DDraceNetwork - developer

DDraceNetwork

Development / developer

Development discussion. Logged to https://ddnet.org/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 2024-08-30 00:00 and 2024-08-31 00:00

[ddnet/ddnet] Pull request opened: #8846 Update spanish.txt

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 (especially base/) or added coverage to integration test
[ ] 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-ad...

[ddnet-web:master] 1 new commit

bf24fa5 Add 6 € funding for UKR by ⚣ПАНК⚣ - def-

github-merge-queue[bot]

[ddnet:master] 2 new commits

099ec44 Update spanish.txt - n0Ketchp 528a382 Merge pull request #8846 from n0Ketchp/patch-21 - def-

07:10

[ddnet/ddnet] Pull request closed: #8846 Update spanish.txt

By the way, "Tee 0.7" tab name is not localisable (edited)

~~I wouldn't localize it in chinese anyway.~~

[ddnet/ddnet] Pull request opened: #8847 Update Ukrainian l10n (18.5)

[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 (especially base/) or added coverage to integration test
[ ] 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--undefine...

TsFreddie

~~I wouldn't localize it in chinese anyway.~~

I would though

no fucking way it's translated as тії

tii

It is, my fellow homo sapiens, it is

chiller doing the lord's work :'0

07:59

giving zCatch some love

chiller are you a 0.7 maintainer already?

Chiller is the 0.7

Does he merge his own prs tho

he's giving zCatch 0.6/0.7 some love

<ChillerDragon> @jxsl13 send good zCatch colors

08:08

<ChillerDragon> @Teero nah i got no powers nowhere :c

look in my mod I think in player.cpp or so

;(

<ChillerDragon> ez steal

ez

08:08

take everything

<ChillerDragon> poggers2

poggers2

08:08

<ChillerDragon> u got good colors from professional designer?

no

08:10

it's Teetime's original color

<ChillerDragon> woah what is this black magix algorithm color = max(0, 160 - GetNumTotalCaughtPlayers() * 10) * 0x010000 + 0xff00;

08:11

<ChillerDragon> how many kills does it support? does it get stuck at some point?

08:11

<ChillerDragon> oke ima try

08:16

<ChillerDragon> yeah hm gets stuck at red

<ChillerDragon> https://zillyhuhn.com/cs/zcatch_colors.gif

Embedded image

08:27

<ChillerDragon> hmmmm

iirc at like 20 kills it was at grey, at least on zCatch 0.7

<ChillerDragon> i need lerato the color wizard to get correct colors mapped from 0.6 to 0.7 xd

[ddnet-web:master] 1 new commit

ed918fb Update translation stats - def-

github-merge-queue[bot]

[ddnet:master] 2 new commits

cdc5b99 Update Ukrainian l10n (18.5) - EGYT5453 3d5b533 Merge pull request #8847 from EGYT5453/ua-18.5 - def-

09:34

[ddnet/ddnet] Pull request closed: #8847 Update Ukrainian l10n (18.5)

my github notifications have grown to an unmanageable level

Image attachment

stop watching whole repos

10:20

then it gets easier

Image attachment

10:25

lerato githubbing skill issue

I also made a big mistake. I forgot to note down the command I was using to generate my initramfs and it ended up getting pushed out of my bash history. If I make a mistake I might not be able to boot back in to try again

No backups?

I'll keep one, but idk if I can get my bootloader to use it, never had to use the fallback with this bootloader

increase bash history size

10:30

hehe

chillerdragon

lerato githubbing skill issue

or you are simply not as relevant as lea, no front

GitHub

Click to see attachment 🖼️

:D

when will we add Denglish as language

Imagine not having your bash history backed up on GitHub.

Replying to @Learath2 I also made a big mistake. I forgot to note down the command I was using…

plis help what new name

10:58

my oCAUNTT

10:58

HELP

whats wrong with your current name?

YES

i see

11:00

F1 -> player_name YOUR_NAME

NO NEW NAME MY OCAUNT DDNET

-# ok then player_name OCAUNT

11:02

can you write your question in your native language?

