core dump

i dont understand why it tells that its a recursion exception

14:51

where is the recursion its talking abotu xD

DDNet

27.11 MB

core.DDNet.1001.a285a523ac49418194a2d2d8a806a006.309363.1648651756000000.zst

32.34 MB

ok will see if i find smth thanks

I think that is a plugin error of my gdb

The recursion exception is the gdb plugin crapping out

can u do info threads

that is while in debug:

sadly gdb seems to stupid to just read the symbols from the DDNet binary :/

ah ok

15:02

@c0d3d3v in backend_vulkan in Cmd_PostShutdown there are two clear commands, can u put them after m_RenderThreads.clear()

15:03

that kinda looks wrong to me

15:06

i wonder how asan didnt detect this if thats the problem

15:06

kinda looks obvious

15:07

on the other side, maybe its too luck based

It crashs here I think because you cleared the list in line 7012 but in the other thread it is still accessed in line 7069

15:09

on the join*

yes thats why i am asking

Jupstar ✪

@c0d3d3v in backend_vulkan in Cmd_PostShutdown there are two clear commands, can u put them after m_RenderThreads.clear()

if this fixes it

ah

15:10

xD

15:10

one momenet

Jupeyy

[ddnet/ddnet] Pull request opened: #4919 Clear command list & helper after the threads joined

thanks to @C0D3D3V for finding this

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](https://github.com/ddn...

can it actually be that for some reason the new ddnet version opens a web page and that is "https://master1.ddnet.tw/ddnet/15/servers.json" xD somehow the page opened a few times when testing and I do not know why.

it downloads it

15:23

im not aware that it opens it tho

15:23

maybe the debugger parses it?

15:23

and just openens it xD

Jupstar ✪

@c0d3d3v in backend_vulkan in Cmd_PostShutdown there are two clear commands, can u put them after m_RenderThreads.clear()

so i tried to build it myself several times now but the bug only comes when i use the PKGBUILD file to build it, even if i use the source code that is downloaded there. Very very strange. But I think that will have been the problem. So I have to wait till the PR is merged xD

sometimes code layout can change runtime on the CPU

15:27

but i'd say the chance is 99.99% that this was the cause

Jupstar ✪

and just openens it xD

ah it is printed in the terminal, maybe I clicked it multiple times accidentally

I have constantly closed ddnet when it was just opened, thats why the link was always well in the clickable area,

^^

man, i cant read all that, i counted like 500 messages

Jupeyy

[ddnet/ddnet] Pull request opened: #4920 Rename 3D analysis config variable, so it resets, bcs the gl major/mi…

…nor configs were renamed

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](https://github.com/ddnet/ddn...

Chairn

man, i cant read all that, i counted like 500 messages

release day

15:41

but to sum up:

fokkonaut mad
ryo has AUR package up to date
codedev found a bug

who r u calling mad

xd

i am about to get real mad over that

15:41

this is sparta

def-

[ddnet/ddnet] Pull request opened: #4921 Revert "Update players.cpp"

This reverts commit ccc23161f639bab498fcdbe9f2cd14d0ceffa89d. Caused inaccuracies. I tried checking what calculations the server does, but couldn't figure it out quickly. @sjrc6 Reported by Zeral: https://user-images.githubusercontent.com/2335377/160875760-e639c752-d2e1-4ace-b98f-772dce689186.mp4

Checklist

[ ] Tested the change ingame
[ ] Provided screenshots if it is a visual change
[ ] Tested in combination with possibly related configuration options
[ ] Wri...

Jupstar ✪

tried it out: 243 files changed, 1742 insertions(+), 656 deletions(-) xd, would need to benchmark if its worth it

might be a good addition to my extern pr https://github.com/ddnet/ddnet/pull/4854

Chairn

might be a good addition to my extern pr https://github.com/ddnet/ddnet/pull/4854

i later found out it made 1 seconds difference in build time (single threaded)

15:47

so doesnt sound too good, considereing i manually had to fix some bugs

Jupeyy

[ddnet/ddnet] Issue reopened: #4207

Weapon of Ghost is not transparent

15:52

bors[bot]

[ddnet:master] 2 new commits

fe530ae Clear command list & helper after the threads joined - Jupeyy 213f721 Merge #4919 - bors[bot]

15:52

bors[bot]

[ddnet/ddnet] Pull request closed: #4919 Clear command list & helper after the threads joined

1

Ham5terzilla (Макс)

