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 2022-02-15 00:00:00Z and 2022-02-16 00:00:00Z

imagine not having loop {} which is like a while(true) BASED

BASED

meh most fixes are using booleans for places where bools are required

[ddnet:master] 17 new commits

ca78c78 Use C++17 (requires Debian 9 to build) - def- 4c8e33e Fix MSVC build - def- 1afd1a3 Fix Linux build with C++20 - def- e31f255 Use ninja instead of makefiles - def- 0d06193 gnu++17 - def- 62756db Enable modernize-make-unique clang-tidy check - def- 984aed2 Add more clang-tidy modernize-* checks - def- 68e4eb2 Add modernize-use-bool-literals - def- 2da5d7d Add modernize-use-emplace - def- b58470f Add modernize-use-equals-default - def- de17e99 Add modernize-unary-static-assert - def- 7da9f1b Add modernize-use-equals-delete - def- dbfc62d Add modernize-deprecated-headers - def- fef7af8 Add some more modernize-* checks - def- 18d9997 Fix clang-format - def- e615d4b Fix tools - def- 218f4a4 Merge #4710 - bors[bot]

07:29

[ddnet/ddnet] Pull request closed: #4709 Use C++17 (requires Debian 9 to build)

07:29

[ddnet/ddnet] Pull request closed: #4710 Add some modernize-* clang-tidy checks

o.O (@deen)

a secret large map, not released yet

08:43

lmao deen is such an professional :D (@Ryozuki)

unknown.png

justatest

wtf xd

can we somehow maybe get estimated number instead of what we currently have

09:38

so that developer doesnt have to die

the color indicates the estimated number

[ddnet-web:master] 1 new commit

e30b178 Fix release block size - def-

Souly

can we somehow maybe get estimated number instead of what we currently have

f1: leak_ip_address_to_all_servers

estimated number

11:12

that command actually leaks ip

real number > estimated

Souly

that command actually leaks ip

of course, that does serverinfo requests the old way by pinging each one manually

ye and that's not safe

Souly

ye and that's not safe

u wont die of it + just restart router or do it with a vpn

11:13

your client caches the pings

11:13

and saved them upon restart

11:13

so do it once and u r fine until new servers come

i will die

stop troll, its hardcore unfunny ^^

11:14

and will not get funnier over time xd

#showroom

watfaka is this live freeze :D i somewhat saw it on github and now after a merge with upstream i get random freezes haha

11:34

is it used in some map already?

not on official maps

hmk

11:35

do you have a tldr how it works?

you can't do movement key inputs with it

when?

11:36

on tile touch?

when you have the effect, one tile turns it on, one off

are you saying i drunk mapped that index somewhere?

yea

lol

I suppose ^^

INDEEED

11:36

144

11:37

HOW DAFAK DID THAT HAPPEN

kek

1

11:37

Ok thanks bois

11:38

@bencie lmao i always thought ur name is benice

11:38

as in ben ice or be nice

you just noticed?

11:38

:D

but bencie is like so odd

11:38

i noticed when i loaded the save game

Although I've heard that before ^^ it's my irl name with an i added

took me 10 mins

hahahahah

had to go to ddnet.tw and typed benice 3 times and somehow the name had no points

11:40

was about to contact ddnet staff that your name was wiped from the database

LOL

11:40

that's funny

If I would try to fake you in game it won't be hard to identify me

ben ice

souli

[ddnet/ddnet] Issue opened: #4711 DDNet doesn't load some visual assets when Streaming through Discord, showing black squares

Happens when using Discord to stream with direct application capture. Doesn't happen with other games or other capturing programs. Assets that aren't loaded seem to be random. Need to fully restart DDNet in order to get the assets. Starting the stream on an already started DDNet will unload assets and replace them with black squares (visible to both player and viewers). Tried to change DDNet settings (OpenGL, full screen, etc) and Discord settings without success. Opening this issue t...

