DDraceNetwork - developer

chillerdragon

owo

gtfo

5

Headshooot

bors[bot]

[ddnet:master] 2 new commits

ac56075 Upgrade resolution list - Jupeyy acb8717 Merge #4107 - bors[bot]

08:00

bors[bot]

[ddnet/ddnet] Pull request closed: #4107 Upgrade resolution list

09:26

chillerdragon: i am haunted aswell xD

white theme programmer

09:41

send help

09:41

true lol

09:42

should install a dark notepad++ theme

09:42

I don't like sublime for some reason thou

i use neovim

09:42

I use vim because I only use git bash

09:44

tree-sitter based highlighting is the future https://tree-sitter.github.io/tree-sitter/

@Robyt3 i don't know what's wrong but you can shorten x ? true : false to x

lol thats true

yes, I don't know why that implementation was changed from vanilla to begin with

git blame

09:45

I think Vanilla did a circle clip, which is a bit useless and expensive

I think in GameLayerClip it's the same

so we instead do a rectangle, since we only support rectangular resolutions

imagine circular resolutions

this is vanilla (after refactoring)

looks more readable i guess

code style question: should I litter everything with const?

for me yes

TileX and TileY could also be const but it does it make a difference to the compiler?

09:48

I try to make everything const when possible so I know it can't change

i think it may help optimize stuff but idk im no expert

@Robyt3 yeah, const should've been the default in C++ and a modifier been used for non-const

Ryozuki

i think it may help optimize stuff but idk im no expert

compilers are pro

09:53

they find it

ye in rust everything is immutable unless specified

09:55

ez for optimization

deen

@Robyt3 i don't know what's wrong but you can shorten x ? true : false to x

same thing in CPlayer::IsReady() as i remember xd

I don't think the compiler is allowed to do much with the const in C++ but it's good to have for the programmer

In Haskell everything is immutable, period.

09:59

Yeah, mainly for documentation for programmer so you know this value will stay the same for the entire function

09:59

/ scope

@Learath2 can you sumup what yours and nuborns discussion with fokkonout ended up with?

That's also annoying to work with imo, default immutable and sprinkling in mutability as needed seems to maximize readability and optimizability

@deen trml is @trml

Yes, he told me on Discord

@Deleted User we want to send a single switch state, that will be the state for the team of the player

and has a letter on the way to him

:o

Learath2

That's also annoying to work with imo, default immutable and sprinkling in mutability as needed seems to maximize readability and optimizability

yeah also not sure if i like it default

u know how much it takes?

deen

and has a letter on the way to him

ez

how long*

I'm a bit annoyed everyone has a different name on Discord and GitHub, but can't complain since I'm part of the problem

Learath2

@Deleted User we want to send a single switch state, that will be the state for the team of the player

do you think somebody will do with within the next days?

I have the same name on every single platform, so I can complain

i dont want to have a weeb name on github monkalaugh

Deleted User

do you think somebody will do with within the next days?

Already done, I'm reviewing it

i have different name everywhere xd

It's different for each country. You can choose a country here and check it: https://www.deutschepost.de/de/b/briefe-ins-ausland/laenderinformationen.html

Länderinformationen

Länderinformationen für Ihren Brief ins Ausland: Was müssen Sie beachten bei Versand und Zoll? Wie lange braucht ein Brief von Deutschland ins Zielland?

yeah you e-mail is funny xd

10:01

your*

spanien

10:02

hmm my german is not good

xddd we were in school, in a computer course, and started trible wars(a browser game) and i needed a name

10:02

and my teenage pro brain comes up with it

heh

@Ryozuki 2-4 working days

:o

bors[bot]

[ddnet:trying.tmp] 1 new commit

e18e741 [ci skip][skip ci][skip netlify] - bors[bot]

10:20

bors[bot]

