00:00 < bridge> I think you should assign assa 00:01 < bridge> i thought u said u wanted to do it 00:02 < bridge> I was thinking about it back in the days when I had time 00:03 < bridge> Now I got wife and kids and am retired from Kirby game 00:04 < bridge> could readd to merge queue? 00:04 < bridge> was failed ci 00:04 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1404948610042298429/image.png?ex=689d0bdb&is=689bba5b&hm=b0c9a363d4031257b59445068a72af8bfcd29f51879a7ca2927f2112f3055fcd& 00:04 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1404948656322510979/image.png?ex=689d0be6&is=689bba66&hm=f765c379a64f6a96b3fd6b48985aac204e79318904cba2791c296e5fb0bc2d46& 00:05 < bridge> ? 00:05 < bridge> Stop faking konsti 00:05 < bridge> https://tenor.com/view/konsti-pink-rat-ddnet-gif-20877998 00:11 < bridge> potato? 00:12 < bridge> @kebscs: Team 0 can have practice? o0 00:12 < bridge> with sv_practice_by_default 00:13 < bridge> maybe after this pr and testing commands could be voted in t0 00:15 < bridge> anyone got any diea how to make a smart rescue algorithm 00:15 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1404951394859089960/image.png?ex=689d0e73&is=689bbcf3&hm=95bf7b6302daaa9b9f695e5cfaa770a0382d82ff35542b218388f449cea9d8f7& 00:15 < bridge> currently i have 8 floodfills around the player (they combine) then the one which is nearest and has line of sight is /tpxy'ed 00:15 < bridge> but in this scenario that just puts me 1 tile up 00:15 < bridge> Selfkill if freezetime souly 00:17 < bridge> No idea what you mean kebso 00:17 < bridge> it also gets the center of each floodfill so uhh 00:17 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1404951930232635432/image.png?ex=689d0ef2&is=689bbd72&hm=54955cc8c7a534df376b4f18fd56d0433e4b15fe11c071d9d51554581839ba36& 00:17 < bridge> @kollpotato: developa 🥔 00:18 < bridge> maybe i can just save positions which are safeish 00:18 < bridge> rip my overcomplex code 00:21 < bridge> done ✅ 00:22 < bridge> also done ✅ this one is a blocker for me, since I want to move and rewrite more stuff 00:24 < bridge> but dobby goes to sleep now, good night ^^ 00:28 < bridge> @essigautomat git works surprisingly well if you base follow up PRs on existing unmerged PRs. Just branch of there. They can even be prd at the same time and merged in weird orders and partially replace each other. 00:28 < bridge> But everything goes to shit if the base pr gets rejected or has to be rebased and force pushed 00:47 < bridge> stop pinging me 😭 00:47 < bridge> Who ghosted ping ? 00:49 < bridge> Me @badoumed not ghost tho .\_. 00:50 < bridge> chillerdragon: i played some chillerdragores today 00:50 < bridge> and perfect corner hook killed me 00:50 < bridge> i hope you are happy 00:50 < bridge> hrhrhr rip solly 00:50 < bridge> Ez troled 00:50 < bridge> Where is the ping I cannot see it 00:51 < bridge> Here 00:51 < bridge> Yea I just saw it 😄 00:51 < bridge> @soulyvevo: I actually mapped that to clip farm 00:52 < bridge> imagine an april fools map thats just some long tight gores full of those tiny corner gaps 00:52 < bridge> Sounds horrible 00:53 < bridge> :troll: 06:32 < bridge> usually the last grounded not freezed position is saved, but you can be smarter about that and consider time and some hookable positions or save multiple positions and delete the last on rescue 07:10 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405056009105313935/image.png?ex=689d6fe1&is=689c1e61&hm=6c0d12e97c5adb7f5d0f773c6754f1a1466412908ce8b56fcd837d652e75440c& 07:33 < bridge> no avx? 07:43 < bridge> yes but the base PR would be a bit unstable, since any knew change might force a rebase. 07:44 < bridge> gumo ^.