11:02

russian or chinese

IM RUSIAN

then ask in russian

DA

Задайте свой вопрос на русском.

POMOGI POMENAT IMA V DDNET

11:04

KAK

11:04

V GOOGL NE DOET

i told you: press F1 player_name oCAUNT

I

Я сказал тебе: нажми F1 player_name oCAUNT

lmao pro

Replying to @Jupstar ✪ i see

DAY TGK

11:05

YA POKOGU

11:06

JUPSTAR

What’s pokogu?

HELP FOR NEW NAME DDNET

Maybe oCAUNT is not his name but means account and he wants to transfer points?

11:08

What is your name?

Replying to @EGOR DDNEt HELP FOR NEW NAME DDNET

i ask him to ask in russian

11:08

but he spams his capslock latin letter russian

Should work too

JUPSTAR

I AM EGOR

11:09

I SPEAK VERY ENGLISH

Google translate should pick it up but it doesn’t find pokogu

chatgpt is overloaded with his bad russian xD

Image attachment

TG

11:09

TELEGRAM GIVE ME

NOWWWWW

You can write here

YOUR TELEGRAM JUPSTAR

it's only for family sry

11:10

u'll have to marry me first

JUPSTAR GIVE TELEGRAM NOW

Xd

I HAVE HIS TELEGRAM

YOU ASK RUSIAN

NO YOU

TEERO

Jupstar ✪

it's only for family sry

We are family, the ddnet family

justatest

1

gigachad

2

@MilkeeyCat: help

GO TO ME GAME DDNET

What is your name?

?

In game

1 SEK

11:13

piYAR55

omg you left the server again xd

hes on kog

11:14

xd

chillerdragon

@MilkeeyCat: help

With wat

this goes above my bed

Image attachment

we need translator

11:16

lmao

<ChillerDragon> ok nvm we solved the language barrier with gaming

11:36

<ChillerDragon> https://zillyhuhn.com/cs/.1725017747.png

Embedded image

chillerdragon: make it so that i can use load_map on demos

12:31

thanks

Sure I’ll open an issue. Weird that it doesn’t work

im motivated

12:48

to do a riscv emulator (edited)

12:48

and run linux on it

почему тут вмсе по английски пишут

i dont hold u back

SnusEbalneka_q1wor3iK

почему тут вмсе по английски пишут

what?

бля

what?

12:49

i cant read your msg

why does everyone write in English?

in what language should they write?

12:50

most ppl here know english we are from all kind of countries

12:50

germany, spain, italy, turkey

12:50

so we speak english

Jupstar ✪

in what language should they write?

I don't know Russian or English well

12:52

and then I realized why there are so many people who speak English here

in #off-topic you can also write russian

12:52

sometimes there are russians

I'll be using a translator then.

Jupstar ✪

sometimes there are russians

but I understood well, it's just that a lot of people write in English

SnusEbalneka_q1wor3iK

why does everyone write in English?

You know what would be better, if everyone was writing in their mother tongue, hell yeeeeeah

13:01

This chat would be all in german

das stimmt

tja

isso weilso is

https://youtu.be/fvz2CxyGcyQ @Avolicious you could learn a thing or two from that guy hehe. tl;dw he is running million dollar startups with this gigachad techstack: 1vps (16core 500€), no staging, no kubernetes, php, jquery, sqlite3

Mischa van den Burg

He makes $3M a year with a $400 VPS

13:16

Actually ddnet setup is quite close lol

13:17

deen also likes bare metal and beefy vps and ddnet.org uses jquery afaik

urgh

13:25

jquery

chillerdragon

https://youtu.be/fvz2CxyGcyQ @Avolicious you could learn a thing or two from that guy hehe. tl;dw he is running million dollar startups with this gigachad techstack: 1vps (16core 500€), no staging, no kubernetes, php, jquery, sqlite3

But my client is running K8s and makes multi billion dollars, not million

13:30

So nearly all my clients are investing into cloud-native, all multi billion dollar companies