Click to see attachment 🖼️

thanks, fixing

Is this smallest bug ever? xd

xd

well, it's one character so kinda, but it could be worse if it was a whitespace error (clang-format commit)

before integrating format toll there was chillerdragon with his hands xd

16:24

tool*

@Jupstar ✪

16:26

16:27

i still don't understand why this only happens when building with pkgbuild

probs some hardening we dont use

16:28

@deen if u havent done the patch already

16:28

give me 10min

Jupeyy

[ddnet/ddnet] Pull request opened: #4922 Reset thread count too, so the implementention knows its not multi th…

…readed at this point anymore

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](https://github.com/ddnet...

c0d3d3v

@Jupstar ✪

thanks for quick testing

16:32

its basically the same error just in a different call xD

def-

[ddnet:16.0.1] 1 new commit

5bc4a62 Reset thread count too, so the implementention knows its not multi threaded at this point anymore - Jupeyy

c0d3d3v

i still don't understand why this only happens when building with pkgbuild

you could look what flags the compiler ueses and see what your manual build uses (edited)

-fstack-protector-strong -D_FORTIFY_SOURCE=2 wiull be the cause

16:39

CPPFLAGS="-D_FORTIFY_SOURCE=2"
CFLAGS="-march=native -mtune=native -O3 -pipe -fstack-protector-strong"
CXXFLAGS="-march=native -mtune=native -O3 -pipe -fstack-protector-strong"
LDFLAGS="-Wl,-O3,--sort-common,--as-needed,-z,relro"

we use _FORTIFY_SOURCE but no stack protector

16:40

-z,relro is also security stuff i think

16:40

o3 could also cause it xd

ah I see makepkg uses the flags in the makepkg config so one of these probably

bors[bot]

[ddnet:master] 2 new commits

2eebb48 Rename 3D analysis config variable, so it resets, bcs the gl major/minor configs were renamed - Jupeyy 8a942a2 Merge #4920 - bors[bot]

16:47

bors[bot]

[ddnet/ddnet] Pull request closed: #4920 Rename 3D analysis config variable, so it resets, bcs the gl major/mi…

@TsFreddie if your desktop resolution is lower than the monitor supports this check fails i think: https://github.com/ddnet/ddnet/blob/8a942a21dd31e9889a1028b22a8923eedc7018fc/src/engine/client/backend_sdl.cpp#L998 I'll just add a !SupportedResolution to it, is that ok?

ddnet/backend_sdl.cpp at 8a942a21dd31e9889a1028b22a8923eedc7018fc ·...

DDraceNetwork, a cooperative racing mod of Teeworlds - ddnet/backend_sdl.cpp at 8a942a21dd31e9889a1028b22a8923eedc7018fc · ddnet/ddnet

sure

Jupeyy

[ddnet/ddnet] Pull request opened: #4923 Check if the resolution is "really" not supported before applying des…

…ktop resoltution probably fixes his problems https://steamcommunity.com/app/412220/discussions/0/3176733964137910697/

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

...

17:08

bors[bot]

[ddnet:master] 3 new commits

36bfe24 Revert "Update players.cpp" - def- 0d4904f Reset thread count too, so the implementention knows its not multi threaded at this point anymore - Jupeyy 5fbef34 Merge #4921 #4922 - bors[bot]

17:08

bors[bot]

[ddnet/ddnet] Pull request closed: #4921 Revert "Hook collision calculation each frame"

17:08

bors[bot]

[ddnet/ddnet] Pull request closed: #4922 Reset thread count too, so the implementention knows its not multi th…

def-

[ddnet-web:master] 1 new commit

d7c9830 16.0.1 - def-

@Ryozuki when update

17:36

aur package

17:36

xd

17:36

later

17:36

im playing games

17:36

Jupstar ✪

aur package

*our

*ur

*u

its funny that there are probably more ppl testing vulkan, than there ever were ppl using opengl 3.3, just bcs it mentions more fps xdddd

17:38

clickbait works

17:39

even the ppl on 32bit suddenly care xD

so no fps?

not for everyone xd

ill try it sometime

soon™️

i have a 1660

17:40

nevr

sounds good

17:40

oldest NVIDIA i tested was gtx 760

17:40

oldest intel hd 550 and oldest amd, was mine xdd

17:40

ah no, amd rx 580 was oldest

used to use a 950

17:40

thank god i dont anymore

and i tested fairphone 3 and a newer phone from my brother (edited)

it was decent for a while. then i turned on genshin impact for the first time and i thought an alarm was going off in my pc

17:41

where is arch logo

17:41

fake ryo

im on windows

17:41

playing

17:41

my gif was removed this is censorship

i know its the amd driver program

17:42

looks quite nice actually

Jupstar ✪

and i tested fairphone 3 and a newer phone from my brother (edited)

you use fp3?

