DDraceNetwork - developer

@Learath2

Ewan

that makes the whole setup process p much install msys -> install a few pkgs -> cmake -> win

if you're familiar with it, maybe you could write a guide

Ok

09:19

I don’t really want to write a forum post, is there somewhere else I should do it

you could add a section to the readme.md

wiki

1

Ok remind me later i'm doing awesome wave stuff

https://twitter.com/cjones3D/status/1678772727413616641

Chris Jones (@cjones3D)

#b3d Tendons https://t.co/7GKJ3svGHR

Likes

47896

Retweets

6325

09:39

@heinrich5991 have u xp compiling rust?

09:39

i know they use x.py

09:39

but it doesnt work for me

09:39

rust on  master via 🐍 v3.11.4 via 🦀 v1.70.0
❯ ./x check
Traceback (most recent call last):
  File "/data1/edgar/rust/x.py", line 50, in <module>
    bootstrap.main()
  File "/data1/edgar/rust/src/bootstrap/bootstrap.py", line 1113, in main
    bootstrap(args)
  File "/data1/edgar/rust/src/bootstrap/bootstrap.py", line 1070, in bootstrap
    build = RustBuild(config_toml, args)
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/data1/edgar/rust/src/bootstrap/bootstrap.py", line 505, in __init__
    self.build = args.build or self.build_triple()
                               ^^^^^^^^^^^^^^^^^^^
  File "/data1/edgar/rust/src/bootstrap/bootstrap.py", line 976, in build_triple
    return config or default_build_triple(self.verbose)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/data1/edgar/rust/src/bootstrap/bootstrap.py", line 259, in default_build_triple
    kernel, cputype, processor = uname.decode(default_encoding).split()
    ^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: too many values to unpack (expected 3)

09:39

xd

09:39

is it cuz python 3.11?

sounds like your target triple is wrong

❯ uname -a Linux gentoo 6.1.38-gentoo #1 SMP PREEMPT_DYNAMIC Mon Jul 10 08:28:17 CEST 2023 x86_64 AMD Ryzen 7 5800X 8-Core Processor AuthenticAMD GNU/Linux

heinrich5991

sounds like your target triple is wrong

how to check it

try uname -smp

❯ uname -smp Linux x86_64 AMD Ryzen 7 5800X 8-Core Processor

$ uname -smp
Linux x86_64 unknown

09:41

the line with the exception expects 3 words separated by spaces

09:42

I suppose that's where the problem comes from?

i guess

09:42

@Learath2 is this some gentoo thing? i remember something about uname

change the line to .split(maxsplit=2)

09:42

and make a pull request

ohhh

09:43

epic

09:43

im rust contributor

09:43

-p, --processor print the processor type (non-portable)

09:43

it says non portable

09:43

so i would say my -p is ok

09:44

my system i mean

it's only used on freebsd, anyway, according to the source

09:44

so .split(maxsplit=2) should not cause problems

seems to work

09:45

gotta fork rust

there's lots of pull requests you can send to rust-lang/rust

09:45

do it every time you find something lacking in the docs, e.g.

what does e.g smile mean?

09:46

confused xD

09:46

actually i never rly searched what e.g means

09:46

"for example"

09:46

ah ok

e.g. = for example (for weird latin reasons) (edited)

09:46

i.e. = that is (for weird latin reasons)

https://github.com/rust-lang/rust/pull/113616

Fix bootstrap.py uname error by edg-l · Pull Request #113616 · rust...

The x.py script fails with ValueError: too many values to unpack (expected 3) when uname -smp gives more than 3 words The error I got: ❯ ./x check Traceback (most recent call last): File "/d...

09:57

10:01

damn their ci is fast

10:01

AMD EPYC 7763 64-Core Processor

10:01

XD

<Jupstar> your first contribution to rust is python

sadge

10:03

wow rust compiled in 2 mins

10:03

i think

10:03

x build

<Jupstar> now say the same about llvm xd

xd

10:05

its amazing

10:05

the x tool

10:05

they setup env for either compiler library tools etc

10:05

and it asks if u want a configured vscode settings

10:06

@Jupstar ✪ when epyc server for ddnet CI

<Jupstar> sounds good, only annoying thing is that u have to setup vm or smth for security

true

10:10

i guess i didnt need to compile llvm 16

10:10

since i got it already

<Jupstar> pff llvm16

10:10

<Jupstar> non-hipster

xd

10:11

arch linux is on 15 still

<Jupstar> i am on 17 since like a month

i think arch linux lacks slots

10:11

gentoo portage slots

10:11

u can have 15 and 16 side by side

10:11

with proper deps

can’t u do that on nix as well

probs

https://media.discordapp.net/attachments/1033076370420535398/1128628894312960030/audiomanip_LKuoc17JiC.mp4

10:12

i love wave

<Jupstar> thanks for destroying my ears

you are welcome

10:16

anyone know of a list of known ddnet spammers/ddossers/bots

10:17

ppl to blacklist

10:17

just banned one from unique usa and wondering if there's a list i can ban all of

Ask noby

he was on fastcap and his name was green i thought we were gonna be buddies

He has lists

names to blacklist?

ips

sounds futile, unfortunately

10:17

ah

ok i ask noby

10:26

lmfao?????

