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 2024-03-13 00:00:00Z and 2024-03-14 00:00:00Z

[ddnet/ddnet] Issue closed: #3994 Feature: allow players to bind commands on connect

00:27

[ddnet/ddnet] Issue closed: #3990 Ninja+bullet in switch layer (off) are predicted incorrectly

<ChillerDragon> @Ryozuki are you saying edlang has a release, a web page, a project creation tool and a debugger but it can not do hello world?

[ddnet/ddnet] Issue reopened: #3994 Feature: allow players to bind commands on connect

why i cant use prediction margin 0?

01:35

only 1 or higher

[ddnet-maps:master] 1 new commit

29fe4b5 M Caverns, M The Floor is Lava - ddnet-maps

ws-client

<ChillerDragon> @Ryozuki are you saying edlang has a release, a web page, a project creation tool and a debugger but it can not do hello world?

yes xd

06:54

https://twitter.com/phoronix/status/1767695611808158035

Phoronix (@phoronix) on X

OBS Studio 30.1 Released With AV1 Support For VA-API & @PipewireP Video Capture https://t.co/F3m5BK651m

06:54

with this obs is better on linux now than on windows

06:54

legit

06:54

u cant do pipewire stuff in windows

Matodor

@TsFreddie ddnet-pvp has any config option for disable room settings changes for spectators?

no, but i think you can just lock the room/team to prevent spectator joining in

07:41

and they can still spectate using /spec player

edlang now supports having both lib.ed and main.ed

13:11

and builds a shared library and a binary pog

13:12

it also doesnt require a top level module

13:13

pub exported fn hello_world(a: i32) -> i32 {
    return a;
}

what do u think about the exported keyword? it disables name mangling and makes it externally visible

13:13

the pub is only within edlang

13:13

or maybe i should just copy rust? pub extern "C" fn

13:13

but i doubt ill support other than extern "C" itself

13:14

Image attachment

13:14

owo

Ryozuki

pub exported fn hello_world(a: i32) -> i32 {
    return a;
}

what do u think about the exported keyword? it disables name mangling and makes it externally visible

Why do you do mangling at all?

@Learath2 because i have modules

13:15

namespaces if u will

13:15

my mangling is rly dumb tho rn

13:16

if i mark it not exported

13:16

the name will be

13:16

define private i32 @"hello_world@0@3"(i32 %0) !dbg !2 {

13:16

0 is like the file id

13:16

3 is the definition id

13:16

which increases when it finds definitions like functions etc

13:16

2 = distinct !DISubprogram(name: "hello_world", linkageName: "hello_world@0@3", scope: !3, file: !1, line: 4, type: !4, scopeLine: 4, spFlags: DISPFlagDefinition, unit: !0, retainedNodes: !7)

13:16

good debug info kek

Do you have overloads?

@Learath2 i will have generics

13:17

no overloads for now

If no overloads, why do you need the definition id?

true

13:17

but definition id could help in generics

13:17

xd

13:18

cuz multiple definitions

13:18

with diferent types

Yeah, but I think you should move to a more deterministic mangling scheme

yeah

13:18

it was just smth quick for now

13:18

to not get symbol conflicts

Understandable

@Learath2 any known algorithm?

13:19

also i think the linker doesnt support the @ symbol

Ryozuki

or maybe i should just copy rust? pub extern "C" fn

I don't think rust disables "name mangling" for extern "C"

oh

you need to add the #[no_mangle] attribute IIRC

well then my exported keyword is the same as extern "C" and no_mangle

13:21

ill change @ to $

Ryozuki

@Learath2 any known algorithm?

You could draw inspiration from rust or c++

13:24

You want the return type, the function name (including module maybe), and argument types in order

oh @ actually is valid

for whatever reason, the gdb demangler can segfault

13:26

and it's kinda known that it's bad code

In the stdcall and fastcall mangling schemes, the function is encoded as _name@X and @name@X respectively, where X is the number of bytes, in decimal, of the argument(s) in the parameter list (including those passed in registers, for fastcall). In the case of cdecl, the function name is merely prefixed by an underscore.

heinrich5991

for whatever reason, the gdb demangler can segfault

Huh

All mangled symbols begin with _Z (note that an identifier beginning with an underscore followed by a capital letter is a reserved identifier in C, so conflict with user identifiers is avoided

13:27

TIL

island is also a reserved name

All mangled symbols begin with _Z (note that an identifier beginning with an underscore followed by a capital letter is a reserved identifier in C, so conflict with user identifiers is avoided); for nested names (including both namespaces and classes), this is followed by N, then a series of <length, id> pairs (the length being the length of the next identifier), and finally E. For example, wikipedia::article::format becomes: _ZN9wikipedia7article6formatE

heinrich5991

island is also a reserved name

xd

13:28

the full text

13:28

ok shouldnt be hard to copy this mangling

i keep setting aside implementing strings and arrays with other stuff

13:54

i need to focus

13:54

justatest

Ryozuki

i keep setting aside implementing strings and arrays with other stuff

Are you doing utf8 strings?