Chairn

you use fp3?

my mum xd

i got one myself

i have very old phone

17:42

bcs honestly i dont care about phones xD

17:42

i only use them for browsing

400€ for a phone

one plus x

i at most spend 200€

17:43

i cant hardly justify spending more when i dont use it much

17:43

but i gadly spend 600€ on a gpu

17:43

17:43

gladly*

Ryozuki

400€ for a phone

for something that i'll use for 10 years, doesn't sound so bad

10 years is hard

that's why i fight here to keep support for old gpu, cpu and systems

17:44

both hardware and software obsolescence should be forbidden

a 200€ phone literally has everything

17:44

the only thing u pay is probs a camera

it's not about having everything, it's about having a reparable phone

i dont agree totally

17:46

cuz i like progress

17:46

and old tech only makes us stuck

we are in a phase where technology is just at the beginning and thats why we have so many new phones etc almost yearly. In many years it will be different, of course that may take decades still, but at some point there will be one state of phone or whatever that you will use for more than 2-4 years

yeah hard topic, if we recycle shit more than its maybe ok to say more often updates

17:48

else its waste xd

17:49

(complete phone replacements, not updates xd)

ppl using blackberry isnt even 10 years ago

17:49

imagine

17:49

xdxd

Blackberry bulgariantroll

Ryozuki

10 years is hard

I’ve never had a phone last 10 years. That’s just not a thing anymore. Planned obsolescence gang ruined that

Learath2

I’ve never had a phone last 10 years. That’s just not a thing anymore. Planned obsolescence gang ruined that

and its not realistic

17:51

10 years ago phone were rly different

mine is december 2015

17:51

i am getting closer

17:51

lets go

u would be stuck with 2g

No battery replacements anymore is also making it insanely hard to get it to last

xd

and then comes apple who pushes updates that destroy hardware on purpose

it even has GLES3 already xd

well apple is special

17:52

i never owned iphone

Me too

17:52

and im happy about it

imagine buying a lot to have less freedom

bors[bot]

[ddnet:master] 2 new commits

843acf6 Use std::size instead of sizeof(a) / sizeof(a[0]) - def- c592e71 Merge #4917 - bors[bot]

17:52

bors[bot]

[ddnet/ddnet] Pull request closed: #4917 Use std::size instead of sizeof(a) / sizeof(a[0])

https://stallman.org/apple.html

17:52

read the bible

17:52

and thank me

17:53

Right to repair

>

Apple Is Lobbying Against Your Right to Repair iPhones, New York State Records Confirm.

>

Apple forbids recyclers of Apple computers from extracting any usable spare parts from them, by imposing nasty contracts.

>

Apple's conduct should be forbidden by law so that no company can ever do this.

>

The iPhone 7 has DRM specifically to brick it if anyone other than an authorized repair agent fixes it.

>

The term "lock" is inadequate to describe this sort of malware. Let's use other words that show what's really going on.

>

Apple faces trial in Australia for bricking devices because they had had an "unauthorized" repair.

>

Apple machines are built with unusual screws that make it difficult for the owner to take them apart.

>

Along with technical barriers, Apple lobbies against "right to repair" laws. Apple is campaigning against right-to-repair laws that weaken the unjust effect of the DMCA.

>

Apple is putting DRM on batteries, and the system software turns off certain features when batteries are replaced other than by Apple. Reevaluating Apple's reputation for good design: design for nonrepairability is not good design.

Wym forbidden by law, are you a filthy socialist?

17:54

How could you suggest that the profit motive isn’t in the consumers best interest?

Apple exploits the app developers mercilessly, aside from a few stars whose role is to give a misleading impression of what developers can expect.

>

I can't sympathize much with those app developers, since they are making proprietary software. They all deserve to fail. However, that doesn't excuse the way Apple treats them.

17:54

i love stallman xdddd

17:54

they literally have drm on batteries

17:54

apple is insane

It’s more money for the shareholders, thus it’s good

thats why steam is so good for now

17:55

no shareholders to appease

17:55

only lord gaben

@deen https://github.com/ddnet/ddnet/issues/2937#issuecomment-1083424923No idea what you mean I isolated haha. Wasn’t @Learath2 working on a refactor that isolated things even on the server side?Also to have complete physics you need more than CGameContext

Teehistorian Replayer · Issue #2937 · ddnet/ddnet

I haven't really started yet, if someone wants to help tell me: Replay modus on server that opens a teehistorian file as inputs Should run faster than realtime, no sleeps, instantly go to n...

we don't have a button to save ghosts before they finished but they will be overwritten as soon as you finish with a better time. Why do we have this save button? is it only active when you have finished but not automatically save the gohst on? https://github.com/C0D3D3V/ddnet/blob/ee5652ce7f64a8642f35f09bf27026ae22046def/src/game/client/components/menus_ingame.cpp#L1201-L1209

ddnet/menus_ingame.cpp at ee5652ce7f64a8642f35f09bf27026ae22046def ...