13:31

So for me it makes the most sense to invest knowledge into kubernetes, multi-tenant environments and all various kinds of technologies

13:32

But sure, it doesnt always make sense to built it upon a hyperscaling infra. But for me its more a learning process & get some expertise based on a non financial ruin side project

Jupstar ✪

das stimmt

Scuse me, no foreign languages outside #off-topic

So chiller, I've looked into some of his websites & he uses other services mostly. He isnt a service provider per se, he just uses other services and is the middleman inbetween. Thats basically not my goal to do in life

13:39

But if you like something like this, to work as a middleware. You can also check some documentaries about dropshipping - same technique ( no legal advise to do from me tho ).

It is my professional opinion that cloud is a scam

does anyone know if llvm has a riscv assembler?

13:44

for riscv assembly

13:45

not llvmir

13:45

btw my emulator name will be rysk

Ryozuki

btw my emulator name will be rysk

ryokan

Ryozuki

does anyone know if llvm has a riscv assembler?

https://llvm.org/docs/RISCVUsage.html looks like it

13:53

ah wait no you still have to convert it to llvmir ig

@Learath2 how to differentiate pointer variable declaration and multiplication expression statement, they are both ident * ident :\

meloƞ

https://llvm.org/docs/RISCVUsage.html looks like it

but thats to assemble llvmir to risc

14:24

i want to assemble assembly to machine

14:24

like riscv64-elf-gcc

if there's nothing in https://github.com/orgs/riscvarchive/ ig there's just nothing

RISC-V Archive has 66 repositories available. Follow their code on GitHub.

MilkeeyCat

@Learath2 how to differentiate pointer variable declaration and multiplication expression statement, they are both ident * ident :\

Precedence

14:34

Well actually, not even precedence. How is dereference ident * ident?

Learath2

It is my professional opinion that cloud is a scam

But companies are paying a lot for cloud resources

14:35

Without naming any company, but I can tell that all my clients are paying more than a million / month for cloud resources

Avolicious

But companies are paying a lot for cloud resources

Making more and more money for the corporations

Yeah, but running on-prem isnt sufficient for serving billions of requests / month

Most people aren't serving billions of requests per month

But my clients arent most people

14:37

They have the need to serve it, so cloud is sufficient for their use-case

And people were serving billions of requests per month way before aws existed

Learath2

Well actually, not even precedence. How is dereference ident * ident?

Wot

But yeah, you dont have to run everything on a hyperscaler... but the linked yt video from chiller mentions "how to run it on a 400 USD vps"

14:39

Even if he has a small team of a few people, its more than enough. But the larger the company gets, the more fine-grained permissions it needs

14:41

If your company now runs several stores across a country, its also bad to have it on-prem. If there is any problem within this infra, all stores are kinda offline. Cloud eliminates it by providing a flexible way to lift & shift

Avolicious

But yeah, you dont have to run everything on a hyperscaler... but the linked yt video from chiller mentions "how to run it on a 400 USD vps"

E.g. the startup I'm currently at would run at one one hundredth the cost if they weren't running on GCP

Learath2

E.g. the startup I'm currently at would run at one one hundredth the cost if they weren't running on GCP

Then your startup has a bad deal tho

They are preparing for a scale that will never come

We are google partners and we can give credits to our clients easily

14:42

Your google key account manager can add funds to your account without worrying about any consequences

There are some operations that will never need this level of hyperscaling. People need to understand their companies will never be the next tik tok

14:42

And you exist to serve the few who truly need it

Learath2

There are some operations that will never need this level of hyperscaling. People need to understand their companies will never be the next tik tok

But traffic is expensive on hyperscalers. Ingress mostly is pretty cheap, but egress is expensive as hell

14:43

So having on-prem and doing a hybrid solution can be more expensive

Avolicious

But traffic is expensive on hyperscalers. Ingress mostly is pretty cheap, but egress is expensive as hell

The compute isn't exactly cheap either

What are you using for your compute?

chillerdragon

deen also likes bare metal and beefy vps and ddnet.org uses jquery afaik

What choice do I have? If I used cloud for DDNet I'd have to pay 500k€/year

Also the only other alternative isn't on-prem. On-prem is a lot of work and will indeed end up more expensive in all the costs associated. But a lot of companies can run on a handful of off premises bare metal or even vps instances.

