02:09 < bridge> @heinrich5991 when you have some time, would you like to go through all the PRs together and close/merge them? 02:15 < bridge> The amount of PRs is rather large, but still a manageable amount. We should get to it before it gets worse and while I'm still very much unemployed 02:37 < bridge> wait I can help 02:39 < bridge> :monkalaugh: 02:41 < bridge> What happened with that PR anyway, was supposed to be pretty simple :/ 02:41 < bridge> It even worked fine iirc 02:41 < bridge> it had some issues 02:42 < bridge> getting the entity ids was not as consistent as I expected 02:42 < bridge> so if two entities entered a teleport on the same tick and both of them didn't get an m_Id they would go to the same spot 02:43 < bridge> also seeding the prng with a small integer is probably not very good for it 08:50 < bridge> Start from the oldest 08:52 < bridge> The two oldest open prs are by this Heinrich and learath guy. Go tell them to get their pr fixed up and merge it 09:27 < bridge> morning C enjoyers and others 09:28 < bridge> morn 10:16 < bridge> im still a certified c++ dev(tm) 10:16 < bridge> :justatest: 10:19 < bridge> Is there some leetcode algorithm to do this without n^2 10:20 < bridge> well messages are max 256 chars 10:20 < bridge> I'm just wondering 10:20 < bridge> oh if u find tell me 10:20 < bridge> i just wanted to get this working(tm) 10:21 < bridge> maybe instead of a vec i could use a set 10:21 < bridge> are there hashsets in c++ 10:21 < bridge> std 10:22 < bridge> maybe iterating the message word by word and doing a set.contains is faster 10:22 < bridge> Yeah but you still pay the hashing price 10:22 < bridge> that would be just n iirc 10:22 < bridge> true 10:22 < bridge> is hashing that expensive tho 10:22 < bridge> what alg does std::set use 10:25 < bridge> @heinrich5991 is using a set better? 10:25 < bridge> saw ur comments 10:27 < bridge> I think you want a tree for the best complexity 10:28 < bridge> If you hash all the banned words you still need to check against the hash of all possible sub strings in the message that are shorter than the longest banned word 10:29 < bridge> which is kinda still n^2 I guess 10:29 < bridge> Maybe that makes the current solution n^3 10:29 < bridge> Well 10:30 < bridge> m*n^2 10:30 < bridge> according to docs std::set is a binary search tree 10:31 < bridge> i guess i need to figure how the json parsing works in ddnet now 10:37 < bridge> it's in general a red-black tree i think 10:37 < bridge> (which is a kind of binary tree) 10:42 < bridge> making it json makes it actually easier xd 11:07 < bridge> @totar actually i forgot a loop there lmao 11:09 < bridge> fixed and now uses json 11:09 < bridge> hot 11:10 < bridge> (without the loop it only censored the first matching same word) 12:03 < bridge> ``` 12:03 < bridge> [build] /Library/Developer/CommandLineTools/SDKs/MacOSX14.sdk/System/Library/Frameworks/ApplicationServices.framework/Frameworks/QD.framework/Headers/ColorSyncDeprecated.h:1000:41: note: 'CMProfileRef' has been explicitly marked deprecated here 12:03 < bridge> [build] 1000 | typedef struct OpaqueCMProfileRef* CMProfileRef DEPRECATED_IN_MAC_OS_X_VERSION_10_6_AND_LATER; 12:03 < bridge> [build] | 12:03 < bridge> ``` 12:03 < bridge> there is lot of msg like this when compiling in macos 12:04 < bridge> What even is the n here? I'd expect it to be something like m * n * k where n is the message length, m is the amount of words in the blacklist, and k is the average length of a blackliated word 12:04 < bridge> due to [build] In file included from /Users/edgarluque/Documents/misc/ddnet/src/macos/notifications.mm:1: 12:04 < bridge> xD 12:07 < bridge> Though there are more efficient ways to do it indeed. You can employ the Rabin-Karp search algorithm 12:08 < bridge> Or build an Aho-Corasick state machine 12:31 < bridge> i leave that for the future 12:37 < bridge> there is also https://en.wikipedia.org/wiki/Commentz-Walter_algorithm 12:55 < bridge> the pr is ready now btw 12:56 < bridge> well there is a log i can remove 12:57 < bridge> done 14:00 < bridge> what is it, just hash pattern matching? 14:03 < bridge> nerdy ass developer channel 14:03 < bridge> i dont think it should have both local and global 14:04 < bridge> why not just have online lists that have varying lebels of censor if ur able to change the online lists url anyways 14:04 < bridge> like censor_slurs, censor_all 14:16 < bridge> no 14:16 < bridge> i prefer to give control to the user 14:16 < bridge> and levels is subjective 14:18 < bridge> ye, then they can host their own censorlist if they really want that much control 14:19 < bridge> seems like an unnecessary additional setting 14:20 < bridge> u could also have a local json that has online: field updated by the online url and a local: that won't get overwritten 14:20 < bridge> @ryozuki does the censorlist support replacing the words? nerdy ass programming language 14:20 < bridge> so ppl who really care to have a local can edit it manually and those who dont care dont have to be exposed to multiple settings in the gui 14:24 < bridge> no users who dont want this just click 1 checkbox 14:24 < bridge> im not gonna change it 14:24 < bridge> u can make a pr, i think my way is the best 14:25 < bridge> it just replaces to * 14:27 < bridge> nerdy *** programming language 14:27 < bridge> ***** 14:27 < bridge> uh ok 14:27 < always> helo 14:27 < bridge> will that feature be enabled on the official servers? 14:27 < always> gg's @everyone 14:27 < always> WTF 14:27 < bridge> dont use spyware 14:28 < always> it removed my @ 14:28 < bridge> *\*\*\*\*\* 14:28 < bridge> its client sided 14:28 < bridge> oh 14:28 < bridge> im a dumbass 14:28 < bridge> xD 14:28 < bridge> then its cool 14:28 < bridge> if u want chat censor u can always go to chn servers 14:29 < bridge> server sided chat censor* 14:29 < bridge> is it enforced by the government? 14:29 < bridge> ye 14:29 < bridge> is there actually a censor?? 14:29 < bridge> classic china moment 14:29 < bridge> on chn there is server side censor 14:29 < bridge> gg 14:29 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1331979131998699623/DDNet_Vgju4GrwFp.png?ex=679395bc&is=6792443c&hm=a30c925d8f7d41c087563dbadf3a2746ea000da5151c2d8c37c54f4832ad7c61& 14:29 < bridge> u cant say taiwan number one 14:29 < bridge> i will test it when i get home 14:29 < bridge> theres no way 14:30 < bridge> i just lost so much social credit 14:41 < bridge> i am the chosen one 14:41 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1331982253810520126/image.png?ex=679398a4&is=67924724&hm=f7ff5aa5fcb880fe59f6f4cb752b5bbbecf4902968de0ab37bab73e20d5cf041& 14:42 < bridge> or am i supposed to say it in chinese 14:42 < bridge> xD 14:42 < bridge> will try again 14:42 < bridge> still works 14:42 < bridge> in chinese 14:44 < bridge> but the external male organ is being censored 14:49 < bridge> xd 14:49 < bridge> idk 15:05 < bridge> dream setup complere 15:05 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1331988176017031268/IMG_3730.jpg?ex=67939e28&is=67924ca8&hm=6c0270e1b6a05215b576e03ae5c7b9fe11f8fbb1f22883ad8c7be915eb0273e6& 15:05 < bridge> dream setup complete 15:05 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1331988176017031268/IMG_3730.jpg?ex=67939e28&is=67924ca8&hm=6c0270e1b6a05215b576e03ae5c7b9fe11f8fbb1f22883ad8c7be915eb0273e6& 15:05 < bridge> now i just need m4 mac mini 15:17 < bridge> well the censorship actually doesn't help lot tho 15:17 < bridge> since names are not censored 15:19 < bridge> and also censor word list is not enough 15:19 < bridge> it is enough for now 15:19 < bridge> so i'd rather not having it at all 15:19 < bridge> we can make it better always 15:19 < bridge> u can disable it so that works for you 15:19 < bridge> I mean the current chn ones 15:19 < bridge> oh ok 15:19 < bridge> idc about china xD 15:20 < bridge> ofc i should care being in china TT 15:20 < bridge> Give team viewer access pls. Need to git bisect smth on Apple silicon 15:22 < bridge> @jupeyy_keks send repro instructions 15:22 < bridge> if usend a git bisect script i can run it 15:23 < bridge> u know u can script it to automatically compile and test and check if its good or bad 15:28 < bridge> lmao i am not even sure yet how i'd reproduce it, it's just that newer moltenvk create some weird pink framebuffer: 15:28 < bridge> https://github.com/KhronosGroup/MoltenVK/issues/2220 15:28 < bridge> 15:28 < bridge> they know all commits where it happened but apparently nobody fixes it xd 15:28 < bridge> I'd maybe simply revert some commits: 15:28 < bridge> https://github.com/KhronosGroup/MoltenVK/compare/v1.2.6...v1.2.7 15:29 < bridge> i just hate apple, these mfs that refuse anything good and let randoms have to fix their shit. 15:29 < bridge> 15:29 < bridge> free slaves for apple 15:30 < bridge> Same should apply to many 15:30 < bridge> apple is defs the worst xd 15:30 < bridge> :nouis: 15:45 < bridge> US dream setup frfr 16:12 < bridge> @learath2 good read https://github.com/bminor/glibc/commit/7a61e7f557a97ab597d6fca5e2d1f13f65685c61 16:13 < bridge> * Do not deallocate the environment array. Instead, keep older 16:13 < bridge> versions around and adopt an exponential resizing policy. This 16:13 < bridge> results in an amortized constant space leak per active environment 16:13 < bridge> variable, but there already is such a leak for the variable itself 16:13 < bridge> (and that is even length-dependent, and includes no-longer used 16:13 < bridge> values). 16:13 < bridge> xd 16:13 < bridge> extern char** environ; 16:13 < bridge> this legacy thing is crazy 17:56 < bridge> Hot 18:03 < bridge> That's only the data repo tho 😂 18:03 < bridge> But yeah commit is in unmerged pr 18:04 < bridge> Idc, hot 18:04 < bridge> 18:04 < bridge> Btw are you planning on doing the freeze timer? I saw the .PNG but no impl 18:04 < bridge> Wanted to do it but I cba learning how to handle external pngs for now :justatest: 18:05 < bridge> you have to use the freeze container. 18:05 < bridge> game/client-containers/src/freezes.rs 18:06 < bridge> almost every asset is build around the container stuff 18:06 < bridge> Yeah I looked at how you did the hook but it's impl looked like a lot of boilerplate 18:06 < bridge> Not saying it's bad! But I was just to tired to do it 18:06 < bridge> :justatest: 18:08 < bridge> besides the texture or exactly the texture? 18:09 < bridge> Besides 18:10 < bridge> Can't reread it so I'm also not sure anymore :justatest: 18:10 < bridge> In bed rn 18:10 < bridge> sure 18:10 < bridge> but i think it should not be _much_ harder than the old ddnet. 18:10 < bridge> xddd 18:10 < bridge> 18:10 < bridge> he said 18:11 < bridge> GG you jinxed it! 18:23 < bridge> @milkeeycat u implemented generics in ur lang? 18:24 < bridge> not yet, i'm still not sure if I need to add one more ir xd 18:24 < bridge> ```rust 18:24 < bridge> struct A { 18:24 < bridge> a: i32, 18:24 < bridge> b: i32, 18:24 < bridge> } 18:24 < bridge> 18:24 < bridge> impl A { 18:24 < bridge> pub fn hello(&self, other: i32) -> i32 { 18:24 < bridge> return self.a * other; 18:24 < bridge> } 18:24 < bridge> 18:24 < bridge> pub fn set_a(&mut self, value: i32) { 18:24 < bridge> self.a = value; 18:24 < bridge> return; 18:24 < bridge> } 18:24 < bridge> } 18:24 < bridge> 18:25 < bridge> pub fn main() -> i32 { 18:25 < bridge> let mut x: A = A { 18:25 < bridge> a: 2, 18:25 < bridge> b: 3, 18:25 < bridge> }; 18:25 < bridge> 18:25 < bridge> x.set_a(4); 18:25 < bridge> 18:25 < bridge> return x.a; 18:25 < bridge> } 18:25 < bridge> ``` 18:25 < bridge> this compiles in the lang im making at work 18:25 < bridge> now i need to do generics 18:25 < bridge> epyc 18:25 < bridge> @milkeeycat how many irs u have rn 18:25 < bridge> i have AST -> ir -> llvm ir -> machine code 18:25 < bridge> ast and 1 ir 18:25 < bridge> the IR will be fully monomorphized already, this means that the ir itself has no generics, so i need to do the work on the ast 18:26 < bridge> can you guys stop being so smart 18:26 < bridge> god damn it. 18:26 < bridge> do you have type inference? 😬 18:26 < bridge> only a tiny bit 18:26 < bridge> its not the priority right now xd 18:27 < bridge> do you do it on ir or ast? 18:27 < bridge> ast, the IR is fully typed and has all the info needed 18:27 < bridge> to make codegen as easy as possible 18:27 < bridge> but for example, i think rust has 3 irs 18:27 < bridge> AST THIR MIR LIR 18:28 < bridge> and i think it does type checking on THIR 18:28 < bridge> where's hir? 🤨 18:28 < bridge> oh THIR and hir are probs same 18:28 < bridge> actually THIR may be something like type higher ir or smth 18:28 < bridge> hm, i remember it being a separate thing 18:28 < bridge> https://rustc-dev-guide.rust-lang.org/thir.html 18:28 < bridge> The THIR ("Typed High-Level Intermediate Representation"), previously called HAIR for "High-Level Abstract IR", is another IR used by rustc that is generated after type checking. It is (as of January 2024) used for MIR construction, exhaustiveness checking, and unsafety checking. 18:29 < bridge> 18:29 < bridge> As the name might suggest, the THIR is a lowered version of the HIR where all the types have been filled in, which is possible after type checking has completed. But it has some other interesting features that distinguish it from the HIR: 18:29 < bridge> isnt this rust 18:29 < bridge> ok so i guess it does type checking on the ast 18:29 < bridge> looks nearly same ye xd 18:29 < bridge> so whats the goal 18:29 < bridge> but this lang will have an embedded async runtime, like erlang 18:29 < bridge> i'm confused, is this still edlang? or something new 18:29 < bridge> the goal is a mix of erlang and rust 18:29 < bridge> @blaiszephyr im also making a language at work 18:30 < bridge> oh right embedded async runtime 18:30 < bridge> lmao 18:30 < bridge> cool 18:30 < bridge> in rust u have to choose 18:30 < bridge> tokio for example 18:30 < bridge> it divides a bit the community 18:30 < bridge> async rust is hell 18:30 < bridge> also tokio needs to be cooperstive, it cant be preemptive 18:31 < bridge> the coop part is the .await points 18:31 < bridge> is where u yield 18:31 < bridge> to be preemptive u need compiler control 18:32 < bridge> u smart af 18:32 < bridge> i dont understand anything u said 18:32 < bridge> xd 18:32 < bridge> :cammo: 18:33 < bridge> @kollpotato you know what a thread is right 18:33 < bridge> you know green threads? 18:34 < bridge> no 18:34 < bridge> a thread that is scheduled by a runtime library or virtual machine (VM) instead of natively by the underlying operating system (OS). 18:36 < bridge> I love your funny words magic man 18:36 < bridge> xD 18:36 < bridge> u are writing a compiler yourself 18:36 < bridge> well processes execute in a thread, and in a cpu core, only one thread at a time can actually run (simplified for explanation), but the kernel needs to manage how much cpu time each process uses, otherwise one program would run always while the others wait forever, for this, when you are in a preemptive runtime, the kernel forcefully yields your thread, giving control back to kernel and it can decide to give run time to another process 18:36 < bridge> u are the magic man too 18:36 < bridge> I'm a cave man magic man 18:36 < bridge> in a cooperative runtime u yield as a user 18:37 < bridge> in rust those are .await points 18:37 < bridge> i probs explained this wrong but idk xd 18:37 < bridge> @kollpotato read this https://en.wikipedia.org/wiki/Preemption_(computing) 18:37 < bridge> read lot of wiki articles 18:37 < bridge> i do this all the time 18:37 < bridge> now i understand 50% instead of 0% 18:37 < bridge> thanks smart man 18:38 < bridge> ive been reading a lot of them recently 18:38 < bridge> mostly about networking 18:38 < bridge> networking is fun 18:38 < bridge> one thread at a time per cpu or per cpu core? 18:38 < bridge> ah right 18:38 < bridge> cpu core 18:39 < bridge> so if i have 4 cores i can only run 4 threads at a time 18:39 < bridge> then i dont understand what i achieved with python 3 years ago when i tried running 1000 threads in parallel 18:39 < bridge> well there are weird stuff like hyperthreading which idk what does but ye 18:39 < bridge> > Intel® Hyper-Threading Technology is a hardware innovation that allows more than one thread to run on each core. More threads means more work can be done in parallel. How does Hyper-Threading work? When Intel® Hyper-Threading Technology is active, the CPU exposes two execution contexts per physical core. 18:40 < bridge> you have to know the difference between parallel and concurrent 18:40 < bridge> is it like fetching and executing instruction of one thread 18:40 < bridge> and then going to another one 18:40 < bridge> yes 18:40 < bridge> to me these words are the same 18:40 < bridge> > A system is said to be concurrent if it can support two or more actions in progress at the same time. A system is said to be parallel if it can support two or more actions executing simultaneously. The key concept and difference between these definitions is the phrase “in progress.” 18:41 < bridge> got it 18:42 < bridge> tokio can be concurrent using a single thread 18:44 < bridge> @kollpotato btw threading is one of the hardest and most interesting stuff in computing, its why u need locks, mutexes, atomics, and need to know well the memory model 18:44 < bridge> i find most of the things in computer science interesting 18:45 < bridge> but i will still avoid async rust for a while 18:45 < bridge> https://doc.rust-lang.org/nomicon/atomics.html 18:45 < bridge> well thats just because i dont need it 18:45 < bridge> ive read that btw 18:45 < bridge> oh yeah u dont need async for threading xd 18:45 < bridge> there was a post learath sent me years ago 18:45 < bridge> but i forgot 18:46 < bridge> @kollpotato implement this https://en.wikipedia.org/wiki/Work_stealing 18:46 < bridge> good for learning 18:47 < bridge> :justatest: 18:47 < bridge> do they teach this in uni 18:47 < bridge> this is what tokio does 18:47 < bridge> idk i didnt go to uni 18:48 < bridge> i will 18:48 < bridge> probably 18:49 < bridge> do it if u can 18:49 < bridge> how old r u 18:49 < bridge> 16 18:49 < bridge> just 3 more years in school 18:49 < bridge> u got lot of.time ahead 18:51 < bridge> i gotta write my own compiler 18:51 < bridge> before uni 18:51 < bridge> thats has been my dream 18:51 < bridge> for 3 years already 18:51 < bridge> @ryozuki what do you do with ir before it's lowered to llvm ir? 18:51 < bridge> that has been my dream 19:00 < bridge> @ryozuki btw this is kinda cool https://edgl.dev/blog/rust-fn-size-trick/ 19:00 < bridge> but why cant the compiler optimize that? 19:01 < bridge> nothing 19:02 < bridge> :thonk: 19:02 < bridge> the ir is block based and kinda like ssa 19:02 < bridge> so it makes it rly ez to translate to llvm ir 19:03 < bridge> it doesnt have if else it has (cond) jumps s between blocks 19:03 < bridge> and blocks have no control flow within 19:03 < bridge> like llvm ir 19:04 < bridge> maybe one day I'll make a generator which will spit out llvm ir in the end 19:15 < bridge> Then what do you generate now 19:15 < bridge> assembly 19:15 < bridge> :poggers2: 19:15 < bridge> when machine code 19:15 < bridge> assembly too high levle 19:15 < bridge> assembly too high level 19:18 < bridge> I got modulo in automap working! 19:18 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1332051778413265006/2025-01-23_13-17-32.mp4?ex=6793d964&is=679287e4&hm=1b69e94cdb4dd32f7f091178bbe3d6507bd76e9cf20b4826f4dd11b43cacb8c9& 19:23 < bridge> Well, kind of. Let me explain: 19:23 < bridge> The logic is inverted. It's supposed to place tile 49 when the modulo is equal to 0, not the opposite. 19:23 < bridge> However, when trying to set it that way, it only places down tile 49 (meaning it's always true). 19:23 < bridge> 19:23 < bridge> Here's what I put down for the code. In reference to auto_map.cpp 19:23 < bridge> ```cpp 19:23 < bridge> void CAutoMapper::Proceed(CLayerTiles *pLayer, CLayerTiles *pGameLayer, int ReferenceId, int ConfigId, int Seed, int SeedOffsetX, int SeedOffsetY) 19:23 < bridge> // [ ... ] 19:23 < bridge> for(size_t k = 0; k < pIndexRule->m_vModuloRules.size() && RespectRules; ++k) 19:23 < bridge> { 19:23 < bridge> CModuloRule *pModuloRule = &pIndexRule->m_vModuloRules[k]; 19:23 < bridge> if((x + pModuloRule->m_OffsetX) % pModuloRule->m_ModX == 0 && (y + pModuloRule->m_OffsetY) % pModuloRule->m_ModY == 0) 19:23 < bridge> { 19:23 < bridge> RespectRules = false; 19:23 < bridge> } 19:23 < bridge> } 19:23 < bridge> 19:23 < bridge> // already defined code below, just for context 19:23 < bridge> if(RespectRules && 19:23 < bridge> (pIndexRule->m_RandomProbability >= 1.0f || HashLocation(Seed, h, i, x + SeedOffsetX, y + SeedOffsetY) < HASH_MAX * pIndexRule->m_RandomProbability)) 19:24 < bridge> { 19:24 < bridge> CTile Previous = *pTile; 19:24 < bridge> pTile->m_Index = pIndexRule->m_Id; 19:24 < bridge> pTile->m_Flags = pIndexRule->m_Flag; 19:24 < bridge> pLayer->RecordStateChange(x, y, Previous, *pTile); 19:24 < bridge> }``` 19:24 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1332053089334595604/image.png?ex=6793da9c&is=6792891c&hm=ba792bd5fc7a4a58b3a6060d3f276b8dfc2bcb870781a293162f3a9514c3f055& 19:25 < bridge> Like, I KNOW it's in this bit of code: 19:25 < bridge> ```(x + pModuloRule->m_OffsetX) % pModuloRule->m_ModX == 0 && (y + pModuloRule->m_OffsetY) % pModuloRule->m_ModY == 0)``` 19:25 < bridge> But if I set the == to != it doesn't set RespectRules to false anywhere. 19:30 < bridge> Well, setting it to !=, and then changing the && to a || works, kind of, but it doesn't accept more than 1 modulo factor that doesn't overlap. 19:31 < bridge> sweet 19:42 < bridge> oh i have an idea actually 19:49 < bridge> nvm 19:54 < bridge> u could probably do it if that's the only goal 19:55 < bridge> it would be really cool to see 20:04 < bridge> am getting 20:04 < bridge> ``` 20:04 < bridge> Error running link command: no such file or directorymake[2]: *** [CMakeFiles/steam_api.dir/build.make:101: libsteam_api.a] Error 2 20:04 < bridge> make[1]: *** [CMakeFiles/Makefile2:399: CMakeFiles/steam_api.dir/all] Error 2 20:04 < bridge> make[1]: *** Waiting for unfinished jobs.... 20:04 < bridge> ``` 20:04 < bridge> when compiling ): 20:04 < bridge> needs make 20:04 < bridge> but also wtf 20:04 < bridge> what command invoked this 20:04 < bridge> never seen that 20:05 < bridge> lmfao pkg broken and im unable to fix it myself so i go to google 20:05 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1332063581419733106/image.png?ex=6793e462&is=679292e2&hm=0049460077e8c2876fe5db0ebba8b45b23f6ee37ba6aa8bf328d073c134956c9& 20:05 < bridge> 1 minute ago 20:05 < bridge> i did cmake ..; make -j24 20:05 < bridge> try not make 20:05 < bridge> okay.... 20:05 < bridge> do u have ninja or something 20:06 < bridge> -GNinja 20:06 < bridge> i do now 20:06 < bridge> and i will try 20:07 < bridge> 👀 20:07 < bridge> works 20:10 < bridge>