@Patiga @Learath2 what is the executable that is run on $optimize? Want to do it locally

edit_map, let me check the options used

13:23

--remove-everything-unused --shrink-tiles-layers

why not --shrink-layers?

because that would touch the game layers. that will change the position of the kill barrier and if the map is shrunk from the left or up it will move the map's floating point precision

13:26

I first made it do --shrink-layers, but I think Ravie and maybe others were against that

ok

-> if you want to do it locally on a normal, map, --shrink-layers should be harmless

@Jupstar ✪ @Learath2 here us caring about 2gb map when they require 12gb ram nowadays

Image attachment

13:32

PepeA

Minimum 12g memory pepeW

pepeW

13:33

We are so in the wrong timeline

i wonder if the game is coded in java or something lmfao

memory is simply the easiest hardware to scale

bad memory usage apologists smh

i understand that you are mad if memory is wasted

13:41

and i agree in this case

13:41

but gameengines probably simply try to not unload textures/vertices etc as much as possible

its hard for me to comprenhend how they need so much ram

is that an MMO or smth?

13:43

open world?

its elden ring

13:43

yes open source

13:43

souls like

13:43

open world

13:43

*

13:43

lmao idk why i said open source

13:43

xd

that was ur desire xd

xd

13:43

imagine looking into the source code of a AAA game

13:43

would be interesting

13:43

or maybe disappointing

its alot of scripting language

13:44

for all events basically

i expect some native fully optimized assembly here and there

13:44

like crypto people

only physics and other stuff that need high performance use lower level language mostly

Ryozuki

imagine looking into the source code of a AAA game

I mean if you are interested unreal engines source is available, it's a massive part of the source code

ah

13:45

i wanna know about the frostbite engine source code

13:45

bf3 and bf4 look too good for their age

13:45

must be some good engineering (edited)

the engines are often very good indeed

I'm more interested in what rockstar has, their open world games have always been above the rest for me

ah ye

13:46

and they perform rly well

13:47

gta 5 is a beast

13:47

the amount of detail

13:47

on a open world game

13:47

it has no match

RDR2 aswell, great performance for a game of that scale

i'd say a lot of approximation for stuff that should look good but not be too realistic

@Learath2 we are oficially in c++17 land

\o/

13:48

I can remove that one ugly void_t hack

i wonder if #4711 is fixed by using vulkan

https://github.com/ddnet/ddnet/issues/4711

DDNet doesn't load some visual assets when Streaming through Discor...

Happens when using Discord to stream with direct application capture. Doesn't happen with other games or other capturing programs. Assets that aren't loaded seem to be random. Need ...

directly used the automatic std::array auto template deduction

13:50

https://github.com/ddnet/ddnet/blob/08a40740b38e9233619e83534fcb1cf59106468b/src/engine/client/backend/vulkan/backend_vulkan.cpp#L2049

ddnet/backend_vulkan.cpp at 08a40740b38e9233619e83534fcb1cf59106468...

DDraceNetwork, a cooperative racing mod of Teeworlds - ddnet/backend_vulkan.cpp at 08a40740b38e9233619e83534fcb1cf59106468b · ddnet/ddnet

why is DOWNLOAD_GTEST on by default in windows

yeah a bit annoying, but heinrich said he did it, because google said you should do it xD

@Jupstar ✪ are u using the c++ vulkan api?

Ryozuki

i wonder if #4711 is fixed by using vulkan

i'd hope so

or the C one

C

there is a c++ one btw

i know

why not the c++?

i dunno i just prefer to write my own wrappers

coffee

Ryozuki, what are your feelings on this?

PES3_Flushed

Heinrich would have a spicy opinion about that too

He is not a big fan of rolling your own almost anything

i would use the c++

13:55

im sure its safer

13:55

they actually use smart pointers iirc

the problem is, you dont gain alot from it

13:56

device->dosmth or dosmth(device)

Jupstar ✪

yeah a bit annoying, but heinrich said he did it, because google said you should do it xD

worst reason to do so

https://github.com/edg-l/SuperSDL/blob/master/src/graphics/renderer.cpp i think i achieved rendering a triangle

13:57

old code xd

\o/

did you at least understand it?