Why on server GER2 and FIN i have 200/400 ping ò.ò

Avolicious

What are you using for your compute?

Everything is running on GCP. I didn't see what sort of instances they are running. But I do see the final cost and how many requests we do

Learath2

E.g. the startup I'm currently at would run at one one hundredth the cost if they weren't running on GCP

Be the change you want to see. I moved our tests from AWS to Hetzner at work.

deen

Be the change you want to see. I moved our tests from AWS to Hetzner at work.

I set up our staging instance on a hetzner vm. It has been a much appreciated cost reduction

Learath2

Everything is running on GCP. I didn't see what sort of instances they are running. But I do see the final cost and how many requests we do

It depends on how you utilize GCP. You can do similar tasks with different tools, they are priced different

Paralix

Why on server GER2 and FIN i have 200/400 ping ò.ò

Problems with peering to Hetzner I guess

Tbh, thats a problem at hyperscalers. Cost transparency & FinOps in an easy way

14:48

You need basically a team of special trained GCP engineers/architects to get a good result which is likely to be cost efficient

Avolicious

Tbh, thats a problem at hyperscalers. Cost transparency & FinOps in an easy way

Figuring out the cost is also so hard to do in advance. With a tinfoil hat on I'd go so far as to claim it's on purpose

Learath2

Figuring out the cost is also so hard to do in advance. With a tinfoil hat on I'd go so far as to claim it's on purpose

I guess your startup doesnt pay this much, but if so, you get a key account manager at google. They can help you in this

Avolicious

You need basically a team of special trained GCP engineers/architects to get a good result which is likely to be cost efficient

With double tinfoil hats on I will make the claim that this entire industry just exists to make it as complicated as possible for non-specialists so there are jobs for specialists who pitch the solutions and so and so on creating an infinite money stream

Learath2

With double tinfoil hats on I will make the claim that this entire industry just exists to make it as complicated as possible for non-specialists so there are jobs for specialists who pitch the solutions and so and so on creating an infinite money stream

Well, there is a lot of money in this industry

Avolicious

I guess your startup doesnt pay this much, but if so, you get a key account manager at google. They can help you in this

We exist at just expensive enough that google won't care about us yet

I would say its multi-billion dollars a month

Learath2

We exist at just expensive enough that google won't care about us yet

Oh, well maybe we can cover you

I'll ask in the gcp chatroom, maybe someone knows who to ask

14:51

They love this tactic to keep you as a consumer

14:51

They basically set your account balance to a small amount, you are happy, google is happy

Avolicious

Oh, well maybe we can cover you

I'll ask in the gcp chatroom, maybe someone knows who to ask

Eh I don't care. It's not my money they are spending

14:54

We phased out an old elixir backend which helped a bit. But if this was my startup I'd definitely run it on just a single bare metal server

Maybe if I believed it'd scale into multiple countries I'd have a script to set up the server quicker

Terraform

14:54

No scripts

14:55

You can even manage bare-metal with terraform on hetzner

Ah that's cool. Last I used terrafrom it was basically aws, do, gcp, azure

You can use any provider, its vendor agnostic

14:56

but aws, do, gcp, azure and a few others are supported by hashicorp itself

14:56

Others are mostly community modules, but they work the same way

14:56

you can also write your own module to support your own provider if needed

Oh that's nice. I didn't know community modules were a thing

https://registry.terraform.io/browse/modules

14:57

Its actually pretty awesome

14:58

We are using it for any scale. SME and big companies

Back when I was last interested in this sort of automation ansible was still relevant

It is still

To be fair not exactly the same thing

There is a diff in Terraform vs Ansible tho

ansible sucks but still widely used

jxsl13

ansible sucks but still widely used

Why?

