DDraceNetwork - developer

ws-client1

<ChillerDragon> sv_gametype ctf

Thaaanks, and are you able to see Discord reacts on WS?

@Skeith hey can you ban SP Someone from DDNet USA vanilla server again pls?

03:06

Or any mod

Pathos

@Skeith hey can you ban SP Someone from DDNet USA vanilla server again pls?

could you #✉-create-a-ticket and provide a reason please? :)

Dealt with already

1

meloƞ

could you #✉-create-a-ticket and provide a reason please? :)

I was told to message here, I also couldn't find that channel so easily

My game is running out of VRAM whenever I'm trying to add skins, anything I can do to add all of them without crashing my game?

04:24

I am literally adding the same skins that I had before I re-installed my game.

ryuma

My game is running out of VRAM whenever I'm trying to add skins, anything I can do to add all of them without crashing my game?

don't add all of them justatest

, the game has no lazy loading, means it loads all skins at once. Maybe try to reduce the quality of some?

meloƞ

don't add all of them justatest

, the game has no lazy loading, means it loads all skins at once. Maybe try to reduce the quality of some?

Yeah, I figured out to just add a few at a time.

04:45

Kinda weird tho, back when I first added all sorts of skins I was able to just drop them in to the folder and thats it.

ryuma

Kinda weird tho, back when I first added all sorts of skins I was able to just drop them in to the folder and thats it.

its not only skins. Gameskins, assets, emoticons, etc all count into vram usage (edited)

meloƞ

its not only skins. Gameskins, assets, emoticons, etc all count into vram usage (edited)

Yeah I know, but I only have 1 of each for gameskins, assets, emoticons. Its basically "just" 1500 skins.

(edited)

04:52

DDNet skin database + ahl full skin folder

ryuma

DDNet skin database + ahl full skin folder

Why not use a skinDB url ? scrumplex.net/skins Or ddstats.tw/skins

meloƞ

Why not use a skinDB url ? scrumplex.net/skins Or ddstats.tw/skins

What is the difference tho?

05:37

I've got 1449 skins rn, my game takes 10 seconds to load already and I can't even go in the assets tab anymore

ryuma

What is the difference tho?

You see the skin when someone is using it, and you can use the skin by writing it in the skin name You can use skin favourites to always load them

05:38

But that will prevent issues with having to many skins

scrumplex.net/skins doesn't work

05:40

I basically just want the majority of skins that people on KoG use which is basically ddnet database + ahl custom skins

ryuma

scrumplex.net/skins doesn't work

ddstats.tw/skins has much more anyway iirc

everything was working before idk why its not working now

05:44

so much trouble for some skins

ryuma

everything was working before idk why its not working now

One skin to much ig ^^

ryuma

so much trouble for some skins

..some??

05:45

1500skins

I literally got the same two folders that I had before and now my game is crashing all of a sudden

i had ~2000 and then i made a pull request to load them faster and it did not get merged

Rename the folder to skins.bak Launch game, drag skins over that you want Try again

so i just put my skins on an http server

05:45

and use the ingame mechanism

05:46

that's all u can really do (edited)

i believe you

1

05:52

I guess I gotta filter some of the skins out

ryuma

Click to see attachment 🖼️

You lied

05:53

I see custom entities and gameskins

05:53

05:53

Also I blame t client

Yeah I added some of them back after I said that to check wether it causes problems or not (edited)

meloƞ

Also I blame t client

I tried with latest DDNet already - same problem. (edited)

?? it is not a tclient issue

1

Ya, the devs added the ability to load skins from a site, rather than requiring them to always be within a folder because of this reason. There's no fix for it

05:55

Just don't have this many skins in your folder

60mb how much is that in vram :0

morning ocaml devs and others

06:18

doing some codewars to learn it

06:18

let explode s = String.to_seq s |> List.of_seq

let rec reverse_chars s =
  match s with [] -> "" | h :: t -> reverse_chars t ^ Char.escaped h

let reverse_string (s : string) : string = explode s |> reverse_chars

06:18

06:21

let reverse_string2 (s : string) : string =
    String.to_seq s |> List.of_seq |> List.rev |> List.to_seq |> String.of_seq

06:21

this is a funny one

06:22

just gotta remember that i guess ocaml calls iterators sequences

MilkeeyCat

Does anyone have ideas for enums syntax which doesn't require pattern matching? feelsbadman

struct Bar {}
struct Baz {}

enum Foo = Bar | Baz;

let foo: Foo = Bar {};

switch foo {
    case Bar bar:
        // bar contains struct of type Bar
    case Baz baz:
        // baz contains struct of type Baz
}

Here's one of possible syntaxes

jxsl13

60mb how much is that in vram :0

Depends souly only resolution not disk size

1

ryuma

My game is running out of VRAM whenever I'm trying to add skins, anything I can do to add all of them without crashing my game?

Don't use resolutions above 512x256. Then you can add around 2000* skins with like 3gb vram (edited)

07:46

But I bet you have skins with 4k resolution. On your full HD monitor

Jupstar ✪

Depends souly only resolution not disk size

@Souly moment (It's solely btw) greenthing

hi Souly! (edited)

1

Now do some math and you notice that 4k skins are for 32k monitors

Skeith

@Souly moment (It's solely btw) greenthing

hi Souly! (edited)

I was wondering why autocorrect couldn't find it xddd

07:47

I blame too much discord

1

Jupstar ✪

Don't use resolutions above 512x256. Then you can add around 2000* skins with like 3gb vram (edited)

Alright, thanks. Will check through them later.