[ddnet:trying.tmp] 3 new commits

5d701d4 Bump GTest version to latest master as of 2021-04-07 - heinrich5991 f95f172 Revert "Manually patch gtest to allow CMake version 2.8.12...3.19.1" - heinrich5991 42446a6 [ci skip][skip ci][skip netlify] -bors-staging-tmp-3761 - bors[bot]

10:20

bors[bot]

[ddnet/ddnet] New branch created: trying.tmp

10:20

bors[bot]

[ddnet:trying] 3 new commits

5d701d4 Bump GTest version to latest master as of 2021-04-07 - heinrich5991 f95f172 Revert "Manually patch gtest to allow CMake version 2.8.12...3.19.1" - heinrich5991 fdf0cc8 Try #3761: - bors[bot]

bors[bot]

[ddnet:master] 2 new commits

b216b67 Add Android build scripts - Jupeyy 7e69796 Merge #4085 - bors[bot]

10:33

bors[bot]

[ddnet/ddnet] Issue closed: #4086 Rare(?) random(?) crash on Android

10:33

bors[bot]

[ddnet/ddnet] Issue closed: #477 Android support for SDL2

10:33

bors[bot]

[ddnet/ddnet] Pull request closed: #4085 Add Android build scripts

ok github caching is really weird

10:35

it still show the ci build for ubuntu 16

10:35

probably because the branch wasnt rebased

10:35

still weird xD

bors[bot]

[ddnet:master] 3 new commits

2191fdc Bump GTest version to latest master as of 2021-04-07 - heinrich5991 d00c5aa Revert "Manually patch gtest to allow CMake version 2.8.12...3.19.1" - heinrich5991 00ff285 Merge #3761 - bors[bot]

10:44

bors[bot]

[ddnet/ddnet] Issue closed: #3388 Stop patching gtest once they resolve the CMake deprecation issue

10:44

bors[bot]

[ddnet/ddnet] Pull request closed: #3761 Bump GTest version so we don't have to patch it anymore

@Learath2 do you time for a small debugging session with macos and newest SDL?

#if defined(CONF_PLATFORM_MACOS) // Todo: remove this when fixed in SDL
                case SDL_WINDOWEVENT_MAXIMIZED:
                    MouseModeAbsolute();
                    MouseModeRelative();
                    break;
#endif

That should preferably removed and i also want to remove for focus changing

Newest SDL as in master or newest release?

in SDL master

11:02

thats what we will use

Mouse capture has been proper broken on macOS since forever, I need to click outside and back inside to get it to properly capture

11:02

The main reason we haven't been upgrading sdl has been the massive performance drop on windows, is that fixed?

yeah

11:03

maybe its fixded for relative mouse mode i made a patch on sdl github

11:03

you can try it too and merge #4095 i will force push it now

https://github.com/ddnet/ddnet/pull/4095

https://github.com/libsdl-org/SDL/pull/4683

Make relative mouse mode inactive when window/keyboard focus is los...

Description Automatically deactivates relative mouse when keyboard focus is lost. No idea if this is a good idea at the keyboard event. But I guess it kinda is related to it in a relative mode Or s...

11:05

thats also going to be merged by our SDL

11:05

so SDL master + ^ ddnet master + #4095

https://github.com/ddnet/ddnet/pull/4095

what happened with embeds?

finally broken

11:14

ok forced pushed once again, since i need to test now on windows too

Jupeyy

[ddnet/ddnet] Pull request opened: #4115 Fix toolchain

broken by the android check

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...

switching from windowed to fullscreen is still not perfect :/ need to unfocus once and refocus and it works, whyever this is needed xD

11:37

so random

At this rate we'll end up replacing sdl with our own wrapper

true xD

11:41

but atleast the mouse stuff seems to work with windows thats nice

11:41

then i can introduce windowed fullscreen once again xD

11:42

multi monitor users will be happy

sddl

11:48

haha

11:48

ok

nothing more annoying than windows updates

why u use windows

"shutdown" ... please wait 5 hours

11:50

i tested newest SDL

bors[bot]

[ddnet:master] 2 new commits

942e54d Fix toolchain - Jupeyy 6cbb0c6 Merge #4115 - bors[bot]

11:58

bors[bot]

[ddnet/ddnet] Pull request closed: #4115 Fix toolchain

Jupeyy

[ddnet/ddnet] Pull request opened: #4116 Update GLEW

If the glew mantainer merges his pr we can remove the ifdefs But we can also use it like this for now, then atleast our official build has no x11 calls

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 physi...