DDraceNetwork, a cooperative racing mod of Teeworlds - ddnet/menus_ingame.cpp at ee5652ce7f64a8642f35f09bf27026ae22046def · C0D3D3V/ddnet

18:03

@Jupstar ✪ your last fix fixed it

no crash anymore

nice, thanks for testing

Changelog should appear in the main menu screen after update. Many softwares have the "What's new" button or popup overlay after update.

it still gets a highlight

1

bors[bot]

[ddnet:master] 2 new commits

7bf967f Check if the resolution is "really" not supported before applying desktop resoltution - Jupeyy d94bc56 Merge #4923 - bors[bot]

18:16

bors[bot]

[ddnet/ddnet] Pull request closed: #4923 Check if the resolution is "really" not supported before applying des…

I have not checked the news once since the new Lay-out haha (@Jupstar ✪)

Jupstar

nice, thanks for testing

18:25

Aaa wrong reply

18:25

I go kms

18:26

Meant this bitsh (@Jupstar ✪)

Jupstar

it still gets a highlight

chillerdragon

I have not checked the news once since the new Lay-out haha (@Jupstar ✪)

xd guess you also too close to development

18:31

for the users its more intresting

No I always loved reading the news when they auto popped up

18:43

But I never noticed them since

C0D3D3V

[ddnet/ddnet] Pull request opened: #4924 fix gohst transparency of hook hand and ninja skin

https://streamable.com/oquu51 reported by lomination fixes #4207 again

Checklist

[x] Tested the change ingame
[x] 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
[x] Changed no physics that affect existing maps
[x] Tested the change with [ASan+UBSan or ...

nightly version on steam is still on 15.9.1 (edited)

yes but it should have most features from 16.0

19:05

the version was just updated today

ok ^^

who is censoring ma backtrace xd

19:18

https://paste.zillyhuhn.com/lR

use RelWithDebInfo

use -DDEV=ON when calling cmake

I use FAKIN Debug

19:20

ou wats DDEV ?

19:21

DDEV didnt help either

19:21

i mean i have debug symbols in my code but just no backtrace

19:22

maybe its in a lib? or i f'ked up the stack?

19:22

q

.#1 0x00000000000001e0 this indeed looks like fucked up stack (edited)

https://github.com/ddnet/ddnet/issues/1

Show time of player you're spectating · Issue #1 · ddnet/ddnet

Probably difficult since spectators don't have a character?

sad dragon noises

DDNet

https://github.com/ddnet/ddnet/issues/1

stupid bot

19:24

just use asan

oh smort

19:24

oke lemme try

for -DDEV=ON with cmake, it's probably better to start in a fresh build folder

19:25

mkdir _build && cd _build && cmake -DDEV=ON -GNinja ..

omagawd ninja

actually, im surprised the call on #1 doesn't segfault

https://github.com/ddnet/ddnet/issues/1

Show time of player you're spectating · Issue #1 · ddnet/ddnet

Probably difficult since spectators don't have a character?

OMG, is this bot this fucking stupid ??

its my stable code using AI and blockchain to recover from crashed on the fly by doing memory analysis

1

what does it do if i wanna sort out my problem #124345939874603460246032643

#1

19:27

have you tried not using discord?

no

lol why are so many issues related to the first issue

19:29

ah because all of them have stack traces with ddnet#1 in it

https://github.com/ddnet/ddnet/issues/1

Show time of player you're spectating · Issue #1 · ddnet/ddnet

Probably difficult since spectators don't have a character?

good ol out of bounds access

i am the #1 player

https://github.com/ddnet/ddnet/issues/1

Show time of player you're spectating · Issue #1 · ddnet/ddnet

Probably difficult since spectators don't have a character?

@Chairn asan did good

stack overflow

hey, anyone know how is generated ddnet-info.json ?

19:32

i mean, is there any script available ?