Cool map idea

anyone know what happened

10:30

i think i have to tell timakro his instagib server is broken

Ewan

i think i have to tell timakro his instagib server is broken

zCatch?

hmm?

10:32

it was unique

10:32

in germany

10:32

i was getting rid of the spammers

instagib isn't exactly a game mode I think, but rather a gametype

10:35

deen recently found a zcatch related crash, so maybe it's that

<Jupstar> isnt that simply unused entities?

maybe. idk what are u talking about murpi, your words make sense but idk how they pertain to this

<Jupstar> is the design also broken?

that bug in my screenshot happened on the unique instagib server in germany. not ddnet affiliated (edited)

10:36

or related to zcatch

You said timakro's instagib server is broken. Instagib is a type of game, not a game mode. You can play instagib in zCatch or iCTF. Does that make sense?

probably ictf

ye

10:39

idk why u are talking about a crash rn

10:40

which i think is related to loading the game after ninja is voted in anyway. and probably on a different version of the game

Ah, I see the confusion. I didn't see the screenshot, only your message saying that x is broken.

oh lol

Oops

idk what that crash was but if it's the one i reported a day or 2 ago i think i had it before as well

10:45

but only recalled after and mistook it for a crash on the vanilla server. to which deen said there were no recent crashes or logs when i asked

10:45

that would explain a bunch

Ewan

Click to see attachment 🖼️

the map got corrupted

well yeah

10:55

i think maps should be hashed before they are loaded tho

<Jupstar> maps have a hash

10:59

<Jupstar> or what do you mean

def-

[ddnet-scripts:master] 1 new commit

1632188 Update server scripts - def-

<Jupstar> if the hash is wrong, then thats a problem after downloading a map, not when loading a map

the only reason i can see it being corrupted is network issues

<Jupstar> mh i doubt that

which would be checked after downloading. which is before they are loaded

11:02

lol

<Jupstar> it must be broken on the server already

11:02

<Jupstar> does the prediction also think the map is broken ? :D

idk im not there anymore

ws-client

<Jupstar> does the prediction also think the map is broken ? :D

yeah i think prediction makes u fall through the floor

<Jupstar> but the server resets your position?

i was in bed all day

11:13

My first time taking a leave to just chilling and napping

11:13

Didn't feel good tho lmao

yeah don't do that

11:13

bad for you

<Jupstar> yeah better continue working 18h a day

that's also bad

It's a paid leave so it was at least worth it

11:19

11:19

I just forgot to stop working till 4am last night and i have trouble sleeping so i just requested a day off in the morning

TsFreddie

I just forgot to stop working till 4am last night and i have trouble sleeping so i just requested a day off in the morning

how one forgets to stop working?

Because unity sucks and i have to make my own emoji atlas and hack my way through a fake font format and when I'm done it's already 4

11:48

Also ADHD probably

but

11:52

only work on work hours

11:52

if it takes longer it takes longer

11:53

or the company hires more

11:53

also if u dont get paid for those hours i find it stupid, and i dont think its cuz u need that job cuz living in a poor region

11:53

but its bad for u health ultimately

11:54

inspirational quote

11:54

#instagram

11:55

#lifestyle

11:55

Ryozuki

Click to see attachment 🖼️

im 14 and this is deep

ikr

12:11

but its true

Ryozuki

only work on work hours

i get bored at home

then become unbored

12:24

Ryozuki

@Learath2

Me fr

xd

ddnet is too laggy with my new internet

skill issue

since when did SDL3 is a thing

a few months

13:10

it's still preview

13:10

very unstable

interesting

13:18

i bought a game on steam which is SDL3

13:18

so i was confused

13:18

wow that is a bad idea

That's hipster

13:22

I like

13:23

As if any sdl2 was ever stable xd

there are degrees to everything

13:23

lol

~~like how i have a master degree~~

weird flex but ok

and it is 37 degrees outside even though it is 9 in the evening

13:25

that's 99 degrees for you americans btw

Ewan

very unstable

where did u hear very unstable

i experienced it

13:27

libinput is completely broke on windows

13:27

build system and conflicts

ah windows

13:27

on linux?

13:27

the only one and only

skill issue

windows is a platform they target and it doesn't work on there so it's an SDL issue (edited)

13:27

words

skill issue

excels and powerpoints

@heinrich5991 regarding the rust pr i made

13:28

im very confused

13:28

apparently another did the same fix

13:28

but mine was accepted

13:28

https://github.com/rust-lang/rust/pull/113616

Fix bootstrap.py uname error by edg-l · Pull Request #113616 · rust...

The x.py script fails with ValueError: too many values to unpack (expected 3) when uname -smp gives more than 3 words The error I got: ❯ ./x check Traceback (most recent call last): File "/d...

13:28

https://github.com/rust-lang/rust/pull/113571

bootstrap: Don't break on uname -p != unknown by clubby789 · Pu...