Someone wants to review #3967 ? I'd like to hit the milestone today, else we end up in a ever waiting loop xd

https://github.com/ddnet/ddnet/pull/3967

Add state if no master was found by Jupeyy · Pull Request #3967 · d...

fixes #3951 by adding a state if no master was found Checklist Tested the change ingame Provided screenshots if it is a visual change Tested in combination with possibly related configuration o...

Deleted User

Someone wants to review #3967 ? I'd like to hit the milestone today, else we end up in a ever waiting loop xd

Eeeeh, it looks fine to me but I really wanted heinrich to take a look at it instead :/

but he is busy

12:20

i doubt he's here any time soon

12:20

we can still revert it, if its horrible xD

Yeah, he is out for another week

12:21

I guess it's safe enough and if @heinrich5991 wants to fix it better when he is back he is welcome to

bors[bot]

[ddnet:master] 2 new commits

bde7e85 Update GLEW - Jupeyy eaad036 Merge #4116 - bors[bot]

12:23

bors[bot]

[ddnet/ddnet] Pull request closed: #4116 Update GLEW

@Learath2 do you also have time to test SDL today?

Um, I can sneak it in if I don't make any mistakes on this fourier series exercise

@deen if you want to upgrade SDL please add https://github.com/libsdl-org/SDL/pull/4306 and https://github.com/libsdl-org/SDL/pull/4683

[DONOTMERGE] video: Prefer Wayland over X11 by flibitijibibo · Pull...

This is a draft patch that will (at last) prioritize the Wayland video driver over X11. This is here purely for internal testing only and is NOT meant for wider community testing; to test SDL Wayla...

Make relative mouse mode inactive when window/keyboard focus is los...

Description Automatically deactivates relative mouse when keyboard focus is lost. No idea if this is a good idea at the keyboard event. But I guess it kinda is related to it in a relative mode Or s...

I don't want to

12:27

If the upgrade fixes things I can do it though

it fixes the relative mouse focus change bug on windows

it should also fix quite a bit on macOS and let people compile on their own

12:29

currently you can't compile and run on macOS unless you have an alien mac like @TsFreddie or build a patched sdl

ok, I'll try it when I have time

hi

hi alien mac

oh ye

12:29

my alien mac

12:29

still don't know why

Oh and it should also let us build for a higher version of macOS, the building for the old sdk thing is quite annoying

problem is really, we kinda have to upgrade, else we'll never be able to support fancy stuff like 3879

12:30

#3879

https://github.com/ddnet/ddnet/pull/3879

Use SDL_FlashWindow to request user's attention by Jupeyy · Pull Re...

We should soon update to SDL 2.0.16 for our bundled libs, but don't merge before this happened. This drops the remaining X11 dependency. We should then also use GLEW with EGL for official b...

is ddrace rom public?

that even works on macOS @Learath2 xd

rom?

