00:01 < bridge> @Robyt3 has this been fixed? https://github.com/ddnet/ddnet/issues/4537 07:09 < bridge> "I'll help anyone interested on how to earn 100k in just 24hours from the crypto market. But you will have to pay me my commission! when you receive your profit! if interested send me a direct message https://t.me/victormarck by asking me HOW 11:45 < bridge> Was this an issue at all? I read this like 5 times a day xd. We should defs check out if the latest 2.x version works 11:51 < bridge> I remember this being an actual issue we verified with a profiler 11:55 < bridge> To me it doesn't even sound input related 11:55 < bridge> More like some weird vsync stuff again xd 11:56 < bridge> Anyway. Maybe windows fixed it. Or SDL in the meantime 11:56 < bridge> We should at least try it out 12:21 < bridge> @Jupeyy_Keks https://github.com/libsdl-org/SDL/pull/4683 doesn't apply anymore, do we still need it? 12:24 < bridge> Yep with new SDL we have to remove it 12:24 < bridge> ok 12:26 < bridge> We also should do a minor patch in our input handling. Somewhere in input.cpp i made some comments about it 12:40 < ChillerDragon> !help 12:40 < chillerbot> https://github.com/ChillerDragon/ddnet-bot-irc eth0=192.168.178.22 commands: !mods, !ping 12:41 < ChillerDragon> sorry too lazy to lookup nmap command (will find a fix for the future) 12:41 < bridge> !ping 12:41 < chillerbot> pong 13:27 < bridge> @Jupeyy_Keks apparently webgpu in chrome doesn't have support for linux yet 13:28 < bridge> really 13:28 < bridge> web browsers still do stuff like that? 13:32 < bridge> Sad. 13:32 < bridge> Sadly even for css stuff. Linux Firefox still hasn't backdrop filter default on. On windows since a few months it is iirc 13:33 < bridge> backdrop filter? 13:33 < bridge> https://caniuse.com/?search=webgpu says that it can be enabled in older chrome versions for linux, windows, macos 13:33 < bridge> so perhaps you can enable it on linux on the newest version as well? 13:34 < bridge> Css attribute 13:34 < bridge> Manually maybe 13:34 < bridge> ye, manually 13:35 < bridge> In Firefox Dev i couldn't run the wgpu thing. Worth a try in nightly chromium 13:35 < bridge> Or it's indeed a wgpu stable issue 13:36 < bridge> Maybe current gh repo works better ^^ 14:00 < bridge> @Jupeyy_Keks @heinrich5991 I somehow forgot to link it https://developer.mozilla.org/en-US/docs/Web/API/WebGPU_API#browser_compatibility 14:02 < bridge> Sad to know that MacOS will take another 100 years before implementing it 14:02 < bridge> Or more iOS 14:03 < bridge> Which doesn't allow a custom browser to run 14:03 < bridge> yea :/ 14:37 < bridge> We can only use SDL 2.0.18 or 2.0.20 for IME though, because later version introduce another bug 14:41 < bridge> Lmao 14:42 < bridge> The only thing that is more broken than ddnet editor is sdl 14:43 < bridge> eh, maybe ping me when that new issue is fixed? 😄 14:43 < bridge> Do they fix it before 3.0 release? 15:35 < bridge> https://en.wikipedia.org/wiki/Elias_gamma_coding 15:57 < ChillerDragon> ryo lets refactor tw pack int with elias coding just to break protocol for kicks 16:00 < bridge> that coding seems super inefficient 16:09 < bridge> its made to be space efficient 16:09 < bridge> for low values 16:10 < bridge> https://www.reddit.com/r/rust/comments/12nw1pc/announcing_bitcode_format_for_serde/ 16:10 < bridge> found it hereww 16:10 < bridge> xd 16:14 < bridge> I think I prefer something like twint, seems more space-efficient to me 16:14 < bridge> only the numbers 1-15 are more space efficient in elias than in twint, and everything else is much worse 16:15 < bridge> (do note that i never said i wanted this for tw xD i just found it interesting) 16:15 < bridge> I see 16:15 < bridge> it seems to be invented for clarity, not for practical use 16:15 < bridge> there is also the alternative of huffman 16:15 < bridge> that had some license limit or smth 16:15 < bridge> arithmetic coding? 