PepeA

or just copy paste xd

some stuff

13:58

not everything

13:58

Image attachment

i mean in the end vulkan still feels similar to OpenGL, its just that OpenGL hides synchronizations and allocations

u rly have lot of control

13:58

xdd

and then you have 10000000000000 lines of code xd

tho if u ask me now i dont remember much

vk::PipelineLayoutCreateInfo PipelineLayoutInfo = {}; for example is even more verbose than c version vkPipelineLayoutCreateInfo PipelineLayoutInfo{};

14:00

xd

this is from when i thought using only C and not c++ was the cool kids thing: https://github.com/edg-l/SimpleGame

(how i used in inside c++ code ofc)

xd

therefor you can code the compiler for you simplegame in 1 month

14:01

while for rust and c++ you'll need years

[ddnet/ddnet] Pull request opened: #4712 CMake: Use /utf-8 for MSVC

Because the good guy microsoft respects the old habit of using local codepages for source code in Asia. (In china and japan MSVC defaults to treating the source code as GBK and SHIFT_JIS encoded still.)

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 o...

weird

TIL I can't compile teeworlds in china

do you own a pure chinese computer now, or what? xd

Actually vk::PipelineLayoutCreateInfo PipelineLayoutInfo{}; should be fine in C++ too, no?

Learath2

Actually vk::PipelineLayoutCreateInfo PipelineLayoutInfo{}; should be fine in C++ too, no?

ah yeah, i only meant the scope

oic

but yeah dunno, i dont want to rewrite it now

14:03

the documention is in c, so i use c

namespaces are one of the few C++ things I actually enjoy

its simply more accessable for me

14:04

i like template programming with c++ most i think

14:04

you can do so much bs with it, and that's what i like

14:04

you never learn out what quirks are possible with it xD

thats why i avoid c++

14:05

what if some quirk is UB

14:05

monkalaugh

its just a meta language

https://stackoverflow.com/questions/56781981/c-variable-template-this-is-ub

C++ Variable Template. This is UB?

I created a simple C++14 variable template that calculates factorials (just for learning). Then I want print the first 12 factorials. template <int n> const int fact = n * fact<n - 1>;

just bcs rust says something isnt UB by documenting it, doesnt mean that its intuitive

14:06

are condition variables in rust FIFO? do they garantuee to exectue immediatly?

14:07

cond.notify_all();
cond.wait(mutex);

what happens here?

14:07

do we wait, or do we notify ourself?

14:07

xd

I think that's well defined, you've missed that notification

it's not well defined in a way that you think its intuitive

14:08

you can signal yourself indeed

14:08

(in c++)

14:08

that's why i am asking

This method will ensure that any current waiters on the condition variable are awoken. Calls to notify_all() are not buffered in any way.

yes it makes no garantue that it executes directly

It has to execute directly if it's not buffered in any way

14:10

Rust is lacking a standard that defines this terminology though, so who really knows

@Jupstar ✪ did u try rust yet

i sometimes read a bit of code, but i still dont like the syntax, i want a c like rust xd

14:12

i have nothing against the concept of rust, defs helps to not shoot yourself

hmm

14:12

i dont dislike the syntax

14:12

sometimes it can get a bit weird

tho, when i'd use it, i'd use safe only code, else it's useless xd

when u have lot of types

14:13

but that also happens in c++

14:13

i only used unsafe code to call C functions

well i am ofc also fine with proven code

14:13

not everywhere safe code can be used ofc

@Jupstar ✪ are you sure that it's not ordered properly in C++?

i feel like rust will look bad until u try it and get to intermediate level atleast

Learath2

@Jupstar ✪ are you sure that it's not ordered properly in C++?

gimme a second

or atleast past the initial stage where u think the compiler trashes u for no reason

C++11 30.5p4:

The implementation shall behave as if all executions of notify_one, notify_all, and each part of the wait, wait_for, and wait_until executions are executed in a single unspecified total order consistent with the "happens before" order.

is "happens before" order defined?