^ 07:59 < bridge> this wouldn't save much on setup time, but I don't see much obvious potential in improving quad loading times. I can load billions of quads. It could however slightly improve the performance of rendering itself especially for ogl1 08:08 < bridge> beatiful this change also collides with 2 other PRs 🙁 09:12 < bridge> does the mapformat has a limit on the number of quads? At least twmap has one 🙈 09:30 < ws-client> **** i am running out of rcon rank levels sos. I currently miss 2 more auth levels. Any chance we could get more or even n more? Or custom access per account name? 09:30 < ws-client> **** maybe wrong place in this chat i should open an issue 09:44 < ws-client> **** @Assa i think rossbit mentioned something around 65k yesterday in this very chat 09:44 < ws-client> **** probably some kind of int max 09:57 < bridge> "And way better than teeworlds!" hits hard 10:14 < bridge> 2^64 auth ranks when? 10:47 < bridge> no it must be much higher, the mona lisa has more 10:47 < bridge> excluding the editor 10:48 < bridge> excluding the editor, because it crashes much much earlier than map rendering or the map format 10:48 < bridge> excluding the editor, because it crashes[1] much much earlier than map rendering or the map format 10:48 < bridge> 10:48 < bridge> [¹] gets stuck at 1 fps and is basically unusable 10:51 < bridge> envelopes, not quads 10:51 < bridge> envelopes, not quads (oops sry robyt didn't want to ping you) 10:52 < bridge> How can I make my own server in ddnet, like clb2s? 10:53 < bridge> $LAN Server 10:53 < bridge> This command was not found. 10:53 < bridge> 😠 10:53 < bridge> $LAN_Server 10:53 < bridge> This command was not found. 10:53 < bridge> https://wiki.ddnet.org/wiki/LAN_Server 11:53 < bridge> Skill issued hard 11:53 < bridge> $wiki lan 12:30 < bridge> $wiki hammerfly 12:41 < bridge> tested at offtopic 12:41 < bridge> $wiki LAN Server 12:51 < bridge> push #9793 12:52 < bridge> https://github.com/ddnet/ddnet/pull/9793 17:06 < bridge> can anyone help a bit? im stuck because i want to compile my rust visualizer of a c99 library to wasm, but my libary does use libc and libc isn't in wasm32-unknown-unknown. i can't use emscripten because i get weird internal errors from eframe/egui. have any of you done something like this before? 17:07 < bridge> llm says the only option i have is to rewrite the libc functions i use but that would be bloated asf for a small library and would generally be a bad idea because i assume the implementations of cosf, sinf, atan2, powf etc. are not short 17:07 < bridge> why u get errors from egui? 17:07 < bridge> egui is rust only 😮 17:10 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405206868401590312/message.txt?ex=689dfc60&is=689caae0&hm=311dd90b72695cfac5f19f9586d9441666a4f7555a129529b702f9ac6ee65ea8& 17:10 < bridge> `cargo build --target wasm32-unknown-emscripten` 17:14 < bridge> gl gamers 17:14 < bridge> melon was in vc and left?? 17:14 < bridge> noo 17:15 < bridge> i couldnt har his voice 17:21 < bridge> @blaiszephyr can u make another voice message 17:21 < bridge> i can see em using chatgpt 17:25 < bridge> no 17:25 < bridge> you'll hear my voice soon stop glazin 17:25 < bridge> there is nothing wrong with using ai for help as long as you understand what the ai is doing and double check 17:25 < bridge> yeah 17:26 < bridge> i use it to find some info about os dev 17:26 < bridge> cuz its hard to find stuff sometimes 17:26 < bridge> like my flags mistake above was found by AI 17:26 < bridge> stream soon??? 17:26 < bridge> because + has priority over bitwise & which seems wrong to me xD 17:26 < bridge> @blaiszephyr do a collab stream with freezy 17:27 < bridge> will you whisper some romantic things in my ear 17:33 < bridge> im not