16:15 < bridge> ye 16:15 < bridge> the patent expired by now I think 16:16 < bridge> nice 16:17 < bridge> zstd uses something in between arithmetic coding and huffman IIRC 19:07 < bridge> @Patiga not working either, either the twgpu implementation isn't correct for webgpu, or i am stupid (changed the ::GL to ::WebGPU), or its wgpu 19:18 < bridge> tips of the day: when you look for files deleted more than 3 months ago, you can still grep for string on /dev/sda 19:19 < bridge> and then? 19:19 < bridge> text files u mean? 19:21 < bridge> yeah, only works for text file 19:21 < bridge> and only if u didnt overwrite the bytes xD 19:21 < bridge> you grep with `grep -b` for byte offset and then extract the file you want with dd 😄 19:21 < bridge> but yeah 19:21 < bridge> still useful 19:28 < bridge> i don't get it, it's way less efficient? 19:28 < bridge> https://en.wikipedia.org/wiki/Universal_code_(data_compression) 20:41 < bridge> https://en.wikipedia.org/wiki/Neobuthus_factorio 20:41 < bridge> @Learath2 @heinrich5991 lol 20:41 < bridge> > The species was named after Factorio, a video game created by the son of one of the researchers who described the species.[1][2] 20:43 < bridge> the world is so small xd 21:04 < bridge> https://www.reddit.com/r/2westerneurope4u/comments/12pc3xd/france_and_belgium_always_the_special_children/ 21:05 < bridge> Do spanish and italian people really use qwerty keyboard? 21:05 < bridge> kataloneons like ryozuki do 21:07 < bridge> Uh, italian is qwerty with some extras 21:07 < bridge> kinda weird for latin language 21:07 < bridge> letter placement is probably suboptimal 21:11 < bridge> Qwerty is suboptimal even for english iirc. I think colemak is supposed to be the efficient one for english 21:33 < bridge> yeah 21:33 < bridge> qwerty with ñ and ç 21:33 < bridge> · 21:34 < bridge> l·l 21:34 < bridge> azerty are the weirdos 21:34 < bridge> im to used to qwerty 21:34 < bridge> cant switch 22:03 < bridge> have you used the compiler options that the wiki tells you to use? 22:03 < bridge> have u a webgpu version hosted? 22:04 < bridge> i mean u probs already tried, didnt ya 22:04 < bridge> no 22:04 < bridge> yes, only once and it didnt work 22:05 < bridge> otherwise maybe another version of wgpu might work, in particular there is a gecko branch which was supposed to work with the unstable firefox implementation (can be enabled as a unstable feature in firefox), didnt work the last time I tried 22:05 < bridge> did you do this? 22:06 < bridge> nah 22:06 < bridge> if it compiles it must work xdd 22:07 < bridge> I don't see hoe twgpu could have anything to do with it. wgpu might be not compatible currently somehow, maybe the js calls changed or smth 22:08 < bridge> hoe? 22:09 < bridge> darn phone keyboard :p 22:14 < bridge> did now, didnt work 22:17 < bridge> wgpu 0.15.1 sadly doesnt compile 22:18 < bridge> ah I think you can find recent (maybe 5 days ago) reddit post on game_dev talking about exactly that (I'm on phone and can't search). they said that wgpu 0.14.0 should work iirc. there was a bit of version breakage, but I think you can just go back a few commits for a working 0.14 22:18 < bridge> huh? you mean with the options? 