Fixes #111562 (comment)

13:28

i guess cuz theirs failed ci, but for unrelated reasons

you're just better

13:30

i do wonder how the guy feels tho, it's exactly the same change.

lmfao

https://github.com/rust-lang/rust/pull/113571#issuecomment-1630946289 like this conversation is pretty funny

ryo’s has a real description

13:32

that’s prob ig

13:32

it

13:32

words

TsFreddie

i do wonder how the guy feels tho, it's exactly the same change.

the other person is a contributor already, they'll be fine

my guess is that was the decision (edited)

13:36

or maybe simply they dont want to rerun ci

13:36

xd

I don't think there was a conscious decision

13:44

these were separate reviewers

13:44

yours just got in faster, then the other was obsolete

13:49

new rust drama blog post incoming

13:49

"pr approved over another pr" (edited)

13:49

13:50

god i cant type

13:51

i also joined zulip

13:51

the open source slack

13:51

looks great

13:51

a bit confusing at first

13:51

https://rust-lang.zulipchat.com/#all_messages

yea, I also liked it

13:51

I wondered whether it'd be a good fit for ddnet

13:51

but I guess github issues is good enough

true

13:52

i wondered if it could replace discord

13:52

but discord is too easy and used byeveryone

it'd be nice if the discord could be mirrored to some open platform

ohh sweet

but I haven't really found a solution for that

open source slack sounds awesome

not sure if it's slack tbh

is it the same API-wise or is it just similar

it sounded more like a replacement for forums

Ewan

is it the same API-wise or is it just similar

definitely not API-wise

well i heard people say its like the open source slack

so like a self-hosted discord?

self hosted slack

13:54

this doesn't look like slack to me

xd

13:54

but it serves the same function

in slack you have a normal chat with optional threads

13:55

this is chat divided by topics

i wonder if zulip has it

idk

or its just how rust does things

Ryozuki

i wondered if it could replace discord

it wont

14:26

discord is too big to be replacrd

it'll be replaced at one point

14:27

all these proprietrary systems are

look at twitter

where are the mastodon gang

Ryozuki

look at twitter

i barely see a difference with twitter threads is a completely different demographic

is it bad to take melatonin every day

ask ur doctor

Voxel

i barely see a difference with twitter threads is a completely different demographic

twitter won't be gone for a couple of years

14:41

it's just so huge

Signs of melatonin overdose include excessive sleepiness, vomiting and trouble breathing. Other side effects of both low and high doses of melatonin can include headaches, excessive sleepiness, blood pressure changes, gastrointestinal problems, changes in other hormone levels and mood problems, McCall says

14:41

since u can get overdose i guess yes

14:41

it can be bad

sleeping is too hard

go to gym

14:43

im no doctor but i guess sleeping can be bad if u simply have a unhealthy lifestyle xd

14:43

eat bad

14:43

do nothing

14:43

etc

14:43

overwork

I walk 30 minutes to work does that count

no xd

18h - 30min or + ? Xd

Jupstar ✪

18h - 30min or + ? Xd

if i work 18h i probably was sleeping at work

14:55

or just didn't go home that day

Ryozuki

https://rust-lang.zulipchat.com/#all_messages

this looks absolutely boring LOL

Jupstar ✪

18h - 30min or + ? Xd

also it was 30 minutes one way so probably an hour in total for commute

idk i think id stick with discord

some gifs are too big to be fun

@Jupstar ✪

Ryozuki

@Jupstar ✪

Robyt3

[ddnet/ddnet] Pull request opened: #6837 Fix editor hotkeys triggering while some editboxes are active

Replace CEditor::m_EditBoxActive which only works with editboxes created from the editor with CLineInput::GetActiveInput which also works for editboxes which are created by generic UI functions, e.g. the value selector editboxes of color pickers.

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 (especially base/) or added coverage to int...

Ryozuki

@Jupstar ✪

texas sharpshooter fallacy troll

Sometimes working with rust fills me with such primal rage, why on gods green earth am I not allowed an empty array without an Option being involved?

18:03

I mean I know why, but still pisses me off

18:07

let me have my uninitialized memory please aaaaaaaa

Learath2

let me have my uninitialized memory please aaaaaaaa

? u can

18:10

maybeuninit

18:10

https://doc.rust-lang.org/std/mem/union.MaybeUninit.html

MaybeUninit in std::mem - Rust

A wrapper type to construct uninitialized instances of T.

it's unsafe, meaning if you use it you'll get lynched

nah

18:11

specially for this use case

18:11

initializing a array

18:11

its well known

18:11

but option is safer

18:11

and usually no overhead

18:11

depends on the type

18:11

https://doc.rust-lang.org/std/mem/union.MaybeUninit.html#initializing-an-array-element-by-element

MaybeUninit in std::mem - Rust

A wrapper type to construct uninitialized instances of T.

Learath2

Sometimes working with rust fills me with such primal rage, why on gods green earth am I not allowed an empty array without an Option being involved?

use case?

18:15

the use case I encounter often is a byte buffer for the read trait

good to know ur using rust instead of C tho

Something far simpler https://docs.rs/twilight-http/latest/twilight_http/request/channel/message/create_message/struct.CreateMessage.html#method.attachments I want to create a message inside a closure but I return it outside, so the backing storage for the attachments must be on the outside scope

CreateMessage in twilight_http::request::channel::message::create_m...