spending time debugging wasm + rust + c99 shit i just want to make it playable in the web so i can put it in the readme so ppl don't have to download and compile to see the demo 17:33 < bridge> i didnt say anything negative 17:33 < bridge> xd 17:33 < bridge> why jump the gun 17:34 < bridge> why are you so negative Ryo? Talking about using a gun 17:34 < bridge> yea damn 17:34 < bridge> why are you so negative Ryo? Talking about using a gun /s 17:35 < bridge> its actually a OG reference to teeworlds 17:35 < bridge> I actually know ^^ i think it was a slogan 17:35 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405213188513075241/Teeworlds_logo.png?ex=689e0243&is=689cb0c3&hm=38ee9bcccd30176b186ca1aa7e52dd3238386708af2fdab5fea58bd1e259a7ef& 17:35 < bridge> :brownbear: 17:36 < bridge> i use claude personally, it helped me debug some os stuff too 17:36 < bridge> altho it was far from having the insight to know the problem actually 17:36 < bridge> morning... or evening for some 17:36 < bridge> Oups 😬 17:36 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405213537751793694/IMG_20250813_173628.jpg?ex=689e0297&is=689cb117&hm=1052e2fdc4b77c0e54899212bf9e102788e6ec284ff05b417dfe788b5c920be0& 17:36 < bridge> yea i switch between them often too see which one gives better answers. gpt5 hallucinates a lot after only little conversation 17:37 < bridge> you often can ask the ai to give you sources and help you find the insights 17:37 < bridge> yeah 17:37 < bridge> i thought gpt5 was supposed to be good st not hallucinating lmao 17:37 < bridge> no, it just hallucinates stuff more professionally now 17:38 < bridge> often I use AI just for 1 prompt and don't actually have a conversation 17:38 < bridge> actually, predicting is the right word here, like a next-door foresight person 17:38 < bridge> that demands 20$+ / month 17:38 < bridge> i asked about tips to parse fat32 17:38 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405213945781817344/image.png?ex=689e02f8&is=689cb178&hm=05387a45005ef0830d7d54cef511bfd55fab739f06ccd297dbb800797d0d0ed3& 17:39 < bridge> maybe it was trained on the fat standard and actually "knows" 17:39 < bridge> ask for a documentary or source of this information, like where it got it from 17:39 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405214149918855260/image.png?ex=689e0328&is=689cb1a8&hm=29c5d62e4ea4d8945f409de1f87f6ee5f95abd96fe41c2adb007b1112b4fd309& 17:39 < bridge> nah i just verify it if needed 17:39 < bridge> but for things with standards that have lots of pdfs, like the fat32 format 17:39 < bridge> its pretty reliable 17:40 < bridge> true why waste time confirming that the source is real, when you can verify it urself 17:40 < bridge> https://tenor.com/view/ladno-gif-20837499 17:40 < bridge> ?? xd 17:40 < bridge> :gigachad: 17:40 < bridge> you can actually find it on wikipedia, just not so nice formatted for my taste 17:41 < bridge> yeah i asked for tables specifically 17:41 < bridge> cuz i love tables 17:41 < bridge> easy info 17:41 < bridge> Surprisingly AI is very good at procuding the shittiest results in the nicest, most professional way possible 17:42 < bridge> anyway i researched a bit up what ppl do with claude, apparently you can have multiple "agents" each defined with a different role and they discusss with each other before giving u an answer, and apparently this results in better results 17:42 < bridge> idk if gpt has this 17:43 < bridge> just treat ai as a powerful google and its ok 17:43 < bridge> this is one of my real sources 17:43 < bridge> https://elm-chan.org/docs/fat_e.html 17:43 < bridge> https://elm-chan.org/ 17:43 < bridge> its from a weeb railways japanese engineer 17:44 < bridge> > My hobby has been a present occupation as it is because I liked electronics relatings since I was little. Railroad, anime, comic and TV game as well. 17:44 < bridge> > As for programming in the hobby, because they are electronic handicraft relations chiefly, the DOS based applications take the most part. However, as for the work, it has been increasing to write Windows based applications. 