14:16

i never rly read a standard much

14:16

i guess they are like legal texts

14:16

u gotta define everything

14:17

@Jupstar ✪ we have 343 issues

14:17

the discussion stuff totally didnt work

14:17

we have to make a purge

Ryozuki

is "happens before" order defined?

Yep 1.10p12

[ddnet/ddnet] Issue closed: #2539 Segfault on Server when trying to load a map that doesn't exist

Basically if an evaluation is sequenced before another (due to a sequence point (like ;)) it happens before or if an evaluation inter-thread happens before another

14:20

When you want to actually analyse whether there is a data race it gets very annoying very quick to figure out what "happens before" what

just use rust 4head

14:21

dont mind me im just here chatting cuz this endless queue

Image attachment

14:21

feelsbadman

https://github.com/teeworlds/teeworlds/pull/2891#issuecomment-877839655 here is a quick example

[sound] remove volatile and use clamp function by vfjpl · Pull Requ...

this produces better assembly https://godbolt.org/z/9T5oTeYh1

Ryozuki

dont mind me im just here chatting cuz this endless queue

everyone is perma stuck in queue, I wonder why amazon just wont run more servers with their bottomless wealth

ye its weird

14:22

@Learath2 i think its cuz the game design

14:22

mayeb they didnt design it to scale that well

Learath2

https://github.com/teeworlds/teeworlds/pull/2891#issuecomment-877839655 here is a quick example

hey i remember we removed volatiles too

14:23

i did it actually xd

Justifying removing them is not actually as trivial as one might initially think

https://github.com/ddnet/ddnet/issues/4216

Volatile cleanup · Issue #4216 · ddnet/ddnet

The volatile types do not provide inter-thread synchronization, memory ordering, or atomicity. we should use _Atomic or std::atomic ❯ rg volatile game/server/gamecontext.cpp 3657: static volatile i...

It's easy to quote some folk tale "Volatile is never what you want"

Both the lock and the volatile can be avoided by just using an atomic here

14:26

this was mostly our case too

14:26

we just needed atomics instead of volatile

14:26

(quoting u btw)

14:28

@Learath2 i like how instead of agreeing with u he calls for mom

Image attachment

14:28

well im joking obviously but it looks like t hat

14:28

monkalaugh

14:30

PepeA

Image attachment

14:30

well its my time bye

o/

Ryozuki

Both the lock and the volatile can be avoided by just using an atomic here

Fwiw I did actually believe that volatile was necessary until I actually went through the steps of figuring out what happens

Learath2

everyone is perma stuck in queue, I wonder why amazon just wont run more servers with their bottomless wealth

I think queues can make something seem more popular and be used intentionally. (Not sure sure if this is the case here though)

i doubt since its a f2p game and the only way to spend money is if ur past the queue inside the game

@Learath2 ok sry

14:48

my example only works with the wait that doesnt take a extra condition https://en.cppreference.com/w/cpp/thread/condition_variable/wait

cpp/thread/condition variable/wait

@Jupstar ✪ does it though? Which part of that page should I read?

the first sentence

"The effects of notify_one()/notify_all() and each of the three atomic parts of wait()/wait_for()/wait_until() (unlock+wait, wakeup, and lock) take place in a single total order that can be viewed as modification order of an atomic variable: the order is specific to this individual condition variable. This makes it impossible for notify_one() to, for example, be delayed and unblock a thread that started waiting just after the call to notify_one() was made." This note derived from the part of the standard I quoted seems to suggest that only a spurious wakeup is possible in that case

14:53

But that has nothing to do with the notify_all imo, spurious wakeups just randomly happen

mhh I'm not so sure about that

Also, no need to be sorry, it is very annoying to figure out what all this means

i could swear i had a bug with condition variables once, and it was bcs the wait returned for no reason

14:54

and notify_one fixed it then xd

Wait does return for no reason at all, called a spurious wakeup, you always have to check a condition when you wakeup

Ryozuki

i doubt since its a f2p game and the only way to spend money is if ur past the queue inside the game

but if people hear that others are waiting in the queue, they'll think "wow, that's a popular game, maybe I should check that out too".