Send a message to a channel.

coming to the light

Ryozuki

good to know ur using rust instead of C tho

Every second I use this language I miss the flexibility of C and C++

rly? i dont

18:17

that feeling of safety

18:17

@Learath2 basically you're trying to initialize an array using a closure?

Ryozuki

that feeling of safety

I honestly hate it, it feels like there is this drill sergeant constantly watching over me. I know when I've initialized my variables I'm not an invertebrate

std::array::from_fn

Learath2

I honestly hate it, it feels like there is this drill sergeant constantly watching over me. I know when I've initialized my variables I'm not an invertebrate

yes yes ur a god

heinrich5991

@Learath2 basically you're trying to initialize an array using a closure?

Yes but the array must be declared on the outside scope

Ryozuki

std::array::from_fn

oh, cool, I didn't know stdlib had this already

18:18

I think @Ryozuki posted the solution

Ryozuki

yes yes ur a god

Don't need to be anything beyond simple human to lift your head from the line you are looking at to a couple lines above and see if you initialized something

18:19

But thanks

heinrich5991

oh, cool, I didn't know stdlib had this already

this is how i init [x; N] arrays

18:19

thagt are not copy

back in my day™ we didn't have these fancy const generics :p

18:19

that's why I don't know these well yet

oh i see

18:19

well rust is getting better

18:19

and learath is learning it

18:19

its 2023

18:19

heinrich5991

I think @Ryozuki posted the solution

This helps initialize an array using a closure, I need to initialize and use the array inside the closure and return a CreateMessage<...> instead

18:20

It's just that the backing storage for the array needs to live the same duration as the CreateMessage<...> so they need to be declared in the same scope

ah

but I can't yet initialize it because I don't know if a) I'll need it b) it's contents

you want to have a struct that has two fields, one pointing into the other? (edited)

CreateMessage<...> holds a reference to an &[Attachment]

yes

what else

Let me write a simple snippet so we can talk over that instead

ye

18:27

@heinrich5991 from_fn is from 1.63.0 (edited)

18:28

so pretty recent

18:28

we cant use it in ddnet code kek

double irc ping poggers

tbf we're not using a lot of rust in ddnet yet

18:28

the only thing we do currently is the mastersrv, unfortunately

18:29

I'm working on doing more in rust

18:29

quic!

nice

18:29

i want to code more in ddnet but idk lazyiness

18:29

and doing stuff in rust is more work cuz there is not a lot of cxx layer built yet

yes

18:30

maybe the cxx approach was the wrong one :/

https://ideone.com/h1I1a7 this is the idea, as you can see this doesn't quite work as the array created on line 12 is a temporary that is dropped too early

Ideone.com

Ideone is something more than a pastebin; it's an online compiler and debugging tool which allows to compile and run code online in more than 40 programming languages.

18:31

I recreated it with a match statement, but closures suffer from the exact same issue

18:32

My inferior C brain immediately tells me that this issue is solved by allocating the backing storage before the block along with the msg, filling it in the match/closure, and then using that

maybe you can do a let attachments; in the beginning

https://ideone.com/1g0ZJB this is the best I could come up with

Ideone.com

Ideone is something more than a pastebin; it's an online compiler and debugging tool which allows to compile and run code online in more than 40 programming languages.

and then in the match, you do attachments = [your_attachment];

18:35

and then pass &attachments to the function (edited)

18:35

that should work

Aha, that does work, what is this magic called?

it's no magic :p

18:37

it's basically doing what you wanted to do, having an uninit variable

attachments would be immutable no?

yes, you can only assign it once

as long as u init the let in every branch it works

I did not know you were allowed an assignment later huh

Ryozuki

as long as u init the let in every branch it works

no, that's not necessary

18:38

you only have to init it before you use it

everywhere where u use it*

18:38

ye

Learath2

I did not know you were allowed an assignment later huh

its initialization

18:38

iirc

It isn't mentioned in the book and I don't read enough rust code I guess

18:39

Nice to know

I don't know how I learned that

18:40

perhaps just like this

https://doc.rust-lang.org/reference/variables.html the reference does mention it however

Variables - The Rust Reference

https://doc.rust-lang.org/reference/statements.html#let-statements

Statements - The Rust Reference

let else is the best thing to happen in a while

let else is indeed nice

18:42

haven't used it yet though. so who am I to judge that ^^

i also never used it

18:42

if let else yes

18:43

let is super powered

18:43

its one of the things that makes me hate old langs

18:43

pattern matching

18:43

The pattern is most commonly a single variable, which means no pattern matching is done and the expression given is bound to the variable. Apart from that, patterns used in let bindings can be as complicated as needed, given that the pattern is exhaustive. See the Rust book for more information on pattern matching. The type of the pattern is optionally given afterwards, but if left blank is automatically inferred by the compiler if possible.

If y'all enjoy the usual rule of never having early returns it's normal that you don't use it

i use early returns sometimes

18:44

actually many times if u count ?