12:31

ddnet gameboy version

uhm

12:31

1 sec

bors[bot]

[ddnet:master] 15 new commits

f66abb7 Added color for strong and weak states - BloodWod-513 49686d8 delete hud and config variables - BloodWod-513 1cf43b4 Add /team -1 - gerdoe-jr eb92bd2 Whoopsie - gerdoe-jr 1661d51 fix display colors - BloodWod-513 b44dabd add bind and diff arrows - BloodWod-513 0aa2e37 check extended data - BloodWod-513 75289d1 add strong/weak icon - BloodWod-513 fd39c88 add strongweak.png - BloodWod-513 adb4729 add description for bind - BloodWod-513 5aafa27 change name file - BloodWod-513 5894c26 refectoring - BloodWod-513 4dd37f6 refactoring - BloodWod-513 d7107da Fix swap resets switches. - edg-l 73e7195 Merge #3960 #4006 #4103 - bors[bot]

12:31

bors[bot]

[ddnet/ddnet] Issue closed: #3806 /swap resets switches

12:31

bors[bot]

[ddnet/ddnet] Pull request closed: #4103 Fix swap resets switches.

12:31

bors[bot]

[ddnet/ddnet] Pull request closed: #3960 Display colors for strong / weak hook states

12:31

bors[bot]

[ddnet/ddnet] Pull request closed: #4006 Add /team -1

oof

12:32

long boi

ok

12:38

how tf fudgy made a tas for this

you mod the server to have the ability to save and replay player inputs

You can also project the results of your shots, noby had a tool to project rocket shots and stuff

4

Ryozuki

only thing missing is the stuff with the switches xd

12:39

rest is all new SDL related

TsFreddie

you mod the server to have the ability to save and replay player inputs

i see

and hopefully you don't do any of that client-side

i just want to make cool runs

ye, then server mod it is

but it looks like setting it up is harder than i antipacited

ye, i think current tasers all uses their own server code

it is quite difficult (thank god, I can't imagine dealing with every skiddy replaying runs)

i wanted to make a server for making tas with a normal ddnet client

12:41

but that wasn't really going anywhere soooo

and how do u mod a server

you clone the source code and change it directly

12:42

in C++

bors[bot]

[ddnet:master] 3 new commits

6101782 Make clicking refresh before current load is finished to calculate best master again - Jupeyy 915ab51 Add STATE_NO_MASTER - Jupeyy 43c3391 Merge #3967 - bors[bot]

12:42

bors[bot]

[ddnet/ddnet] Issue closed: #3951 master servers don't refresh if they weren't found at startup

12:42

bors[bot]

[ddnet/ddnet] Pull request closed: #3967 Add state if no master was found

TsFreddie

you clone the source code and change it directly

oh no

but you made me want to make the tas server really bad. but i really don't have more free time to allocate for a new project right now

don't make it open source

or learath will hate u

12:44

i thought he had like

12:45

converter from teehistorian/demo format

12:45

to mp4 or another thing to see what happens

mhh

13:22

ddnet server under windows with steam writes into the steam directory?

13:22

is that really wanted?

13:22

permission problems pre programmed

What does the server write?

13:24

sqlite file?

13:25

Do you propose to put it in the config_directory instead?

i'd say so

I agree with keeping the Binary dir clean, would be better

steam dir is often in program files (x86)

13:25

ddnet-server.sqlite is the file

13:26

autoexec_server.log too

def-

[ddnet-web:master] 1 new commit

fc1fc4b 2718 - def-

13:31

def-

[ddnet-web:master] 1 new commit

303056e Fix country flags - def-

I.K.U

but it looks like setting it up is harder than i antipacited

u can do 0.25x speed runs easily

13:34

but u need smth to save inputs and im not sure if u can do it without downloading sketchy clients

@Learath2 you are right, SDL_WINDOWEVENT_SIZE_CHANGED might be better to adjust the viewport then, instead of at the resize call, but sadly this doesnt fix the random resize stuff with windows fullscreen :/

random resize?

15:26

Are you getting spurious size changed events?

no, i resize manually, wait for the SDL_WINDOWEVENT_SIZE_CHANGED to adjust the viewport and sometimes the viewport is still wrong until i once onfocus and refocus

15:27

then its magically fixed xD

15:29

screenshots also dont show this, so the viewport must be correct

15:30

well maybe just a driver bug xDD, i'll try with dx12

15:34

even happens with dx12 :/

Deleted User

no, i resize manually, wait for the SDL_WINDOWEVENT_SIZE_CHANGED to adjust the viewport and sometimes the viewport is still wrong until i once onfocus and refocus

Is it that there is a missing event or that you get invalid size when you request the new size on the event

15:35

?