Avolicious

There is a diff in Terraform vs Ansible tho

Yeah very

automating a lot of customer configuration is not feasible

14:59

with ansible

jxsl13

ansible sucks but still widely used

What would be your goto in 2024?

and staying up to date

?

there is no open source alternative, tho

What would be your corposource alternative?

we do our own internal tool for that

And thats feasable

we had a lot of problems with ansible

Ah, rolling your own. My favourite. I approve

Learath2

Ah, rolling your own. My favourite. I approve

Somehow I believe that all DDNet devs have this mindset

So ansible is still the goto for deploying servers? (edited)

Just building everything their own, everytime

Avolicious

Somehow I believe that all DDNet devs have this mindset

Ofc. This is one of the last havens for free-minded developers

How complex are your setups?

15:02

I guess not that complex, no?

My team is doing the tooling for our datacenters, so rolling our own is an option

I would say that ansible is not working for 2 servers, but having 50 servers its highly recommended

Yeah if you are deploying 2 honestly you'd probably be wasting time writing the playbooks

Exactly

15:04

But writing own tooling is time consuming, you have to pay people to do it & to fix possible bugs

yep

15:05

who fixes your ansible bugs?

jxsl13

who fixes your ansible bugs?

Which bugs? Bugs in modules?

in external modules

15:05

for example

Either external maintainers or you & you PR it to help others

also costs money and time

15:06

might be way less than spinning your own

But less than building your own, maintaining it your own & its not helping others

it's not open source, yep

If it would, you need to afford it as company as well

helps our customere

15:07

s

You need active maintainers and contributors, otherwise its a dead project as soo many others

you need to afford that for sure to make such tools open source. Sadly our company does nearly nothing in the open source regard

But thats not sad. Its because they wont take this effort

15:09

Open Source isnt free as people often describe it

yep. companies sink billions into open source

You need to actively maintain the project. Either by answering issues/prs or contribute yourself -> company pays salary for employee

15:10

And mostly the open source community likes to say: Hey, just open source it, so you dont have to take care. Well you basically HAVE to take care

Otherwise it will get forked and your initial open source idea is a dead project blinks his eyes at ddnet

2

[ddnet/ddnet] Pull request opened: #8848 Fix map downloading overriding more recent map change messages

Following sequence of events was possible: 1. HTTP map download for map X is started (e.g. client initially connects to server). 2. Map change message for map Y arrives while downloading (e.g. vote for another map has passed on the server). 3. Client already has map Y and loads it directly without starting or resetting the map download. 4. HTTP map download task for map X finishes. 5. Map X is loaded, overriding the map Y that should be loaded. This is one cause for #2804. This coul...

github-merge-queue[bot]

[ddnet:master] 2 new commits

0597017 Fix map downloading overriding more recent map change messages - Robyt3 f754054 Merge pull request #8848 from Robyt3/Client-Map-Download-Reset-Cleanup - def-

16:39

[ddnet/ddnet] Pull request closed: #8848 Fix map downloading overriding more recent map change messages

Learath2

Well actually, not even precedence. How is dereference ident * ident?

I don't get it xd

Я русский

[ddnet/ddnet] Pull request opened: #8849 Set correct end line in twlang.py for strings with context

The script didn't set the correct end line for strings with context. This caused the ending line to be on the next strings context line, which it would delete and then the next line would have no context and therefore be considered unused. Closes #6971

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 (especially base/) or added coverage ...

MilkeeyCat

I don't get it xd

What is the ident on the left in a deref expression?

Learath2

What is the ident on the left in a deref expression?

I wasn't talking about deref expression

18:29

But about variable declaration

18:30

Like MyStruct* ptr;

Русские есть?

qwalreex

Русские есть?