18:45

let (Ok(i) | Err(i)) = [1, 2, 3].binary_search(&2);

18:45

fn f((Ok(i) | Err(i)): Result<i32, i32>) {}

The else block has to diverge, so it kinda only has a use for guards and stuff

match value {
    Some(p @ 2 | 3 | 5 | 7) => println!("{p} is a prime"),
    Some(sq @ 0 | 1 | 4 | 9) => println!("{sq} is a square"),
    None => println!("nothing"),
    Some(n) => println!("{n} is something else"),
}

18:45

binary_search is funny cuz it returns a result

18:46

Result<i32,32>

18:46

If the value is found then Result::Ok is returned, containing the index of the matching element. If there are multiple matches, then any one of the matches could be returned. The index is chosen deterministically, but is subject to change in future versions of Rust. If the value is not found then Result::Err is returned, containing the index where a matching element could be inserted while maintaining sorted order.

This bot is so bad, I'm ashamed of having written it

xd

18:48

what bot

18:48

im interested

18:48

i made RyoChan

18:48

but it just has 1 command

ah right tomorrow last exam

18:49

for learath

18:49

then bot release

Learath2

This bot is so bad, I'm ashamed of having written it

no open source?

18:49

18:50

Learath2

If y'all enjoy the usual rule of never having early returns it's normal that you don't use it

I like early returns

Ryozuki

no open source?

it is open source

https://github.com/rust-lang/rust/pull/113624

Rollup of 8 pull requests by matthiaskrgr · Pull Request #113624 · ...

Successful merges: #113353 (Implement selection for Unsize for better coercion behavior) #113553 (Make Placeholder, GeneratorWitness*, Infer and Error unreachable on SMIR rustc_ty_to_ty) #113598 (...

18:52

rollup of my pr

18:52

im rust contributor now

18:52

https://github.com/Learath2/discordbot don't look at it though, it's actually embarrasing code

GitHub - Learath2/discordbot

Contribute to Learath2/discordbot development by creating an account on GitHub.

@Learath2 in my book if it doesnt pass clippy with pedantic its embarrasing

18:53

so ye the bar is high

and the commit I'm about to push, don't judge me for it, I just need to fix the bug mods are complaining about before they kill me

that's a stupid rule IMO

18:53

anyone can push new clippy lints

#![deny(missing_docs)]
#![deny(warnings)]
#![deny(clippy::nursery)]
#![deny(clippy::pedantic)]
#![deny(clippy::all)]

18:53

add this

Ryozuki

so ye the bar is high

I have warnings even without clippy

@heinrich5991 i knew u would say this, but i dont care

18:53

and im half joking

18:53

pedantic is a lot

18:53

but atleast make clippy pass

and they only get reverted once enough people announce that the lint is actually not improving stuff

(i unjokingly enable pedantic in my projects tho)

18:54

so far the only annoying ones i found is too complex types and too many arguments (edited)

whys it called clippy

cuz windows clippy

18:54

its a linter

lmfao

18:55

the cult of anti windows praises the windows mascot

18:55

many rust users use windows

18:55

its a separate cult

18:55

actually im not 100% sure its cuz windows tho

i always saw them as the same tbh

wasn't clippy ms office, not windows?

18:55

is that a matrix bot gone wrong?

Voxel

i always saw them as the same tbh

idk what u say, the only rust community u seen is us here xd

heinrich5991

wasn't clippy ms office, not windows?

windows xp?

not AFAIK

i saw that

i missed it

Ryozuki

idk what u say, the only rust community u seen is us here xd

fair but i like it this way

Ryozuki

but atleast make clippy pass

e.g. check this wonderful lint that introduced bugs in people's projects: https://github.com/rust-lang/rust-clippy/pull/9484

Implement manual_clamp lint by Xaeroxe · Pull Request #9484 · rus...

Fixes #9477 Fixes #6751 Identifies common patterns where usage of the clamp function would be more succinct and clear, and suggests using the clamp function instead. changelog: [manual_clamp]: Impl...

@Learath2 atleast u using sqlx

heinrich5991

e.g. check this wonderful lint that introduced bugs in people's projects: https://github.com/rust-lang/rust-clippy/pull/9484

if a buggy lint got into pedantic its a misrelease by them, iirc the experimental lints are on nursery

18:58

some may have false trues in stable but its ok

this was a warn-by-default lint

i dont see the relation

this is a case where blindly following clippy's suggestions introduced bugs in people's code

ok but i still dont see the problem

with clippy's default settings

software can have bugs

okay

also many ppl use clippy check

19:00

on ci

I just want to say that you should critically evaluate the lints

instead of a bare check

19:00

i do

19:00

so far they make lot of sense

if even bugs can get through, surely bad suggestions can also get through

i think u should be more leniant

I thought I was being more lenient than you are on this

19:01

you're suggesting a "you need to fix all clippy warnings" policy (edited)

i always seen u as a anti clippy guy

19:01

yes, so far it worked good for me

and I'm providing a counterpoint

19:01

why fixing all clippy warnings is not always a good idea

19:01

as it even introduced bugs in the past

19:02

code linting tools are good, clippy is one of them

which get fixed and is fine

19:02

99% of cases are good to fix

please don't make up percentages

19:02

most of the cases are good to fix

19:02

I agree

19:02

but you should still think while doing so

19:02

if a suggestion doesn't make something better, don't do it

i was about to make a joke but better not xd

19:04

i aproach these tools like this: im certainly they will in most if not all cases now better than me, obv i also check if they make sense, and i read many lints descriptions