Learath2

Are you doing utf8 strings?

justatest

13:58

for now ill do just "strings"

13:58

and not zero terminated

13:58

but ill probs do utf8

13:58

actually i wanna do CStr and String

13:58

like rust

13:59

i think ill end up doing rust2

13:59

justatest

13:59

i have 0 originality

pepeW

thats why i ask for ideas

13:59

but nobody here has

Rust is just perfect as it is

xD

13:59

indeed

Do unsafe by default rust

justatest

13:59

zig comptime is interesting

14:00

i just need to think a feature that makes edlang different

14:00

but for now im happy to have a working language

Ryozuki

i just need to think a feature that makes edlang different

fearful concurrency

i wanna know if there are cutting edge papers

14:01

on borrow checkers

Cutting edge borrow checkers are probably so much graph theory that your brain will melt out of your ears

rust with good async

garbage collected rust

Learath2

garbage collected rust

hey thats going too far

Mark and Sweep gc isn't too hard to implement

we dont do that here

I'm about to learn some go btw

nice

14:02

u only need 1 day

14:02

literally

14:02

its the only pro of the language XD

i hate how it handles errors

Why? Is it similar to some language I already know?

makes it all so ugly

14:03

@Learath2 google made it for interns

14:03

to learn how to code

14:03

and have people working fast in something

Learath2

Why? Is it similar to some language I already know?

hmm u return a tuple with the error and the result

14:03

if the error is not null u use the result

14:04

u have to check the error on all ur function invocations

14:04

an entire if block

14:04

every time

14:04

there is no sugar

14:04

u can use the result without checking the error

14:04

no safety

14:04

also it has nulls

14:04

so its already a bad lang

14:04

justatest

14:05

https://go.dev/blog/error-handling-and-go

Error handling and Go - The Go Programming Language

An introduction to Go errors.

Rust honestly has it mostly right, I see no point to inventing new languages anymore. If I were making my own flavor of rust, I'd have it unsafe by default, find some way to allow self referential structs and maybe try to improve on the async

f, err := Sqrt(-1)
if err != nil {
    fmt.Println(err)
}

14:05

im sry but i find it rly ugly

14:05

u have to do this

14:06

func Sqrt(f float64) (float64, error) {
    if f < 0 {
        return 0, errors.New("math: square root of negative number")
    }
    // implementation
}

Ryozuki

f, err := Sqrt(-1)
if err != nil {
    fmt.Println(err)
}

Hm, it's just lacking some sugar, isn't this the same approach as rust Results?

having no algebraic data types is a no go too

14:06

u have to use a default value

14:06

when returning a result

14:06

bad bad bad

Ryozuki

u have to use a default value

I guess it doesn't guard as well as Results against you accidentally forgetting to check that there were no errors

Learath2

Rust honestly has it mostly right, I see no point to inventing new languages anymore. If I were making my own flavor of rust, I'd have it unsafe by default, find some way to allow self referential structs and maybe try to improve on the async

the yoke crate is a nice way of doing self-referential structs, even though it has weird terminology

@Ryozuki do fractional number representations

https://ed-lang.org/

The edlang programming language

Experimental statically-typed compiled programming language made with LLVM and Rust

14:21

new design

14:21

actually i just changed the classless css framework i use

louis

@Ryozuki do fractional number representations

xd

louis

@Ryozuki do fractional number representations

you mean as fractions?

yes @heinrich5991

infinite precision arithmetic

btw why is no language doing that?

https://code.pyret.org/editor

14:26

a = 3/5
b = a * 14/2990
print(b)

Teero

btw why is no language doing that?

it seems not particularly useful as a standard representation to me because sooner or later you're going to pass the value to a non-polynomial function

14:29

and then your result can't be precise anymore

store the expression as a tree then evaluate it when it needs to be

my android calculator has arbitrary precision (edited)

14:33

I like that

14:34

python has fractions btw: https://docs.python.org/3/library/fractions.html

fractions — Rational numbers

Source code: Lib/fractions.py The fractions module provides support for rational number arithmetic. A Fraction instance can be constructed from a pair of integers, from another rational number, or ...

Teero

btw why is no language doing that?