15:35

Have you tried using the size sent along with the event?

i change the size over the ingame menu and that calls SDL_SetWindowDisplayMode

15:36

that is apparently made for fullscreen

15:36

the when window got resized, i update the viewport, which works, bcs screenshots look correct

15:40

i mean you dont resize too often in fullscreen, but would be nice if one version just works ™️

15:41

in windowed mode everything works fine as intended

Sounds so stupid, I wonder why it happens :/

mhhh also happens in our current version tho

15:44

still annoying af

15:44

maybe i should test it with a software renderer, maybe its a driver thing after all xd

NO enough driverblaming

happens with software renderer too

15:46

so dunno

This happens in real fullscreen, right?

yes

15:46

i start in lower resolution and update it

So, what generates the resize event I wonder

and then window is broken af

15:46

the menu sends it

15:47

to the graphic object and that calls SDL

Nah, I mean inside SDL, what actually is the origin of the change size event, is it generated inside sdl since the window manager isn't in the way?

15:49

when i once onfocus and go in again, its normal

15:49

so weird dxD

Wait that looks like borderless fullscreen not real fullscreen?

its real

15:50

thats the weird thing

15:50

i dunno why it shows half of the desktop

15:50

thats part of the bug xD

I am very very confused with that, there should be no window manager to render the half desktop

yes, but it even goes black screen when i unfocus

15:52

so there is a video change

15:52

so it kinda must be fullscreen

Mh, does the half desktop actually move? maybe it's just stale data in the framebuffer

i assume so too, should test it

15:53

its updating indeed, really weird xD

That is the oddest thing I've seen, this is windows right?

yes

15:54

i start with low resolution, click a high one and get it

Windows has a weird mode where it can render over fullscreen applications but it looks awful, not crisp like this

i dunno, when i start with high res and go low, then the window is not centered(but no desktop is shown)

15:55

already tried forcing the window to be in mid, but doesnt help

15:56

completly random, but atleast it always was like that, or my windows is broken somehow xD

hm, how can we make sure it is actually in fullscreen mode?

well it defs does a video change, so atleast windows assumes it is

15:57

but no idea how to proof it actually is, in that moment

does a video change imply fullscreen mode?

real window can set the video mode normally

15:58

would be annoying af if a windowed window can set desktop video mode

I guess there is a reason why most games still require a restart when you change resolution, this is cancer

yeah :/

15:59

i should try with gfx_resizable 0

16:00

ok no chance, same

You could try asking in the sdl irc, but it's weird

it somehow doesnt set the right window size in fullscreen

16:01

thats what i assume

16:01

bcs my mouse is also kinda clammped

16:04

tested cs:go lmao, it resized like 3 times before having the right mode xD

16:04

dont want to know what hacks they applied

16:05

they also show the desktop for a moment xD

16:05

almost looks like they fixed it, how i fix it manually xD

16:05

maybe they go out of fullscreen first, i test that now xD

16:07

haha funny, that actually works xDD

16:08

ez hackz

lol

16:11

How do these engineers go to sleep is my question

16:11

I make one silly mistake and I stay up the whole night to fix it

xD

Jupeyy

[ddnet/ddnet] Pull request opened: #4117 Make fullscreen changes better (especially under windows)

This bug already existed in our current release, so its not related to some SDL update

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+U...

bors[bot]

[ddnet:master] 2 new commits

7234c04 Make fullscreen changes better (especially under windows) - Jupeyy a8c3654 Merge #4117 - bors[bot]

16:39

bors[bot]