19:04

and i find them reasonable

I also read lint descriptions

19:04

and when I don't find them reasonable, I don't apply them

19:04

seems like what you're doing, too. except we have different levels of what we deem sensible

yea

but I wouldn't exactly call my stance "more extreme" than yours ^^

lol i saw that

matrix spam?

ye

19:05

briefly xD

how do I check if I'm compiling for windows, in meson?

never rly used meson

19:06

host_machine.system()

19:06

i think

19:06

https://mesonbuild.com/Compiler-properties.html

19:07

https://mesonbuild.com/Cross-compilation.html

thanks

that was 1 quick google tho

19:08

what u doing with meson?

19:08

porting ddnet?

hmm, don't wanna reveal other identities

19:08

no

a ok

typical need to define NOMINMAX

19:09

on windows

19:10

because the windows.h header thinks it's a good idea ot define min/max macros, breaking the c++ stdlib, among others

xd

Ryozuki

@Learath2 atleast u using sqlx

fwiw I only used it because I was interested in it's Any interface, turns out it's not so good and it really isn't even a great idea to pursue that

i never liked multi sql support

19:19

its better to focus on one

19:19

and use its neat features

19:19

mainly psql

Learath2

fwiw I only used it because I was interested in it's Any interface, turns out it's not so good and it really isn't even a great idea to pursue that

do you know anyhow btw? (edited)

Next time I might try tokio_postgres

color_eyre*

Learath2

Next time I might try tokio_postgres

noo just use sqlx with postgres

would make code like this easier to write:

    let Ok(msg) = msg else {
        return Err(Error("CME"));
    }

compile time checked queries

heinrich5991

do you know anyhow btw? (edited)

Yeah, I use it in newer stuff I do, that bot is left from the very early times when I was learning Rust. So I also used it as a playground where I messed around with doing errors on my own without anyhow

color_eyre is a fork of anyhow with better end user errors

19:21

https://docs.rs/color-eyre/latest/color_eyre/

color_eyre - Rust

An error report handler for panics and the eyre crate for colorful, consistent, and well formatted error reports for all kinds of errors.

in what way "better end user errors"?

Ryozuki

compile time checked queries

I never quite figured out how best to package that for a project btw, I just set DATABASE_URL and call it a day

19:22

I think there is a way to generate a json file with the schema

Learath2

I never quite figured out how best to package that for a project btw, I just set DATABASE_URL and call it a day

ye thats how i do it too but u can also do the json

19:22

u need to install the sqlx cli tool

Ryozuki

Click to see attachment 🖼️

I also saw the documentation, it's not clear to me why this is "better end user errors"

which btw, its a rly good tool to manage migrations

I was hoping for some small summary

heinrich5991

I also saw the documentation, it's not clear to me why this is "better end user errors"

It looks pretty? idk

u need the tool with the feature enabled

eyre adds backtraces, I'm guessing color_eyre adds colors to eyre

ye

19:23

backtrace full looks pretty

19:23

beware tho

19:23

in some scenarios due to eyre adding backtraces, there is a issue with lifetimes

I can imagine backtrace full looking extremely ugly in async programs

i found this in some async contexts

19:23

returning some errors

19:23

i forgot what i did

is ddnet async

it does some stuff in threads

19:24

but it doesnt have a async runtime

it's doing manual async

Voxel

is ddnet async

tokio, a async runtime and executor, uses a work stealing scheduler to run ur routines

19:24

tasks*

19:25

learning to make ur async runtime can be fun

19:25

i have to do it

19:25

i only dabbled into it when i was doing ryoOS

Async usually implies a runtime making guesses about where best to run your tasks. It'll start a bunch of threads and as you submit tasks it'll assign them to threads

https://os.phil-opp.com/async-await/

Async/Await | Writing an OS in Rust

In this post, we explore cooperative multitasking and the async/await feature of Rust. We take a detailed look at how async/await works in Rust, inclu…

Learath2

Async usually implies a runtime making guesses about where best to run your tasks. It'll start a bunch of threads and as you submit tasks it'll assign them to threads

the neat part is spawning a task is cheap compared to spawning a thread

I'm still very skeptical about how good of an idea this async thing is, it does make writing async code trivial but how can it be as performant idk

isnt tokio pretty perfomant

19:27

it beats go which is designed for this, in some areas

Locality principle is extremely useful in squeezing out performance, but if my task moves from core to core I might get hundreds of cache misses for no reason

I think postgresql also wanted to move to some async model

i forgot the blog post

Activating VS 17.6.3 — why does this take >1 second?