#off-topic

either russians are trolling us, or we should move "Development" category below the "DDraceNetwork" one

they're too young to understand (edited)

zhn

they're too young to understand (edited)

who would have said that

1

github-merge-queue[bot]

[ddnet:master] 2 new commits

d40cc86 Set correct end line in twlang.py for strings with context - furo321 914867d Merge pull request #8849 from furo321/fix-twlang-endline - def-

19:04

[ddnet/ddnet] Pull request closed: #8849 Set correct end line in twlang.py for strings with context

19:04

[ddnet/ddnet] Issue closed: #6971 Translation updating script sometimes breaks translations with context

hi, ive recently installed ddnet on linux mint through zipped file (tar), cuz ddnet on package manager is outdated. When i unpacked it, it has no icons, how can i make icons appear?

linux magic is required

19:26

gotta google that

19:26

some config file in some folder is required for that

TimePause

hi, ive recently installed ddnet on linux mint through zipped file (tar), cuz ddnet on package manager is outdated. When i unpacked it, it has no icons, how can i make icons appear?

unpack all files

i did unpack all files, how can i unpack part of files? xD

unpack the .tar too

C in meraki stands for consistency

Image attachment

19:52

(i ran the same program 10 times and it gives different results each time)

[ddnet/ddnet] Pull request opened: #8850 Update Czech and Slovak translations for 18.5

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 (especially base/) or added coverage to integration test
[ ] 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-ad...

тут люди которые разговаривают на русском есть?

vladislav_durex

тут люди которые разговаривают на русском есть?

why did you decide to write it in #developer ?

cause it's the first channel they see

20:33

where they can write

No Russia?

MilkeeyCat

C in meraki stands for consistency

time to call big boi gdb

Я ВАС НЕ ПОНИМАТЬ

ИДИ В #off-topic

БЛЯ

20:34

пацаны

20:34

помогите

bluestripe

ИДИ В ⁠off-topic

20:34

ИДИ В ⁠off-topic

20:34

ИДИ В ⁠off-topic

я гонял с читом признаю свою ошибку забанили но разбан должен быть еще 3 часа назад

kek

Image attachment

vladislav_durex

я гонял с читом признаю свою ошибку забанили но разбан должен быть еще 3 часа назад

https://en.wikipedia.org/wiki/Coordinated_Universal_Time

MilkeeyCat

I wasn't talking about deref expression

You have finally arrived at a point that I actually don't know. I have only a clue. I think C doesn't allow expressions on their own. So this can't be an expression

feelsbadman

1

thonk

Image attachment

Where can I get this information as json on ddnet.org?

Image attachment

MilkeeyCat

thonk

Ok, I'm wrong. Then idk how this is handled for once

20:44

I'm also extremely drunk

20:44

Like about to pass out. So maybe tomorrow I'll know

Learath2

I'm also extremely drunk

do you want to read fun fun assembly

MilkeeyCat

do you want to read fun fun assembly

Show me. Idk if my in-brain assembler is still operational though

my neanderthal brain can't do anything even with gdb xddd

I remember C being context free, but perhaps it's not? If it's not context free then left side wouldn't be ident

20:47

It would be type * ident vs ident * ident then there is no conflict

Image attachment

20:47

it returns sometimes true, sometimes false

MilkeeyCat

it returns sometimes true, sometimes false

First non-deterministic machine let's goo :D

MilkeeyCat

Click to see attachment 🖼️

Too long for the drunk version of the assembler sorry

bye

zaerin

Where can I get this information as json on ddnet.org?

found it.

vladislav_durex

bye

Tis not an airport, you don't need to announce your departure

bye bye

Learath2

Too long for the drunk version of the assembler sorry

It's so joe over

MilkeeyCat

Click to see attachment 🖼️

Omfg, it moves 0 into r14b but later in cmp uses r14d

github-merge-queue[bot]

[ddnet:master] 3 new commits

c10b466 update czech translations for 18.5 - dobrykafe 29a205a update slovak translations for 18.5 - dobrykafe 60688bf Merge pull request #8850 from dobrykafe/pr-18.5-translations - def-

22:22

[ddnet/ddnet] Pull request closed: #8850 Update Czech and Slovak translations for 18.5

Exported 351 message(s)

Timezone: UTC+0