Fractional representations of numbers are less performant to work with, at the end of the day when you do need the actual final value of your operation it'll incur an expensive DIV to get it and multiplying fractions together is two MULs instead of one, adding fractions with different bases is variable cost, incurs atleast 2 muls and an add

1

oh ok

why is this not added in new update? https://github.com/ddnet/ddnet/pull/7521

Add /mode allowing teams to behave like team 0 by dobrykafe · Pul...

Maybe more of a concept? Not sure if this is too hacky and can't guarantee this doesn't break things. Adds /mode that can enable teams to behave like team 0 with the added benefit of being ...

16:18

am I right that for 2 month a name can't be chosen and it is the only problem?

pisisik

why is this not added in new update? https://github.com/ddnet/ddnet/pull/7521

Why would you add something like this to the game? I don't really get whats the use case for this

Stepfunn

Why would you add something like this to the game? I don't really get whats the use case for this

Some players enjoy doing t0 speedruns. That would be the new kind of private servers.

Stepfunn

Why would you add something like this to the game? I don't really get whats the use case for this

as written there by Robyt3 "We should better support t0 runs without having to manually set a password on servers." but I can also add that it might be helpful if you still want that t0 gameplay but not with everyone so you can go in team with this t0 mode and kick annoying people from it without it being funvote

Stepfunn

Why would you add something like this to the game? I don't really get whats the use case for this

people getting mad about t0 runs cuz they're not allowed to ban everyone who joins the server

16:28

and no one will set a password for them since t0 runs are unsupported

i mean sometimes admins do and it has happened a few times, but its a hassle for both the runners and the admins (edited)

bencie

i mean sometimes admins do and it has happened a few times, but its a hassle for both the runners and the admins (edited)

yes, and it looks like the only thing that left is to decide on a name and to add it in current update cause I guess there is still time

ah alright that makes a lot of sense now

brownbear

1

[ddnet/ddnet] Pull request opened: #8098 Replace IGraphics::FreePNG with CImageInfo::Free

The engine graphics do not need to be involved with freeing image data. The Free function now also ensures that all other member variables are cleared when freeing the image data.

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 integration test
[ ] Considered possible null pointers and out of ...

[ddnet/ddnet] Issue closed: #8094 UI walkthrough for new players

@murpi skindb bot is down

21:07

do you know why it's down so often?

watfk this button means

Image attachment

cyberFighter

watfk this button means

reloading

its reloading way too long then

i cant play guys (edited)

cyberFighter

its reloading way too long then

is it just the button not changing?

cyberFighter

its reloading way too long then

you cant reach the master server im guessing

1

masters messed up? (edited)

[ddnet/ddnet] Issue closed: #7895 Changelog for mod-makers to keep up with the high development pace

Ewan

masters messed up? (edited)

works for me. what do you see?

2024-03-13 16:26:36 E http: https://master2.ddnet.org/ddnet/15/servers.json failed. libcurl error (28): Failed to connect to master2.ddnet.org port 443 after 10000 ms: Timeout was reached
2024-03-13 16:26:36 E http: https://master3.ddnet.org/ddnet/15/servers.json failed. libcurl error (6): Could not resolve host: master3.ddnet.org
2024-03-13 16:26:36 E http: https://master1.ddnet.org/ddnet/15/servers.json failed. libcurl error (22): The requested URL returned error: 525
2024-03-13 16:26:37 E http: https://master4.ddnet.org/ddnet/15/servers.json failed. libcurl error (22): The requested URL returned error: 404

can you visit https://master1.ddnet.org/ddnet/15/servers.json in the browser?

Sometimes master1 works after refreshing for me

Also saw this 525 once (and it gave this link): https://developers.cloudflare.com/support/troubleshooting/cloudflare-errors/troubleshooting-cloudflare-5xx-errors/#error-525-ssl-handshake-failed

Ewan

2024-03-13 16:26:36 E http: https://master2.ddnet.org/ddnet/15/servers.json failed. libcurl error (28): Failed to connect to master2.ddnet.org port 443 after 10000 ms: Timeout was reached
2024-03-13 16:26:36 E http: https://master3.ddnet.org/ddnet/15/servers.json failed. libcurl error (6): Could not resolve host: master3.ddnet.org
2024-03-13 16:26:36 E http: https://master1.ddnet.org/ddnet/15/servers.json failed. libcurl error (22): The requested URL returned error: 525
2024-03-13 16:26:37 E http: https://master4.ddnet.org/ddnet/15/servers.json failed. libcurl error (22): The requested URL returned error: 404

does it work again for you?

btw https://ddnet.org now also intermittently returns 525 errors. curl -I https://ddnet.org

yea, but that isn't as vital

heinrich5991

does it work again for you?

Yes

22:45

ah yes it did die

Exported 191 message(s)