Learath2

Locality principle is extremely useful in squeezing out performance, but if my task moves from core to core I might get hundreds of cache misses for no reason

i think u can spawn a non movable task iirc

what exactly needs 4'000'000'000 clock cycles there?

Ryozuki

isnt tokio pretty perfomant

I'm talking about async runtimes in general versus manually threading your application thinking about where a thread is useful

https://docs.rs/tokio/latest/tokio/task/struct.LocalSet.html

LocalSet in tokio::task - Rust

A set of tasks which are executed on the same thread.

Learath2

I'm talking about async runtimes in general versus manually threading your application thinking about where a thread is useful

postgresql is thinking about it

heinrich5991

Activating VS 17.6.3 — why does this take >1 second?

What is this?

heinrich5991

postgresql is thinking about it

Is there some discussion somewhere I can read? In their case performance is critical so they might have some of the concerns I have and there might be some answers

Learath2

Is there some discussion somewhere I can read? In their case performance is critical so they might have some of the concerns I have and there might be some answers

lwn

Learath2

What is this?

meson getting the cl compiler and link linker from the environment

is meson made in python

yes

@Jupstar ✪ justatest

19:30

when something like meson in rust

build systems are an annoying problem

19:31

soo many different platforms

yesterday i found this neat crate https://crates.io/crates/dll-syringe

19:31

wanted to try if i can port a mod platform from c# to rust

19:31

for ffxiv

Ryozuki

when something like meson in rust

build systems have to be slow, so better use the slowest language available

heinrich5991

meson getting the cl compiler and link linker from the environment

Maybe it's trying the compiler out? Most build systems check if it can create an executable and stuff

yo the spammer is back

I'd guess it's an extremely slow batch file from msvs

wrong server sry

but I guess I'll debug this for a minute now

i thought heinrich was in a asian timezone

19:38

but maybe he is american

19:38

in*

19:38

idk why i thought asian

it's VC/Auxiliary/Build/vcvars64.bat

19:38

that's taking >1 s to execute

r u on windows?

heinrich5991

typical need to define NOMINMAX

.

19:39

I have a windows VM

19:39

to check that stuff compiles on windows

ah ok

19:39

well one could also think @Learath2 is in a american timezone given his active hours

19:39

19:40

hi deen

1

19:40

wen c++20

19:41

ive seen some news around c++23

Ryozuki

i thought heinrich was in a asian timezone

heinrich is on the space station, that explains his always changing timezone

i failed to get it working

1

cyberFighter

wrong server sry

i mean i guess itw as the right server? the rainbow with random generated names guys

19:42

joined our srv

@deen this reminds me, u soon wont be here at all right?

19:43

can heinrich or learath make releases

19:43

multiview looks pog

1

19:44

@heinrich5991 https://blogs.windows.com/windows-insider/2023/07/12/announcing-windows-11-insider-preview-build-25905/

Announcing Windows 11 Insider Preview Build 25905

Hello Windows Insiders, today we are releasing Windows 11 Insider Preview Build 25905 to the Canary Channel. We are releasing ISOs for this build – they can be downloaded here. Starting with to

19:44

https://news.ycombinator.com/item?id=36700076

pjmlp

Windows 11 Insider Preview Build 25905 brings a bit of Rust in the kernel

19:45

Rust in the Windows Kernel

>

Rust offers advantages in reliability and security over traditional programs written in C/C++. This preview shipped with an early implementation of critical kernel features in safe Rust. Specifically, win32kbase_rs.sys contains a new implementation of GDI region. While this is a small trial, we will continue to increase the usage of Rust in the kernel. Stay tuned!

>

[We are beginning to roll this out, so the experience isn’t available to all Insiders in the Canary Channel just yet as we plan to monitor feedback and see how it lands before pushing it out to everyone.]

19:46

im dying of heat

19:46

it says it rains but its fake

With rust this bug wouldnt have happened

Ryozuki

can heinrich or learath make releases

no, not right now

19:48

i.e. we don't have deen's setup

We could technically make a release I guess, it would just take a lot of effort since we don't have it set up

i think for ddnet future proof sake

19:49

we need to standarize this

We have access to everything needed

i want to be able to do a release on my server (for myself for example)

Ryozuki

we need to standarize this

There is a release script if you want to do it yourself

it has many hardcoded paths iirc

where is deen going?

he isnt going anywhere

(It might be nice to do this in docker actually, I did attempt it once but then it was way too much work)

he will just be busy

19:50

xd

19:50

he mentioned it on this channel but idk if i should repeat it

u can link to it

19:50

or i can just scroll

that would be the same as repeating

19:50

i meant it more as in privacy

19:51

why did u delete

old screenshot

19:51

fulscren bug

xd

pro racer

so pro he does 2 at once

why dummy on race tho

19:52

maybe he is testing

it's just Green with their name changed LMFAO

19:52

the spammer

lol

Malicious

Ryozuki

we need to standarize this

I think that's "only" a question of effort. I don't think anyone would be against it

<ChillerDragon> when finally open source ddnet scripts

20:41

<ChillerDragon> not this half baked some scripts copied over into a git repo

what

heinrich5991

i.e. we don't have deen's setup

Maybe you can make CI scripts (GitHub Actions and/or GitLab CI) producing suitable packages? I see it won't cover everything (e.g. Steam) but maybe the way to overcome this is a better (and more accessible) automation?