ChillerDragon

[ddnet/ddnet] Pull request opened: #4925 Ignore ubsan/asan logfiles

does anyone know bitbucket issues?

can anyone help me with ddnet settings? activated new vulcan settings just for trying and now it wont open again

Rene(Killer)

can anyone help me with ddnet settings? activated new vulcan settings just for trying and now it wont open again

what is ur gpu

1060

20:12

gtx

update drivers from nvidia

did is most actual

from nvidia directly?

also tried reinstall

20:12

y

win10?

11

20:13

but never had any probs bout this with gaming

what driver version does it report?

by my gpu?

20:14

512.15

20:14

dropped 22/03/2022

ok

20:14

then use "gfx_backend opengl" in steam start parameters

how tf i open them ? xD

weird, it shouldn't crash with a modern gpu

right click ddnet

are you using 32 bits version?

properties

20:15

then its there

20:15

win11 has no 32bit

i mean 32 bits ddnet

32bit ddnet has no vulkan xd

dang

Jupstar ✪

then use "gfx_backend opengl" in steam start parameters

worked like wtf bro

@Rene(Killer) one question tho, do you use a laptop?

no

ok

ty tho

np

20:16

mhh my bro's girl friend has a gtx 1060 on a win10 laptop, lets see if that also causes problems

xDDDDDDDDDDDD

@Rene(Killer) can you quickly look in %appdata%/Teeworlds/dumps if there are crash logs created?

there are

can u send them to me?

out

https://zillyhuhn.com/OpenTube/videos/users/chiller/term-ux.mp4

1

20:25

chillerbot-ux curses mode now in the aur term-ux-git

does ncurses also work on windows?

na

20:26

its linux only client as of right now

too bad

20:26

would be a nice replacement

i mean somehow probably possible to get it working but i dont really have the motivation to boot into windows in my freetime anymore xd

20:27

Idk what vim uses but vim runs pretty well on windows

20:28

@Jupstar ✪ the windows user

i just want best experience

20:29

even for windows users xd

replacement for what tho?

for windows console window

its still a console window :D so you want to launch a curses client as console window next to the GUI client instead of just the logging to stdout?

20:33

i doubt windows users have a use case for that anyways. Its more for terminal nerds like me or ppl who need a client on a vps

i mean for ddnet server

now i am confused

could add commands in server, like minecraft server xDD

aaaaaa

20:34

that yes

20:34

yea thats cool in minecraft ngl

20:34

when rewrite server in java?

@Chairn if you care to know, it probs was too many skins loaded

20:35

yeah teeworlds in java

20:35

1gb ram to start the client

ah thx for the info

ah not client

20:35

server in java client in rust

we really need lazy skin loading

graphics backend in bash

generally assets

[quakenet] ChillerDragon

graphics backend in bash

there is a ascii renderer

20:36

ever tested that?

renderer of what?

20:37

tw specific?

no one that can convert pixels to ascii

20:37

its even in VLC under linux XD

20:37

watched animes in it xD

i converted images to ascii

20:37

and videos to images

ah yeah, i've seen the ascii matrix movie

and video images t ascii gifs

20:38

but never used a full ascii renderer taking video files directly

20:38

wait so VLC can output ascii? xd

yes

20:39

even colored asciii

20:39

but that never worked for me, dunno what lib is broken xD

vlc -V caca fileToOpen

20:40

that didnt work for me

20:40

ez

but did you use UI?

20:40

if you gotta use UI to activate ascii you already lost xd

yes

20:41

i am ui nobo

20:41

i thik later i used mplayer to use colored ascii right

20:41

it was just broken in vlc

20:41

i also found commands to resize the frame buffer

20:41

was actually quite high quality

sjrc6

[ddnet/ddnet] Issue opened: #4926 unnecessary hookline change revert?

The player hookline that is rendered in the https://github.com/ddnet/ddnet/pull/4921 bug report video is not from the local player, I don't see how my change could possibly have caused this bug. I did a test and from what I can tell calculating the local hookline from the mouse position is identical to the old way except that it updates quicker. https://user-images.githubusercontent.com/22122579/160928974-75d80554-4c7a-4085-9644-b6ee17f94ecb.mp4

[quakenet] ChillerDragon

https://zillyhuhn.com/OpenTube/videos/users/chiller/term-ux.mp4

pretty cool nicely done

thanks :)

nice ChillerDragon

[quakenet] ChillerDragon

https://zillyhuhn.com/OpenTube/videos/users/chiller/term-ux.mp4

so cool

def-

[ddnet/ddnet] Pull request opened: #4927 Revert "Revert "Update players.cpp"" (fixes #4926)

This reverts commit 36bfe24522f468b06af401a9e6ee352b47599451.