17:44 < bridge> his english is broken 17:44 < bridge> ``` 17:44 < bridge> cargo:warning=In file included from ../build_libccar.c:2: 17:44 < bridge> cargo:warning=../libccar.h:45:10: fatal error: 'math.h' file not found 17:44 < bridge> cargo:warning= 45 | #include 17:44 < bridge> cargo:warning= | ^~~~~~~~ 17:44 < bridge> cargo:warning=1 error generated. 17:44 < bridge> 17:44 < bridge> --- stderr 17:44 < bridge> 17:44 < bridge> 17:44 < bridge> error occurred in cc-rs: command did not execute successfully (status code exit status: 1): LC_ALL="C" "clang" "-O0" "-ffunction-sections" "-fdata-sections" "-fno-exceptions" "-g" "-fno-omit-frame-pointer" "--target=wasm32-unknown-unknown" "-I" ".." "-I" "/usr/share/wasi-sysroot/include" "-Wall" "-Wextra" "--sysroot" "/usr/share/wasi-sysroot" "-o" "/home/teero/software/libccar/car_demo/target/wasm32-unknown-unknown/debug/build/car_demo-09df86eb4 17:44 < bridge> ``` 17:44 < bridge> compiler is now starting to contradict itself 17:44 < bridge> i love it 17:48 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405216343657021500/message.txt?ex=689e0533&is=689cb3b3&hm=bee0de8798b40823d3b96ecb4f1856f8c55811e031a2ec5ef21ef9e42468f033& 19:21 < bridge> chat gpt gg 19:21 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405239830605992018/Discord_PGb72ASaz4.png?ex=689e1b13&is=689cc993&hm=40d86f4cd48761faf9523426626be1e39549d74ec0335ea9e30636d8dc3e2694& 19:22 < bridge> grok, gpt, claude, gemini 19:22 < bridge> 19:22 < bridge> teero has them all 19:22 < bridge> how much does bro spend on tokens 19:23 < bridge> pog 19:23 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1405240357393797140/image.png?ex=689e1b91&is=689cca11&hm=ca6875d1b0f413178ff5de74d36ed793928936d7e3d551bbd52b07b5d5617fac& 20:04 < bridge> 0 man i hate wasm 20:05 < bridge> friendship ended with wasm before it started 20:36 < bridge> what does it have to do with wasm 20:38 < bridge> Thank you Robyt :deen_star: 21:36 < bridge> Apparently 32 GiB of RAM aren't enough anymore. Firefox takes like 13 GiB idle in the background, then Ubuntu VM takes a few GiB when used and randomly VS Code decides to hang during code completion and takes all remaining memory. This causes the Desktop Window Manager to run out of memory so it keeps crashing as well, because it can't show an error message about memory being low. :pepeW: 21:41 < bridge> how many tabs do you have open to use 13 gb xd 21:42 < bridge> running a vm also is expensive 21:43 < bridge> Idk, 32gb is enough for me 21:44 < bridge> You can use swap to prevent crashes 21:44 < bridge> Although i think it's better to just use less ram 21:49 < bridge> No, we need more RAM.. at least 3 TB 21:49 < bridge> Per PC 21:49 < bridge> then all windows ui could be written in react, yey 21:52 < bridge> Data throughput is still a limited resource even with unlimited RAM 21:52 < bridge> @ryozuki I'll rewrite an ir for my compiler for the 4th time, surely this time will be the last one :greenthing: 21:52 < bridge> It seems like CPUs already are like 90% cache lol 21:53 < bridge> replace all ram with l1 cache 21:53 < bridge> pure performance 21:56 < bridge> you'll need quite a lot of power and of cooling to do that 21:56 < bridge> something around like 20kw iirc 21:57 < bridge> But it would be worth it 21:57 < bridge> take it to space 21:57 < bridge> not