Ryozuki

@deen this reminds me, u soon wont be here at all right?

it's not like I'm dying

help math

// samplerate is 48000, can be any positive number tho
double running = 5; // (0, 5)
const double growthConstant = 1. + (1. / (samplerate / 3)); 
// (1, 5 * growthConstant); (2, (5 * growthConstant) * growthConstant)
for (int i = 0; i < samplerate; i++) { // Test every X between 0 and 48000
  audiodata[i] = running;
  running *= -growthConstant;
}

21:44

why does this create a curve where f(x) = 100 (edited)

21:44

what's special about 3 and 5 in this case

I don't see f(x) = x in that image

21:48

where do you plug in the 3, the 5?

i've been kinda trying to turn this into an equation but the whole growthconstant thing eludes me

21:51

5 is the y intercept, 3 is the divisor for the growth value (edited)

21:53

sry i forgot it's not f(x) = x it's f(x) = 100

I also do not see the constant function f(x) = 100

Top of the 'graph' is 100

>>> (16001/16000)**48000 * 5
100.41827035443303

21:56

>>> (1 + 1 / (48000 / 3))**48000 * 5
100.41827035443303

wouldn't that be 100 if not for imprecision

21:57

idk

that's not exactly 100, no

so this is just a coincidence

yes

i see

ah

21:59

you're approximating 5 * e**3 with this formula btw

22:00

(1 / 1 + (high_number / x))**high_number tends to e^x as high_number tends to infinity

hmm

22:02

i wonder how i would write it to make it not an approximation

short answer: 2.718281828459045

22:14

long answer: you can't represent e exactly, e is an irrational number

i see

@Learath2 are empty compilation units forbidden by the C standard?

23:14

why is C2183 a thing and why is it only being thrown when you disable MS extensions?

an empty file? It should be fine iirc

23:15

Atleast in c99

23:18

Well let me clarify, atleast I don't remember reading anything that forbids it

https://learn.microsoft.com/en-us/cpp/error-messages/compiler-errors-1/compiler-error-c2183 (edited)

Compiler Error C2183

Learn more about: Compiler Error C2183

23:21

thanks, ms?

MS is very well known for their non compliant C compiler, no?

I need to enable /Za for stdlib.h not to define min/max macros

23:21

/Za means "turn on standard compliance"

23:21

but /Za also turns on this error

23:21

very funny

Huh, interesting, maybe I remember wrong, let me take a look at the standard

bencie

[ddnet/ddnet] Issue opened: #6838 Killfeed not showing deaths in teams

If a tee dies in an unlocked team mid run, they will not show up in the kill feed. https://github.com/ddnet/ddnet/assets/78111412/4060fe48-5795-4101-a5ed-47136c726a72

why can’t you define NOMINMAX before including Windows.h?

heinrich5991

/Za means "turn on standard compliance"

Yeah I don't see any issue, 5.1.1.2 even has a mention about "A source file that is not empty" implying if it is empty it's fine (IMO)

Ewan

why can’t you define NOMINMAX before including Windows.h?

it comes from stdlib.h, not windows.h for whatever reason

23:26

took me quite some time to debug where this stupid macro is coming from

hmm

heinrich5991

it comes from stdlib.h, not windows.h for whatever reason

gcc -E?

23:26

msvs should have something similar

what if you just -D?

https://stackoverflow.com/q/26541150/870079

Warning: ISO C forbids an empty translation unit

In a header file I have the following code which gives me the error in the title, while trying to link. #ifndef BOOLEAN_H #define BOOLEAN_H #ifndef FALSE #define FALSE 0 #endif #ifndef TRUE #define

Learath2

gcc -E?

doesn't help, that strips all macros

Ewan

what if you just -D?

-D what?

heinrich5991

doesn't help, that strips all macros

but it should show where includes start

heinrich5991

-D what?

NOMINMAX

Learath2

but it should show where includes start

doesn't in cl.exe

23:27

the equivalent flag

sry, gcc -D. dunno what it is in cl

Ewan

NOMINMAX

stdlib.h doesn't check for NOMINMAX

hmm

also -D//D

it's just /D on cl?

-D also works

23:28

yes

heinrich5991

https://stackoverflow.com/q/26541150/870079

@Learath2 gcc apparently also thinks so; seems like this might actually be something

heinrich5991

-D also works

they are the same

bors[bot]

[ddnet:master] 3 new commits

da2da7b Fix editor hotkeys triggering while some editboxes are active - Robyt3 6cfdbd9 Handle editor input events in OnUpdate instead of OnRender - Robyt3 573d9ca Merge #6837 - bors[bot]

23:44

bors[bot]

[ddnet/ddnet] Pull request closed: #6837 Fix editor hotkeys triggering while some editboxes are active

heinrich5991

@Learath2 gcc apparently also thinks so; seems like this might actually be something

I actually don't see it, weird

23:46

Maybe it's an indirect result of another rule because I took a look at all the sections that I think governs translation