Learath2

Wait does return for no reason at all, called a spurious wakeup, you always have to check a condition when you wakeup

ok yeah, then i accept this too, but this is still unintuitive(i guess that was our starting point of the discussion)

I actually read some ancient discussion on spurious wakeups once. I think the arguments were that 1) it allows for much higher performance and 2) when you wakeup you have to check a condition anyway, since a broadcast waking up someone else before you might have changed the condition

14:58

Oh and 3) I think a POSIX guy argued that it forces people to have "correct" code if you threaten them with spurious wakeups

xD

[ddnet:master] 2 new commits

3d450e2 Use /utf-8 for MSVC - TsFreddie b4d699d Merge #4712 - bors[bot]

15:03

[ddnet/ddnet] Pull request closed: #4712 CMake: Use /utf-8 for MSVC

@Learath2 would you say we should also guard the variables in our code, it's probably safer, i think i'll create a pr

guard?

check the conditions correctly

If we are using a cond var anywhere, we definitely should be

15:18

Either with wait + if or the predicate form, either is okay with me

@Learath2 what does CAutoreleasePool AutoreleasePool; fix btw?

[ddnet/ddnet] Pull request opened: #4713 Give the condition waits proper conditions

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...

I think SDL is internally creating some refcounted types and if we don't have an outer pool to catch and release those they leak

15:34

Some obj-c refcounted types that is

weird stuff xD

15:35

i wonder if that's actually still needed

idk what tool was used to find the initial leak, so idk how you'd check if removing it changes sth

oh ok

Each thread creates a thread-wide autorelease pool automatically, but it's not drained until the thread is dead, so the gfx thread having a local one makes sense

i wonder why they even use such things, but anyway, thanks for explaination

Well obj-c is implementing modern language with C aesthetics

15:44

The staple of a modern language is the ability to just allocate stuff on the heap and then forget about it

15:49

It does allow some very interesting things, obj-c is wild

15:49

You can monkey-patch stuff, literally replacing running code with other compiled code at runtime

I think we should consider the tutorial map and vulkan support as a major improvement and release that as 16.0

3

15:59

and use that for the last Steam marketing round we have available (not sure if they have much of an effect anyway)

Learath2

The staple of a modern language is the ability to just allocate stuff on the heap and then forget about it

is this rly it? or its sarcasm xd

automatic memory management is definitely something that unites modern high level languages

16:15

it cuts out a main cause of bugs and crashes at the cost of performance and memory usage

16:16

(the original term "high level language" was for C, FORTRAN, Pascal, etc, so maybe should say https://en.wikipedia.org/wiki/Very_high-level_programming_language)

is obj-c using a gc?

[ddnet/ddnet] Issue opened: #4714 LAN server crashes on map finish on Windows

Multiple reports on Discord in #bugs. Can't reproduce, happens with 15.9

Ryozuki

is obj-c using a gc?

It uses both approaches, it has autorelease pools that will deallocate everything in themselves when they go out of scope, and they have gc for some cases

[ddnet/ddnet] Pull request opened: #4715 Initialize m_aErrorString earlier (fixes #4707)

Just a guess that there might stil be garbage in this array

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 memc...

[ddnet/ddnet] Pull request opened: #4716 Try to revert #4586 (fixes #4714)

Just to get a build and test if this is actually the cause

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]...

[ddnet/ddnet] Pull request closed: #4716 Try to revert #4586 (fixes #4714)

[ddnet/ddnet] Pull request opened: #4717 Try to fix #4714

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...

https://dkb.io/post/google-search-is-dying

Google Search Is Dying | DKB

Reddit is currently the most popular search engine. The only people who don’t know that are the team at Reddit, who can’t be bothered to build a decent search interface. So instead we resort to using Google, and appending the word “reddit” to the end of our queries.