really, access time would also increase dramatically 21:59 < bridge> I think we need 3D CPUs 21:59 < bridge> A flat surface is quite limiting 21:59 < bridge> 3D gives you infinitely more possibilities 22:02 < bridge> That's an interesting bug 22:07 < bridge> 3D chip is also surprisingly hard to make cuz current silicon technology was designed for 2D chip 22:10 < bridge> cube cpu 22:10 < bridge> It would be so cool 22:11 < bridge> it would be so hot, it's almost impossible to extract heat out of a cube 22:12 < bridge> That's why you'd have to have holes for water cooling 22:12 < bridge> Btw why do CPUs get so hot 22:12 < bridge> And why do they use so much power 22:14 < bridge> bcs transistors 22:14 < bridge> their heat output scales by square or smth 22:14 < bridge> increase a bit of voltage and you have a lot more heat 22:16 < bridge> P_dyn = C\*V\*F^2 22:16 < bridge> or CV^2F 22:16 < bridge> But there are also power efficient CPUs 22:16 < bridge> C is capacity of the chip (in Farad), V is power voltage, F is frequency of the chip 22:17 < bridge> Then they slow 22:19 < bridge> I am sure it makes sense somehow 22:19 < bridge> But it still feel weird how my PC required fans all the time 22:19 < bridge> But phones can just work 22:19 < bridge> It doesn't 22:19 < bridge> It will just be slow 22:19 < bridge> xD 22:20 < bridge> Pretty much all modern desktop cpus have rather good overheat protection 22:20 < bridge> well my ryzen 9950x3d burnt to death 22:20 < bridge> but i'd say it was not my fault 😉 22:20 < bridge> So how did it burn 22:21 < bridge> If there are protections 22:21 < bridge> it most likely is a bug in the commuication between motherboard and cpu 22:21 < bridge> u can read the asrock subreddit if you want to go that rabbit hole 22:21 < bridge> Wow that's actually interesting 22:22 < bridge> I have 0 idea how it works 22:22 < bridge> CPUs have so many pins 22:22 < bridge> I wonder how they're used 22:23 < bridge> i mean i dunno what you want to know. i also only have a slight overview 22:23 < bridge> but there are VRMs on the motherboard that need to very quickly change voltage depending on what the CPU needs etc. 22:23 < bridge> VRMs are what makes a good motherboard good vs bad 22:23 < bridge> VRMs are what makes a motherboard good vs bad 22:24 < bridge> You might hear of how nice the USB slots are etc. but VRMs are what e.g. overclockers care more about 22:26 < bridge> ```rust 22:26 < bridge> pub struct AsyncMutex { 22:26 < bridge> lock: AtomicBool, 22:26 < bridge> queue: Arc>, 22:26 < bridge> data: UnsafeCell, 22:26 < bridge> } 22:26 < bridge> 22:26 < bridge> impl AsyncMutex { 22:26 < bridge> pub fn new(value: T) -> Self { 22:26 < bridge> Self { 22:26 < bridge> queue: Arc::new(SegQueue::new()), 22:26 < bridge> lock: AtomicBool::new(false), 22:26 < bridge> data: UnsafeCell::new(value), 22:26 < bridge> } 22:26 < bridge> } 22:26 < bridge> 22:26 < bridge> #[inline(always)] 22:26 < bridge> pub fn lock<'s>(&'s self) -> AsyncMutexLockFuture<'s, T> { 22:26 < bridge> AsyncMutexLockFuture { 22:26 < bridge> lock: &self.lock, 22:26 < bridge> queue: self.queue.clone(), 22:26 < bridge> data: self.data.get(), 22:26 < bridge> } 22:26 < bridge> } 22:26 < bridge> } 22:26 < bridge> 22:26 < bridge> pub struct AsyncMutexLockFuture<'a, T: ?Sized> { 22:26 < bridge> lock: &'a AtomicBool, 22:26 < bridge> queue: Arc>, 22:26 < bridge> data: *mut T, 22:26 < bridge> } 22:28 < bridge> @pilonpl if you have a ryzen 5000+. 22:28 < bridge> 22:28 < bridge> The safest option to overclock is to undervolt. 22:28 < bridge> 22:28 < bridge> It won't go above the spec clock, but it will more easily hold the CPU clock high, bcs less voltage => less heat => easier for the CPUs to not throttle 22:28 < bridge> Undervolting also doesn't really dmg anything, but can lead to unstable operating 22:28 < bridge> Or just don't change anything 22:29 < bridge> And it's probably good enough 22:29 < bridge> It's more effective to buy a modern PC 😄 22:29 < bridge> But yeah, it's also fun to see how much more % you can squeeze out of your hardware 22:30 < bridge> With undervolting you easily get 5% without much trouble 22:30 < bridge> -20mv 22:30 < bridge> Nowadays i think even cheap CPUs are great? 22:31 < bridge> CPUs generally are rather good 22:31 < bridge> But GPUs differ extremly 22:31 < bridge> The iGPUs from intel just almost all suck 22:31 < bridge> GPUs are terrible 22:31 < bridge> But yeah even for CPU depends on task 22:31 < bridge> Way too expensive 22:32 < bridge> 16 cores is defs nice compared to 8 for developing 22:32 < bridge> For gaming it wont matter 22:38 < bridge> ok this was wrong, fixed some waker stuff 22:38 < bridge> 90% are for dram though or PCIe interface 22:39 < bridge> takes a lot to reach like 100GB/s bandwidths 22:39 < bridge> remaining ones are mostly power pins 22:39 < bridge> yeah something that big can't get all its power from one pin, having that much power electronics inside the chip would be foolish 22:40 < bridge> yeah something that big can't get all its power from two pins, having that much power electronics inside the chip would be foolish 22:40 < bridge> I know nothing about electricity 22:41 < bridge> I hope we get smth better than transistors soon. 22:41 < bridge> 22:41 < bridge> The performance is already declining per year 22:41 < bridge> I don't know much about this sort of electronics, but the idea is that it's really hard to keep your voltage stable when your device is switching that quickly 22:41 < bridge> I don't want to die before the next big thing 😬 22:42 < bridge> like, different instructions need different parts of the chip, and the power draw can fluctuate a lot even on like microsecond scale 22:43 < bridge> Okay i guess 22:43 < bridge> It's not obvious what is consuming power in a CPU 22:43 < bridge> Transistors surely 22:44 < bridge> But it feels like it should be basically free 22:44 < bridge> I think it's mostly just resistance of like *everything* 22:44 < bridge> there are like a bazillion miles of wire inside a CPU 22:44 < bridge> all of that has resistance 22:44 < bridge> Well that's unfortunate 22:45 < bridge> That's why superconductors are awesome 22:46 < bridge> igpus are great for transcoding video streams. 22:47 < bridge> people have theorized carbon chips that would have much lower resistance and better heat conductivity 22:47 < bridge> big problem is nobody has any idea how to make them lol 22:48 < bridge> silicon isn't exactly the greatest heat conductor 22:48 < bridge> Is it even possible to make something much better in terms of performance? 22:48 < bridge> in theory yeah much better 22:49 < bridge> current CPUs aren't even *that* limited by density, though scaling things smaller still helps 22:49 < bridge> power density is a really big issue too 22:50 < bridge> if you just made a chip as dense as you felt like, it'd heat up so quickly that it'd be impossible to keep cool 22:51 < bridge> Time complexity is what's annoying. You can have the fast CPU in the world and you could still choke it with a n^2 algorithm 22:51 < bridge> Why is reality so annoying? 22:51 < bridge> @totar im not the only one you see? 22:51 < bridge> and NP is probably not P so rip bozo 22:52 < bridge> ? 