[ddnet/ddnet] Pull request closed: #4117 Make fullscreen changes better (especially under windows)

https://www.tomshardware.com/news/samsung-reveals-ddr5-7200-512gb-ram

Samsung Reveals DDR5-7200 512GB RAM

The new era for RAM

17:07

lel

17:08

Samsung has announced at Hot Chips 33 that the company has developed the industry's first DDR5-7200 512GB memory module. In comparison to DDR4, Samsung's DDR5-7400 memory module offers 40% higher performance with twice the capacity at a mere 1.1V.

lol nice

17:12

finally that bottleneck gets some improvements

you have trouble with std ? :D (@fokkonaut)

fokkonaut

chillerdragon: i am haunted aswell xD

no xe

isnt that some intimate diesease? :D

18:05

i posted my SDL license plate pic because i saw it afte rdebugging SDL for 2 days straight :D

18:05

when update to SDL 1.2 ?

18:05

image.png

since when does SDL2 use opengl? XD

18:06

they have so many backends

18:07

i bet they default is dx or vulkan

i thought its good meme

18:08

found in #sdl :D

where is that even from

18:08

ah ok

seems like a dead channel anyways

funny thing is

18:09

sdl 1.x uses sdl2 xD

i idle in there to rant about fullscreen bugs

18:09

ikr

18:09

sdl compat

18:09

its supported by sdlvm

18:10

your sdl version manager of choice

https://github.com/libsdl-org/sdl12-compat

GitHub - libsdl-org/sdl12-compat: An SDL-1.2 compatibility layer th...

An SDL-1.2 compatibility layer that uses SDL 2.0 behind the scenes. - GitHub - libsdl-org/sdl12-compat: An SDL-1.2 compatibility layer that uses SDL 2.0 behind the scenes.

ye ik

18:11

omagawd bridge delay is so troll

18:11

when do you trolls migrate from discord to matrix?

does it have a webapp finally

18:12

what even is the current normal client, last time i used it it was called element or smth

chillerdragon

when do you trolls migrate from discord to matrix?

when everyone does

element is fakin electron i was told the webapp even has better support for screen sharing

and who will host the matrix server we all join?

18:15

if we all use matrix.org, we can also use signal

fstd ofc

18:16

hrhrhr

18:17

i love the drive by signal rant btw <3

18:17

i love ranting messangers in general

what rant

18:17

matrix.org with a centralized server is like signal

18:17

if everyone uses matrix.org nothing is won

yea

18:18

so you shoot agianst signal

18:18

i like it

actually i shot against not using matrix potential correctly

usually ppl like to praise signal cuz snowden fanbois them

18:18

but drive by shoot signal

18:18

matrix has no potential

18:18

its bloat spyware

18:19

but other than irc it has a proper iOS client :D

18:19

matrix is bad enough to be widley adopted :D

18:23

Watafak Debian11 wants to install updates when shutting it down?

18:24

If I want that I use Ubuntu or windows xD

what programm should even force it?

void CClient::SwitchWindowScreen(int Index)
{
    // Todo SDL: remove this when fixed (changing screen when in fullscreen is bugged)
    if(g_Config.m_GfxFullscreen)
    {
        SetWindowParams(0, g_Config.m_GfxBorderless);
        if(Graphics()->SetWindowScreen(Index))
            g_Config.m_GfxScreen = Index;
        SetWindowParams(g_Config.m_GfxFullscreen, g_Config.m_GfxBorderless);
    }
    else
    {
        if(Graphics()->SetWindowScreen(Index))
            g_Config.m_GfxScreen = Index;
    }
}

not only that oy from 6 years ago

@heinrich5991 https://github.com/ddnet/ddnet/pull/3928 doesnt fix the issue for me, i still need to define the functions in antibot_null.cpp as with ANTIBOTAPI, only having them in antibot_interface.h does not work.

21:54

Ah wait, it does work indeed, but one has to include antibot_interface.h to antibot_null.cpp

ChillerDragon: