03:32 < ws-client> watefek who dis? https://github.com/Jipok 03:32 < ws-client> found him while browsing non tw github repos ?! 03:33 < ws-client> tw fame? 04:19 < ws-client> the gpl they are talking about here what is it? Is the the license? https://forums.developer.nvidia.com/t/linux-6-7-3-545-29-06-550-40-07-error-modpost-gpl-incompatible-module-nvidia-ko-uses-gpl-only-symbol-rcu-read-lock/280908/55 04:19 < ws-client> did a fucking license break the nvidia compile during my debian update!? 04:31 < ws-client> bruv kernel panic on boot. And debian is stable they said^^ 04:48 < ws-client> Yea so if there are any debian users here that have the nvidia package installed I recommend holding back updates for a while. 07:21 < bridge> As soon as it gets the extended sever info 07:21 < bridge> So in this case the master server script 07:22 < bridge> But maybe Heinrich didn't add this 07:49 < bridge> chillerdragon: had this at work yes 07:49 < bridge> With my RTX A500 08:24 < bridge> funny tho debian tells me it's a A1000 08:25 < bridge> funny tho debian tells me it's a A1000, maybe it is 08:27 < bridge> well what do you expect it to do 08:27 < bridge> what encoder & settings are you using 08:27 < bridge> can't really blame obs 08:28 < bridge> libavcodec does all the heavy lifting as far as video encoding is concerned 08:43 < bridge> I expected storage not to stall the entire process when you're only running on memory 08:43 < bridge> Yes myself was the issue 09:11 < bridge> Maybe this looks bad, but it looks fine to me. My first hud in lua xdd 09:11 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207600024582623263/image.png?ex=65e03c29&is=65cdc729&hm=e23a34a1d56b5fe0790d2f608ad3ab92a2cd745fb1391352e86387909f1407e9& 09:12 < bridge> Hot 09:17 < bridge> You need to recv the client extend serverinfo msg. 09:19 < bridge> The client send a packet with NET_HEADER_EXTENDED 09:20 < bridge> The packet like this mode 09:20 < bridge> [NET_HEADER_EXTENDED] [Extra Data] 09:21 < bridge> You need to get the extra data, and 09:22 < bridge> analyze it 09:23 < bridge> broke ass 09:23 < bridge> MRPG0.6 ? 09:23 < bridge> i don't speak broke 09:24 < bridge> Nope 09:25 < bridge> Oh 09:29 < bridge> Are you planning to hard fork like ninslash or nodes? 09:36 < bridge> is this client <-> server lua communication or is it just clientside lua binding baloney 09:41 < bridge> The server sends the lua code its full on remote code execution axaxaxa 09:42 < bridge> At least that’s what I understood from his previous message 09:45 < bridge> Lmao 09:46 < bridge> No, I think 09:47 < bridge> Server is avaible for standard ddnet client 09:47 < bridge> But custom client can receive lua code from server and execute 09:48 < bridge> axaxaxa 09:49 < bridge> How are you going to advertise this rat? @kekomonter 09:50 < bridge> :justatest: 09:50 < bridge> Jokes on me by the way 09:50 < bridge> I gave you the same access on my vps lmao 09:50 < bridge> Yeah be a bit careful xd 09:50 < bridge> Yikes 09:50 < bridge> xDDD 09:50 < bridge> xd 09:50 < bridge> The Russian hackers 09:50 < bridge> This opens a huge door for hackers 09:51 < bridge> Lua itself also has quite some cve apparently 09:51 < bridge> Use wasm with rust 09:51 < bridge> 😏 09:52 < bridge> I think it was lerato who once said „client side rce is a feature so servers can send better prediction code to reduce lags“ 09:52 < bridge> Rust? 09:52 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207610367241363546/ryo.png?ex=65e045cb&is=65cdd0cb&hm=dceb898a09d28a3a3b6ec72f823819618530fd2fbe2ebe6197104edfc187dd1d& 09:54 < bridge> Fake rust doesn't require parenthesis after a if or for token 09:55 < bridge> Else indeed looks like rust 09:56 < bridge> Are there any more PVE mod like Zod, Monster? 09:57 < bridge> theres a syntax error on line 12 09:58 < bridge> https://cdn.discordapp.com/emojis/1170067019446632548.webp?size=48&quality=lossless 10:03 < bridge> Huh, why? 10:03 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207613132239802428/image.png?ex=65e0485f&is=65cdd35f&hm=eb204f554187cbac18f5901d3d0ba847f78e61a6ab9d0ffa1fd1c53b7406243e& 10:04 < bridge> What is your question 10:04 < bridge> thursday vulkan? 10:04 < bridge> Different styles xd 10:04 < bridge> Oh yeah 10:05 < bridge> Ah yeah xd 10:05 < bridge> Who cares 10:05 < bridge> https://tenor.com/view/programming-rust-c-lang-cpp-cplusplus-gif-26743262 10:05 < bridge> @learath2 how has your vk experience been so far? 10:05 < bridge> @tsfreddie 10:06 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207613770608807946/how-do-you-bath-v0-3b2t7eqpfrgc1.png?ex=65e048f7&is=65cdd3f7&hm=bd606acf38fc01fd26d18cf7230707b36c52ce3ce36a37ef9ea9311227867a1f& 10:06 < bridge> looks like js 10:06 < bridge> I couldn't work on it for a couple days. I was in germany 10:06 < bridge> no semicolons, recurring 5-character red identifier used as the only one makes me think `const` 10:06 < bridge> It looks like it wants to be written RTL instead of LTR 10:08 < bridge> No way 10:08 < bridge> Would be very funny 10:09 < bridge> Opening parentheses the other way around 10:09 < bridge> wtf it's Arabic programming language ? 😄 10:10 < bridge> wtf is it Arabic programming language ? 😄 10:10 < bridge> Maybe xd 10:11 < bridge> RTL language on LTR idea noice 😄 10:12 < bridge> @hussainx3 can u read left to right just as good as right to left? 10:12 < bridge> 10:12 < bridge> Is one easier/harder? 10:31 < bridge> https://edgarluque.com/ 10:31 < bridge> updated the look a bit 10:31 < bridge> control-f5 to remove cache 10:31 < bridge> :owo: 10:34 < bridge> how can I get my bot's ID in the twnet_parser library 10:34 < bridge> pyson andy spotted 10:34 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207621058576777236/image.png?ex=65e04fc0&is=65cddac0&hm=2cf6a3e45067cb03e973808e1f59e0d26c2cdc65ae8b6c1533e2f5be30e4c945& 10:37 < bridge> remote bot clients :thonk: 10:40 < bridge> haha i did double precision and i can literally see a bug xDDDD 10:40 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207622360614703205/image.png?ex=65e050f7&is=65cddbf7&hm=25251b20f7f4f396528979a667e2ed287e47bd622c03471da03b5ed19cf4da07& 10:40 < bridge> i was wondering what those shapes where 10:42 < bridge> and it takes so long to compute that i timeout ...... 10:42 < bridge> 🤔 10:42 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207622887670943796/image.png?ex=65e05174&is=65cddc74&hm=32a2ab75756bab765735e67af80f91406781ec41b6cdedac55423c729cb9eed1& 10:43 < bridge> I don't think walking through walls has been added yet 10:43 < bridge> i dont think my screen res is high enugh on this laptop 10:44 < bridge> i don't think my screen res is high enough on this laptop 10:45 < bridge> ehhh ooga booga language please 10:49 < bridge> Are we openly sharing client bots in developer now? 10:50 < bridge> yes visuals very bot! 10:51 < bridge> only pathfinding no bot 10:51 < bridge> thanks for the heads up :p 10:57 < ws-client> @va_ll you are using the flood.py as base right? Then this should work https://paste.zillyhuhn.com/Fz 10:57 < ws-client> https://zillyhuhn.com/cs/.1707991049.png 10:57 < ws-client> irc code snippets be like 11:01 < ws-client> axaxaxax i love how i converted two client skids to 0.7 users 11:01 < bridge> Thank you 11:01 < ws-client> :3 11:06 < bridge> <4ipset> what is "ws-client" 11:16 < bridge> its already there like i can see max and current players on browser but doesnt go more than 16 11:25 < ws-client> @4ipset https://chat.zillyhuhn.com/ 11:31 < ws-client> huhuhuh 11:32 < bridge> ChillerDragon: i can't register pls help 11:32 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207635425846366248/image.png?ex=65e05d22&is=65cde822&hm=87e678cbe46e15a0bbc7ebfd081a72c053ff3c05c43464a085dafe26ca0ebb10& 11:33 < bridge> <4ipset> i too cant 11:34 < bridge> token is nimrocks 11:34 < ws-client> <4ipset> hi guys 11:35 < bridge> ah lol i thouht it was a password confirm box xddd 11:36 < ws-client> Hello World! 11:36 < ws-client> <4ipset> hi 11:39 < ws-client> hi\ 11:39 < ws-client> <4ipset> yo 11:39 < ws-client> <:brownbear:346683497701834762> 11:39 < ws-client> <:hisnail:768893210726367232> 11:39 < ws-client> <4ipset> <:tee_thinking:478629518358085653> 11:40 < ws-client> oh 11:52 < ws-client> hey 11:52 < ws-client> hi 11:54 < bridge> my bot only writes the first message it sees 11:54 < bridge> ```python 11:54 < bridge> if (time.time() - self.last_send_time) > 0.1: 11:54 < bridge> self.send_msg(CtrlKeepAlive()) 11:54 < bridge> for msg in packet.messages: 11:54 < bridge> if msg.message_type == "game" and msg.message_name == "sv_chat" and msg.client_id != -1 and msg.message.find('!test') != -1 and msg.client_id != bot_info["id"] and msg.message != "Do you know someone who uses a bot? Please report them to the moderators.": 11:54 < bridge> print(1234) 11:54 < bridge> msg = MsgClSay(message = msg.message) 11:54 < bridge> msg.header.flags.vital = True 11:54 < bridge> self.send_msg(msg) 11:54 < bridge> ``` 11:55 < bridge> but he writes message 1234 to the console 11:55 < bridge> a friend just walked by me and saw me playin tw and he compared it to coolmathgames...... 12:02 < bridge> now he only writes if there is 1 letter 12:02 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207643194817519696/image.png?ex=65e0645e&is=65cdef5e&hm=c776ce0863493740d17feace1a48af1cd3a8b86f9865847355bf73d142c6c244& 12:03 < bridge> and the Russian doesn’t see 12:19 < bridge> @va_ll: is your code public somewhere? 12:20 < bridge> A person called vali using flood.py to connect bots to tw servers. Anyone else getting nostalgia from this? C: 12:22 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207648113968283708/message.txt?ex=65e068f3&is=65cdf3f3&hm=2119078631d6cbce45381d61b4ab9a19b01848764b1802c773dd327e6daf164f& 12:23 < bridge> *vall 12:23 < bridge> xd 12:23 < bridge> Oh 12:23 < bridge> nostalgia hitting hard 12:23 < bridge> xd 12:24 < bridge> <4ipset> why does everyone call Vall by the name Vali? xd 12:24 < bridge> Looked like an I 12:25 < bridge> I just don’t expect the name vall 12:26 < bridge> Differentiating between I and l is mostly context for me 12:26 < bridge> https://pastebin.com/nZmhy3uM 12:26 < bridge> <4ipset> a lot of people called Vall named Vali ingame 12:26 < bridge> I was talking about a very limited language though. Something perhaps intentionally not turing complete even 12:26 < bridge> Sure sure 12:27 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207649306564239440/image.png?ex=65e06a0f&is=65cdf50f&hm=df7ca3b8442a7278d5dc6494f54947cbb5b5fb49e0ba51b39188009a0e9951eb& 12:30 < bridge> he also sometimes writes messages to the logs 2 times 12:30 < bridge> but it's probably my ping 12:30 < bridge> :wtf: 12:32 < bridge> <4ipset> :what: 12:34 < ws-client> @va_ll looks like chance to me 12:35 < ws-client> ah wait no hmm 12:35 < ws-client> ah yes it is chance 12:36 < ws-client> ``if (time.time() - self.last_send_time) > 0.1:`` this seems to be the problem i assume 12:36 < ws-client> last_send_time gets set on every packet send including keepalives 12:36 < ws-client> <4ipset> what is it 12:37 < ws-client> you put all the if statements with your reply logic into the keep alive branch 12:38 < ws-client> <4ipset> bot have ping 999 for no reason 12:39 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207652381320548363/image.png?ex=65e06cec&is=65cdf7ec&hm=703e535c6a6e6dd4010bd3c31de5d5b55d4169666dfa356af19012ec71127176& 12:39 < bridge> I removed if (time.time() - self.last_send_time) > 0.1 12:40 < ws-client> you wanna keep the keepalive 12:40 < ws-client> just put your logic outside of the if statement like all the other if statements 12:41 < ws-client> and lmao there is a bug in the parser my bad 12:42 < ws-client> <4ipset> the bot still cant write letters 12:42 < bridge> <4ipset> https://cdn.discordapp.com/attachments/293493549758939136/1207653138442747955/image.png?ex=65e06da1&is=65cdf8a1&hm=0a8e52b7f5808082c8b10f17362e003b99cd666bf19ffae9bc37392e73b5952c& 12:43 < bridge> he answers for a while and then stops 12:43 < ws-client> oh and you overwrite msg 12:44 < ws-client> wasnt a bug in my parser you just confused me 12:44 < ws-client> you send a chat message and store it in ``msg =`` 12:44 < ws-client> then you do a if statement based on msg as if it was a server sent message 12:44 < bridge> Ouch 12:45 < bridge> <4ipset> https://cdn.discordapp.com/attachments/293493549758939136/1207653969753800765/image.png?ex=65e06e67&is=65cdf967&hm=8830e4aceaca5f7ddf7d1b5e225f4a326998be4da01898999693aa0228905420& 12:48 < bridge> @va_ll You are creating bot in python? xdd 12:49 < bridge> Then I want to connect it to my discord bot 12:49 < bridge> Lel 12:49 < bridge> <4ipset> they want to create bot with ai 12:49 < ws-client> https://zillyhuhn.com/cs/.1707997758.png 12:49 < ws-client> ah your code is chatgpt? 12:49 < ws-client> that explains things xd 12:49 < bridge> <4ipset> no 12:49 < bridge> Huh 12:50 < ws-client> https://paste.zillyhuhn.com/A1 12:50 < ws-client> lmao i got A1 paste like AI 12:50 < ws-client> i made your code somehow functional 12:50 < ws-client> ah i swapped out the ip to localhost 12:50 < ws-client> you might wanna change that back 12:52 < ws-client> https://zillyhuhn.com/cs/.1707997906.png 12:52 < bridge> oh thank you he started answering 12:52 < ws-client> no code shaming and all but like watfk bro 12:52 < ws-client> <:poggers2:1008007455936094328> 12:53 < bridge> Hmmm 12:53 < ws-client> and a new bot dev is born poggies 12:53 < bridge> If 3 is greater that three I will get Dev role in ddnet 12:54 < bridge> If 3 is greater that 3 I will get Dev role in ddnet 12:54 < ws-client> int(dev) 12:54 < bridge> <4ipset> no i mean that it will be a bot with artificial intelligence that will answer questions in the chat 12:55 < ws-client> i see 12:55 < bridge> By sing openai API? 12:55 < bridge> By using openai API? 12:55 < bridge> <4ipset> i think yes 12:55 < ws-client> i should really build an chat bot example idk why i did not do it yet. 12:55 < bridge> <4ipset> we already try to build it 12:55 < bridge> Yea add another repo to you 165 12:55 < ws-client> nah just a file in examples/ folder 12:56 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207656644855472158/image.png?ex=65e070e5&is=65cdfbe5&hm=eea0b5f157f745de2dcb2ae13de4af1fb87158be3e7ac056d605b654b8eae885& 12:56 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207656668398358548/image.png?ex=65e070ea&is=65cdfbea&hm=6da18dfbb59b8aec531a136f34675d54d1386e388453f41f741e9e12ced5cf3c& 12:56 < bridge> again he doesn't like something 12:57 < ws-client> did you touch the code? :D 12:57 < bridge> Not 12:57 < bridge> I just changed the IP and port 12:57 < ws-client> ah i can reproduce 13:02 < bridge> @ryozuki `error: internal compiler error: compiler/rustc_middle/src/ty/generic_args.rs:898:9: type parameter `U/#1` (U/#1/1) out of range when substituting` 13:02 < bridge> 13:02 < bridge> found a compiler bug in rust stable 13:02 < bridge> 13:02 < bridge> 13:02 < ws-client> seems like it does not auto calc size 13:02 < bridge> epic compiler moment 13:02 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207658186048413766/image.png?ex=65e07254&is=65cdfd54&hm=3e8525318555669ea06e47261ce0e8fd8f502f152785ca1c337cbae4048ed720& 13:03 < bridge> only responds to my first message 13:03 < ws-client> yes the message length calc is off 13:05 < bridge> If you create a bot, either ask the server provider for their consent and mark it as bot 13:06 < bridge> this is the server @4ipset 13:07 < bridge> I'm the admin there 13:07 < bridge> Fair enough. I was talking about a general process like that 13:08 < bridge> If you are an admin you can approve yourself, but also mark it as bot, otherwise players might think that someone is on your server 13:08 < ws-client> @va_ll yea im too lazy to fix that now but thats a bug on my end thanks for testing 13:08 < bridge> avo is scarred by iF xD 13:08 < bridge> I dont care about iF tbh 13:08 < bridge> Its more the player experience 13:09 < bridge> this is a temporary nickname until the bot is not fully operational yet 13:11 < ws-client> oh i think its a pyson momemnt tbh 13:11 < ws-client> where it remember last constructor arguments 13:11 < ws-client> or something like that 13:14 < bridge> You can also announce `"is_player": false,` in the HTTP register. But to do this you need to rewrite some code# 13:17 < ws-client> @va_ll okay so for now you can only ever send the same message length. Or set ``msg.header.size`` manually but its a bit trappy. I will not fix it today but its on my todo now. 13:18 < bridge> Fine 13:23 < bridge> hi 13:24 < bridge> hi 13:27 < bridge> :poggers2: 13:29 < bridge> <4ipset> hey 13:29 < bridge> нi 13:32 < bridge> wazzup 13:37 < bridge> the moon 13:44 < bridge> No it's down 13:44 < bridge> it’s up right now 13:44 < bridge> Where's up? 13:44 < bridge> above me 13:45 < bridge> You are standing sideways 13:46 < bridge> definitely not 13:46 < bridge> If you're a line and you take the normal at the height of your feet and then draw a plane is it on the upper side or the other oeb 13:46 < bridge> If you're a line and you take the normal at the height of your feet and then draw a plane is it on the upper side or the other one 13:46 < bridge> If you're a line and you take the normal at the height of your feet and then draw a 3d plane is it on the upper side or the other one 13:47 < bridge> The moon is above me dude 13:47 < bridge> If you're a line and you take the normal at the height of your feet and then draw a 2d plane is it on the upper side or the other one 13:47 < bridge> Ah ok cool 13:47 < bridge> What direction are you pointing in? 13:47 < bridge> UP 13:47 < bridge> Hmmm. 13:48 < bridge> away from the gravitational pull we describe as DOWN 13:48 < bridge> Woah cool so the moon is above me too? 13:49 < bridge> Since it is not inside the planet? 13:49 < bridge> It is in that direction 13:49 < bridge> Below 13:49 < bridge> Moon is down for u 13:49 < bridge> What is moon for earth? 13:49 < bridge> Below or up? 13:50 < bridge> Depends on where you make the measurement 13:50 < bridge> there is no objective up or down in space 13:50 < bridge> you need an anchor of some kind 13:50 < bridge> I believe the earth isn't perfectly circular so highest point is the head 13:50 < bridge> Ok 13:50 < bridge> /anchor 13:51 < bridge> *anchor 13:51 < bridge> Not really 13:51 < bridge> That’s not how anyone defines a head 13:51 < bridge> It is the furthest point on a generally spherical mass 13:52 < bridge> My toes are furthest out from my center of mass so they must be my head 13:52 < bridge> Yea do the average of all points on the earth and then check which is furthest 13:52 < bridge> That is the truth 13:53 < bridge> Okay mount Everest with 8849m is the anchor 13:54 < bridge> So where is the moon rn? 13:54 < bridge> Above you? 13:54 < bridge> If so we need to know where you are 13:54 < bridge> It’s above me yes 13:54 < bridge> Not you 13:54 < bridge> It’s below you 13:55 < bridge> Also not sure what furthest point has to do with it 13:56 < bridge> So if we assume you are the moon and are represented as a 2d vector pointing from the center of the earth 13:56 < bridge> And I am the same but somewhere else where the moon is not above me 13:56 < bridge> I’m Not the moon 13:56 < bridge> I’m me 13:57 < bridge> So if we assume you are the moon and are represented as a 2d vector of angles pointing from the center of the earth 13:58 < bridge> Since the moon is above you I can say you are the moon +-90° 13:58 < bridge> No 13:58 < bridge> The moon is above me because it’s night time 13:58 < bridge> So in the current time you are almost the moon 13:58 < bridge> *+-90° 13:59 < bridge> * +-90° 13:59 < bridge> I am not the moon 13:59 < bridge> I’m not almost the moon 13:59 < bridge> But in my head you are and you can do nothing against it 14:00 < bridge> You rock 🪨 14:17 < bridge> C++23 still doesn't have compound literals... 14:23 < bridge> I know it worked in msvc once 14:23 < bridge> Idk if it still does 14:24 < bridge> you could pass vec2(x+x,y+y) as a reference to a function directly. If that's what you mean with compound literals 15:18 < bridge> what a cool yavascript feature. took only few hours to find it :pepeW: 15:18 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207692391759216641/image.png?ex=65e09230&is=65ce1d30&hm=35fdbf7db5256b998b5466f2dfd22a5e87a85512804e9accf7e784311d6ee1d8& 15:38 < bridge> use map) 15:38 < bridge> it can finally draw something :poggers2: (coordinates are fucked up tho) 15:38 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207697359425773628/image.png?ex=65e096d0&is=65ce21d0&hm=9269755dc9685e1e48c776e4c47a56b5d2f60848b21bfe21e89de32b00c660ee& 15:38 < bridge> xd, had something similar on pyson 15:39 < bridge> it was just weird that it went thru loop one time and all 50k elements have the same value :justatest: 15:44 < bridge> Had something similar on lua 15:44 < bridge> And this was memory leak, cuz gc in lua 5.4 is shit, LuaJIT better 15:54 < bridge> To quote ryo. Wouldn't have happened with rust 15:55 < bridge> U'd have need to call clone for each and the type would be Rc 15:55 < bridge> Watcha working on 15:56 < bridge> https://github.com/MilkeeyCat/ddnet_demo_web_viewer :justatest: 15:56 < bridge> Epic 15:56 < bridge> i can somehow parse data 15:56 < bridge> but cant draw anything 15:57 < bridge> Does your loader support glsl es? 15:57 < bridge> That's cool 15:57 < bridge> am looking at ddnet graphics code and making it in js 15:59 < bridge> Do you already know typescript? 15:59 < bridge> yea xd 15:59 < bridge> I c 16:00 < bridge> https://github.com/DDNet-Team-Searcher/website is in ts 😏 16:28 < bridge> what was the url to ddnet client compiled to wasm> 16:28 < bridge> what was the url to ddnet client compiled to wasm? 16:29 < bridge> Ddnet.github.io 17:00 < bridge> xd its on version 15.9.1 17:03 < bridge> Yeah nobody uses it anyway xd 17:03 < bridge> true 17:03 < bridge> Not wasting time with updating 18:51 < bridge> when the hint hints the hint 18:51 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207746105190654054/image.png?ex=65e0c436&is=65ce4f36&hm=ddade555eb73c5ad0582bcfab4a40003491ff163b8db65dc272750bfbb68270b& 19:00 < bridge> server auto kicks afk players without a timer when it hits player cap feature when 19:02 < bridge> today i learned floats are 2 to 5 times slower than integers 19:04 < bridge> in what task? 19:04 < bridge> multiplication 19:04 < bridge> so when is 2 times slower and when 5 times? 19:05 < bridge> at multiplication 19:05 < bridge> *according to llm 19:05 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207749630914203728/image.png?ex=65e0c77e&is=65ce527e&hm=922814f4de424382658e34397e675617830e8ec30f5e3372c25da3fe551b4332& 19:06 < bridge> last few months i dont trust any word chat gpt says 19:07 < bridge> its very bad at coding ik 19:07 < bridge> xdd 19:09 < bridge> and to my tests the higher the floating point numbers are the slower it gets 19:10 < bridge> @jupeyy_keks when we finally convert all those vec2s into ivec2s in ddnet code we can do 512 players xd 19:11 < bridge> DDNet 2.0 19:11 < bridge> ask your chatgpt if that also applies for fixed point numbers at division 19:13 < bridge> spouting the same shit again 19:13 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207751502006063104/image.png?ex=65e0c93d&is=65ce543d&hm=94fd4fba7d1f24291312b088e101f30a114c7d95544969e2fff75d1a1252f113& 19:14 < bridge> it's not even gpt4 is it? 19:14 < bridge> nop i dont have xddd 19:14 < bridge> gpt3 is dumb af. better not to use it than use it xD 19:14 < bridge> true 19:14 < bridge> gpt4 at least sometimes is correct xD 19:14 < bridge> i don't trust a word from it when it comes to coding 19:15 < bridge> and it doesnt even know what optimization is.... 19:15 < bridge> and it cant find memory leaks ;( 19:16 < bridge> it doesn't even know what optimization is.... 19:16 < bridge> and it can't find memory leaks ;( 19:43 < bridge> i think both are wrong on x64 and on any high end cpu 19:43 < bridge> or most desktop cpu 19:43 < bridge> floating point division or multiplication implies 23 bits or 51 bits operation while integer one is 32 or 64 19:46 < bridge> hmm, 3 or 4 cycles for integer ones, 5 cycles for floating point ones 19:46 < bridge> https://www.agner.org/optimize/instruction_tables.pdf 19:47 < bridge> (for haswell architecture, already quite old with jupstar's standard) 19:48 < bridge> however, floating point division is faster (22-96 integer vs 10-24 for floating point) 19:49 < bridge> but i guess it also depends on how exactly the calculations are used 19:49 < bridge> 19:49 < bridge> in a tight simd sceneraio ofc every cycle matters 19:49 < bridge> 19:49 < bridge> but if u mix calculations with other stuff (pointer acces etc.) the calculations are often offloaded to the floating point unit while other stuff happens simultaniously 20:18 < bridge> does anyone have experience in analyzing discrete data? 20:18 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207767946399842314/image.png?ex=65e0d88d&is=65ce638d&hm=fb454624cb073ee7509f2be20325f8da417bc7302d2a7bffae39dec0f58d558e& 20:18 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207767946701836298/image.png?ex=65e0d88d&is=65ce638d&hm=f7d0b659b9b9b4bd9dbd31ec0155e8f363a7014d841000f33c265ac5aab865bf& 20:18 < bridge> I need to split the data into approximately the following intervals (segments with a beginning and an end time) 20:18 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1207767961939480597/image.png?ex=65e0d891&is=65ce6391&hm=08bc0129bd39e376c21fe013a0cbe3c4b2d41460e989b93f6e200a5326ff0183& 20:33 < bridge> fft, retrieve main frequency, sample at the correct interval 20:33 < bridge> you can also look at the derivative to detect sign changes 20:34 < bridge> which is where this kind of stuff might occur 20:34 < bridge> (but it's not 100%) 20:34 < bridge> Damn u woke chairn 20:34 < bridge> what you trying to achieve exactly? 20:34 < bridge> determination of useful work based on the current strength of the press 20:35 < bridge> determination of useful work based on the current generic of the press 20:35 < bridge> determination of useful work based on the generic current of the press 20:35 < bridge> shouldn't you try some computation and integration then ? 20:37 < bridge> now a very stupid algorithm has been made, to consider everything above a certain value as useful work, this would be enough, but on different days this limit either rises or falls, such an algorithm is not suitable, a more universal one is needed 20:37 < bridge> fft its Fast Fourier Transform ? 20:37 < bridge>