Jupstar ✪

Depends souly only resolution not disk size

Jupstar ✪

Depends souly only resolution not disk size

hi

2

pattern matching is the future of languages (edited)

11:17

https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2024/p3477r0.html

11:21

P3477R0 There are exactly 8 bits in a byte

Heh, finally someone put it in writing

Jupstar ✪

@Ryozuki https://blog.rust-lang.org/2024/10/17/Rust-1.82.0.html update

woo! float operations in const functions

Jupstar ✪

@Ryozuki https://blog.rust-lang.org/2024/10/17/Rust-1.82.0.html update

ye best

Patiga

woo! float operations in const functions

That sounds difficult to get right. Does the compiler have a table of how floats behave on every target or did they settle on a standard common behaviour?

https://blog.rust-lang.org/2024/10/17/Rust-1.82.0.html#floating-point-nan-semantics-and-const

Announcing Rust 1.82.0 | Rust Blog

Empowering everyone to build reliable and efficient software.

With this release, Rust standardizes on a set of rules for how NaN values behave. This set of rules is not fully deterministic, which means that the result of operations like (0.0 / 0.0).is_sign_positive() can differ depending on the hardware architecture, optimization levels, and the surrounding code. Code that aims to be fully portable should avoid using to_bits and should use f.signum() == 1.0 instead of f.is_sign_positive(). However, the rules are carefully chosen to still allow advanced data representation techniques such as NaN boxing to be implemented in Rust code. For more details on what the exact rules are, check out our documentation.

I didn't go through it all, but my understanding is that float operations are very standardized, except for NaN behavior

With the semantics for NaN values settled, this release also permits the use of floating-point operations in const fn. Due to the reasons described above, operations like (0.0 / 0.0).is_sign_positive() can produce a different result when executed at compile-time vs at run-time. This is not a bug, and code must not rely on a const fn always producing the exact same result.

12:25

important bits

and they decided on some rules regarding Nan stuff, which they use for const

the &raw mut syntax

12:26

u dont need to use addr of mut now

12:26

macro

Ryozuki

u dont need to use addr of mut now

Ah that is so nice. Macros were so ugly

Ryozuki

With the semantics for NaN values settled, this release also permits the use of floating-point operations in const fn. Due to the reasons described above, operations like (0.0 / 0.0).is_sign_positive() can produce a different result when executed at compile-time vs at run-time. This is not a bug, and code must not rely on a const fn always producing the exact same result.

you copied 2 out of the three paragraphs imo the link does a better job here ^^

hmm idk

12:29

its better to not click links

12:29

xd

12:29

assume the end user is lazy

12:29

and stupid

it can also amplify laziness, also in a #dev channel I think you can have higher expectations, especially as people are interested in the subject here

i guess

12:31

its more something i do cuz this way i know ppl actually see something

12:31

with a link u cant be sure they clicked and looked it

12:31

xd

yea, I kinda get that. I also know another person who does that, but I much prefer links: less spam, more complete (at least if it links to well-organized information) :)

they also added lot of new stable methods, some intrinsics

12:33

https://doc.rust-lang.org/nightly/std/boxed/struct.Box.html#method.new_uninit

Box in std::boxed - Rust

A pointer type that uniquely owns a heap allocation of type T.

12:33

not bad

Due to the reasons described above, operations like (0.0 / 0.0).is_sign_positive() can produce a different result when executed at compile-time vs at run-time. This is not a bug, and code must not rely on a const fn always producing the exact same result.

13:26

I think it might be better worded as "we don't consider this a bug"

13:26

because usually, one would expect the same behavior independent of whether the compiler chose to evaluate a function at compile time or not

I am not. I usually launch matrix to comfirm a reply is to a message of mine. On my mobile matrix client I only see „that“ it was a reply not to „what“. But reading all messages it’s usually not ambiguous.

Replying to @Pathos Thaaanks, and are you able to see Discord reacts on WS?

14:05

Not sure if my image got sent I got a 404 error back from the discord bot -.-

heinrich5991

Due to the reasons described above, operations like (0.0 / 0.0).is_sign_positive() can produce a different result when executed at compile-time vs at run-time. This is not a bug, and code must not rely on a const fn always producing the exact same result.

Eeeeeeeh, I guess they are kinda ignoring the hard stuff

15:01

I wonder why they don't have a table of all target behaviours so that this can be guaranteed

yea, it's a bit subpar IMO

I'm done with the GRE so this will probably be the last useless english word of the day, I'll do 3 for this special occasion:

hackneyed = lacking creativity
rapacious = very greedy
scintillating = impressively clever, amusing or witty

scintilla is (was?) a code view for windows

15:05

perhaps the name was derived from this word?

15:06

https://en.wikipedia.org/wiki/Scintilla_(software) — notepad++ uses it

Scintilla (software)

Scintilla is a free, open-source library that provides a text editing component function, with an emphasis on advanced features for source code editing.

15:06

a friend of mine used it for some code editor, too, IIRC

Apparently it's a word of it's own but both from the same latin root

15:08

It's actually in italian too, scintillare = to sparkle

im scintillating

same