22:19 < bridge> i just tried newest version in cargo.toml 22:19 < bridge> but probs the deps are out of sync then 22:20 < bridge> wdym? with the version going a few commits back should still be guaranteed to compile (at least for native) 22:21 < bridge> ah, the most recent version o.o 22:21 < bridge> is it a compile error in the crate thiserror talkong abt smth unstable? 22:21 < bridge> or some version mismatch? 22:22 < bridge> ``` 22:22 < bridge> error[E0412]: cannot find type `GpuUncapturedErrorEvent` in crate `web_sys` 22:22 < bridge> --> /home/jupeyy/.cargo/registry/src/github.com-1ecc6299db9ec823/wgpu-0.13.1/src/backend/web.rs:1758:62 22:22 < bridge> | 22:22 < bridge> 1758 | let f = Closure::wrap(Box::new(move |event: web_sys::GpuUncapturedErrorEvent| { 22:22 < bridge> | ^^^^^^^^^^^^^^^^^^^^^^^ not found in `web_sys` 22:22 < bridge> ``` 22:22 < bridge> I am not sure tbh 22:22 < bridge> actually i thought they can have their own web_sys deps 22:22 < bridge> and it should™️ just compile 22:22 < bridge> ah lol it also uses 0.13.1 now 22:22 < bridge> i c 22:23 < bridge> i have to change in all cargo.toml 22:23 < bridge> since they all have wgpu dep 22:23 < bridge> yea I updated to 0.14 and the right after 0.15 came out, so I went there 22:24 < bridge> --> twgpu/src/map/envelopes.rs:128:46 22:24 < bridge> | 22:24 < bridge> 128 | let texture = device.create_texture(&TextureDescriptor { 22:24 < bridge> | ^^^^^^^^^^^^^^^^^ missing `view_formats` 22:24 < bridge> 22:24 < bridge> 22:24 < bridge> they still added some stuff, but not too much 22:24 < bridge> hopüefully 22:24 < bridge> I really dislike assisting debuggin from mobile, with limited internet 22:25 < bridge> wait you do 0.13->0.14? I thought that switch should also be in the git history 22:26 < bridge> or what version did you just change to break it 22:28 < bridge> there should be a commit 0.13->0.14 and 0.14->0.15 right after 22:29 < bridge> from u? 22:29 < bridge> yes, for twgpu 22:29 < bridge> ah i am in matrix-optimization branch lel 22:29 < bridge> how did i end up there 22:29 < bridge> yea that one is long forgotten ^^ 22:30 < bridge> if those changes arent on the main branch, then they gotta be in `dev` 22:34 < bridge> sadly i still get no suitable adapter found 22:36 < bridge> also on 0.14.x? then unlucky. people on that reddit post had success iirc, so if you want to continue your tour that might be a suitable next stop 22:38 < bridge> mhh seems like neither master nor dev are up to date with wgpu 22:39 < bridge> neither is at 0.15.x? 22:39 < bridge> both on 0.13.1 22:39 < bridge> tee-rendering is 22:39 < bridge> ow, the go on the tee-rendering branch 22:39 < bridge> ye 22:40 < bridge> and go back to those version bum commits I'd suggest ^^ 22:40 < bridge> *bump 22:43 < bridge> Caused by: 22:43 < bridge> failed to read `/home/jupeyy/Dokumente/dev/libtw2/gamenet/ddnet/Cargo.toml` 22:43 < bridge> lol it reads from other dir rn 22:44 < bridge> yeye go back to those commits 22:44 < bridge> which commits? 22:44 < bridge> bump? 22:44 < bridge> bump wgpu to ... 22:47 < bridge> dokumente 22:49 < bridge> yeah bro i installed KDE in german 22:49 < bridge> bcs i am gansta 22:49 < bridge> af 22:50 < bridge> @Patiga how can i see which backend is used? or should i simply force WEBGPU again 22:51 < bridge> ah terhe is a version string in console 22:51 < bridge> still didnt work xD 22:52 < bridge> this lib is harder to use than c++ 23:03 < bridge> https://wgpu.rs/examples-gpu/?example=boids 23:03 < bridge> 23:03 < bridge> those examples also dont work for me 23:37 < bridge> i THINK i did the final update for #6437. Can anyone review to see if it's any good or not? 23:37 < bridge> https://github.com/ddnet/ddnet/pull/6437