18:25

https://news.ycombinator.com/item?id=30347719

dbrereton

Google Search Is Dying

18:26

i also always put "reddit" at the end of my queries

18:26

specially for reviews

18:26

real people usually give more honest opinions than stupid paid websites

18:26

Image attachment

18:27

So how can we regain authenticity? What if you want to know what a genuine real life human being thinks about the latest Lenovo laptop?

>

You append “reddit" to your query (or hacker news, or stack overflow, or some other community you trust).

>

Google is dead.

>

Long live Google + “site:reddit.com”.

18:28

chillerdragon i found ur girlfriend https://sdomi.pl/weblog/15-witchcraft-minecraft-server-in-bash/

My thoughts on writing a Minecraft server from scratch (in Bash)

I wrote a working Minecraft server in Bash! wait, why did I do that

Ryozuki

https://dkb.io/post/google-search-is-dying

i more and more find stuff i want to know on reddit rather than stackoverflow xD

[ddnet:master] 2 new commits

1972a1b Initialize m_aErrorString earlier (fixes #4707) - def- dc52f81 Merge #4715 - bors[bot]

19:01

[ddnet/ddnet] Issue closed: #4707 Problems with saved config file recently

19:01

[ddnet/ddnet] Pull request closed: #4715 Initialize m_aErrorString earlier (fixes #4707)

Ryozuki

i also always put "reddit" at the end of my queries

same, depending on what I need site:reddit.com, or site:ycombinator.com, etc. but it's not necessarily google's fault, but of the people doing "seo"

I've actually noticed that I don't really enjoy this new smart search stuff that's been getting more and more popular for the last 5 years

you mean the sidebar?

20:17

There were some funny stories about that too: https://hristo-georgiev.com/google-turned-me-into-a-serial-killer

Google turned me into a serial killer · Hristo Georgiev

Hristo Georgiev's Website

[ddnet:15.9.1] 16 new commits

d153ed9 Version 15.9 - def- 5e8a337 Update spanish.txt - n0Ketchp 7084b0d Update translations - def- 775faf0 Update German translation - def- 526a64e newJumps → NewJumps - heinrich5991 fcd559c Add "groundjump-only" tile - heinrich5991 6f26ec4 Client: Fix projectiles rendering for demo playback - Kaffeine 8bb3d18 Update Chinese translations (by cheeser0613) - def- 9952cde Delay double click handling one frame after item activation - Robyt3 83db924 Lower the sound volume even further after #4568 - heinrich5991 8d66175 Update Portuguese translations (by Rockus) - def- 78d8ef9 Always destroy bullets/lasers on kill when in practice team (fixes #4694) - def- 7798bdc Add live freeze graphics by Soapy Sandwich - def- 054bb0b Initialize m_aErrorString earlier (fixes #4707) - def- e54319d Fix #4714 - def- e64eb8f Version 15.9.1 - def-

20:20

[ddnet:master] 1 new commit

18ee351 Version 15.9.1 - def-

[ddnet/ddnet] Pull request opened: #4718 Updated french translations

Added missing french translations

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/ddnet] Issue closed: #4714 LAN server crashes on map finish on Windows

21:22

[ddnet:master] 2 new commits

97a79c1 Fix #4714 - def- 84d896a Merge #4717 - bors[bot]

21:22

[ddnet/ddnet] Pull request closed: #4717 Fix #4714

[ddnet-web:master] 1 new commit

0d3288e Version 15.9.1 - def-

[ddnet/ddnet] Pull request opened: #4719 Better fix for sqlite3_expanded_sql

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...

22:14

[ddnet:master] 2 new commits

aca3029 Updated french translations - Chairn 22dd3f1 Merge #4718 - bors[bot]

22:14

[ddnet/ddnet] Pull request closed: #4718 Updated french translations

[ddnet-web:master] 1 new commit

cc5af72 Update translations - def-

Exported 327 message(s)