22:52 < bridge> i ranted a few days ago in vc didn't i? 22:52 < bridge> quantum computers can help... if it's even possible to make one that stays coherent 22:53 < bridge> how does an instruction set for a quantum processor looks like 22:53 < bridge> is there even one 22:53 < bridge> yes 22:53 < bridge> There are quantum gates 22:53 < bridge> uhhh 22:53 < bridge> a bunch of wacky quantum gates that do like fun complex number operations on your qubits 22:53 < bridge> watch the 3blue1brown video it was pretty interesting for a noob like me 22:54 < bridge> basically you replace all your bits with these funny complex number things (is it just a pair of complex numbers or was it more involved than that?) 22:55 < bridge> And then magically n^2 becomes n 22:55 < bridge> That's a really good magic trick ngl 22:56 < bridge> all your inputs and outputs are still normal classical bits, but the internal state of the computer gets really wacky 22:56 < bridge> random number generator xd 22:57 < bridge> yeah, some hardware RNGs (maybe even most?) use quantum mechanical effects to produce randomness 22:58 < bridge> I think lava lamps are the best way 22:58 < bridge> lava lamps are a really fun way 23:02 < bridge> Is there some good video about video codecs? 23:03 < bridge> I've seen some about images 23:03 < bridge> But video would be interesting 23:05 < bridge> yeah my Firefox has been sitting at like 12GB lately but I just have a shitload of tabs. ironically since I've had 64GB I've rarely used VMs whatsoever 23:06 < bridge> idk how zed compares to vscode when it comes to memory, but it's worth a try 23:07 < bridge> I haven't really missed any extensions so far minus weird specific things like a postgres db viewer 23:07 < bridge> Zed is meant to be native and fast and responsive so it's probably decent 23:11 < bridge> zed is nice if you are not on windows 23:12 < bridge> comparing to vscode it feels extremely fast 23:12 < bridge> uses ~3 less memory for me 23:12 < bridge> but it also depends on language servers 23:12 < bridge> rust analyzer can just eat 4 gb at least 23:18 < bridge> zed ran just fine on windows if you self compile 23:20 < bridge> my friend had lots of issues 23:20 < bridge> but it was a few months ago 23:21 < bridge> what kind of issues :o - i've been using zed on and off for a year now on both linux and windows now 23:21 < bridge> what kind of issues :o - i've been using zed on and off for a year now on both linux and windows 23:26 < bridge> @aoetw: nice find. Btw I edited your issue to include empty lines around the video links so they get a preview 23:27 < bridge> Didn't know how to do a preview, but now I do. Ty!! 23:29 < bridge> language server issues 23:29 < bridge> auto complete is not consistent and does not always show up 23:29 < bridge> Hmm, no I actually don't. How did you make the preview? 23:30 < bridge> @aoetw: links to images and videos get a preview if there is an empty line before and after the link 23:30 < bridge> which lang, rust? 23:30 < bridge> c++ can be annoying if you dont use (c)make to generate the paths for clangd 23:30 < bridge> Gotcha, thanks 🙂 23:31 < bridge> i suppose js/ts 23:31 < bridge> i dont really wanna bother him with these questions rn 23:31 < bridge> he uses vscode for now but still likes zed 23:32 < bridge> zed holds a special place in my heart 23:32 < bridge> first rust project i took a deeper look into 23:32 < bridge> i want to learn vim 23:32 < bridge> `h j k l z y p u b B a w W` 23:32 < bridge> especially vim motions 23:32 < bridge> gz, you learned vim 23:32 < bridge> yes 23:33 < bridge> i hate hjkl movement 23:33 < bridge> maybe my fingers are short 23:33 < bridge> but it just does not feel right 23:34 < bridge> xd try emacs! 23:35 < bridge> emacs is an operating system 23:35 < bridge> change my mind 23:35 < bridge> a friend of mine writes his emails in emacs 23:35 < bridge> but i also looked into helix 23:35 < bridge>