00:10 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1203115242922582036/GFWGSw5WgAAtstB.png?ex=65cfeb62&is=65bd7662&hm=8a8b84bed72ac6801c71fd5542760096158315bce6efc7d8a20fe19bc6bb279c& 00:10 < bridge> :gigachad: 00:10 < bridge> :gigachad: 00:11 < bridge> What sum? 00:19 < bridge> cumulative 00:19 < bridge> ah, it says so 00:21 < bridge> @murpi ❤️ 00:22 < bridge> Can't wait for things to break left and right again 00:22 < bridge> Testing the bot changes was incredibly difficult. You can't even change a map channel twice without being rate-limited 02:23 < bridge> well, turkey kebab 02:23 < bridge> :gigachad: 02:23 < bridge> they built different fr 02:45 < bridge> @milkeeycat: ddnet does not appreciate your insane coding skills. Don‘t worry they will get to know you. 02:45 < bridge> 02:45 < bridge> We the discord-irc team thank you for your amazing services! We could not have done it without you. 05:41 < ws-client> htoprc is the greatest thing in the world 05:41 < ws-client> https://zillyhuhn.com/cs/.1706935266.png 07:22 * ChillerDragon a 08:15 < bridge> ChillerDragon: is there any updated version of the ddnet 0.7 client besides the headless one? I've been using the 17.2.1 release to play vanilla as a usa player with eu players but I can't see votes, which I see is now crossed off on the list of bugs. 08:47 < ws-client> woah @pisstine i did not expect anybody to actually use that how did you even hear of it :D 09:04 < bridge> I saw a vid you did of how to make a server dummy n make them do things n then saw the ddnet7 download thing after looking on github for a bit, its very helpful 09:05 < ws-client> common github stalking W 09:05 < ws-client> @pisstine windows user i assume? 09:07 < ws-client> I made a new release extra for you <:heartw:395753947396046850> @pisstine 09:07 < ws-client> https://chillerdragon.github.io/ddnet7-client/ 09:10 < bridge> chillerdragon thx 🥹 09:14 < bridge> oh yeah windows user, thank u very much!!! 09:29 < bridge> mmm 10:10 < bridge> lol 10:10 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1203266229175844924/GFXhwmCacAAWw4Z.png?ex=65d07800&is=65be0300&hm=55af00806eefd07c1c2e37e98769bd805dc4dcebbd0343aae7196a8219011e94& 10:10 < bridge> @tsfreddie take notes of ur weird lang 10:11 < bridge> old news 10:20 < bridge> @learath2 https://neovide.dev/index.html 10:20 < bridge> it has cursor animations 10:20 < bridge> which are impossible in a normal terminal 10:21 < bridge> cuz terminals dont rly know what a cursor is 10:21 < bridge> it makes following vim motions nice 11:03 < bridge> wdym? Terminals definitely know what a cursor is :O 11:04 < bridge> I'm sure you could port that animated cursor to Alacritty or Kitty if you really wanted 11:09 < ws-client> Yea implementing a shell and a terminal emulator is on my bucket list for sure. To fully understand who is responsible for what. 11:09 < bridge> i made a function to convert hsla to int, but i think it's broken o_O 11:09 < bridge> this is the function 11:09 < bridge> ```cpp 11:10 < bridge> int CSkyb::HSLAToInt(std::uint8_t H, std::uint8_t S, std::uint8_t L, std::uint8_t A)//convert hsla color to int 11:10 < bridge> { 11:10 < bridge> std::uint32_t color = 0; 11:10 < bridge> color += H; 11:10 < bridge> color += S * 0Xff; 11:10 < bridge> color += L * 0Xffff; 11:10 < bridge> color += A * 0Xffffff; 11:10 < bridge> return static_cast(color); 11:10 < bridge> } 11:10 < bridge> ``` 11:10 < bridge> it's giving me rainbow kinda effect when i call it this way. 11:10 < bridge> ```cpp 11:10 < bridge> pPlayer->m_TeeInfos.m_ColorFeet = skyb.HSLAToInt(0,255, snum++, 0); 11:10 < bridge> ``` 11:10 < bridge> note: sum goes 0 to 255 in loop. 11:10 < bridge> can someone help ? 11:15 < bridge> We literally already have this in the code 11:15 < bridge> looks like you're not correctly scaling it correctly 11:15 < bridge> looks like you're not correctly scaling it 11:15 < bridge> Use `ColorHSLA` and the `Pack` function 11:15 < bridge> oh wait this is for ddnet - ah gg 11:16 < bridge> You also have the components all in the wrong places. Iirc we have our A on the MSB 11:16 < bridge> Oh wait. Those are multiplications 11:17 < bridge> hm 11:17 < bridge> I remember that 11:17 < bridge> Seems buggy because `A * 0Xffffff` is promoted to signed int, which can overflow. It should be `A * 0Xffffffu` to avoid any signed types in the calculation 11:17 < bridge> there is a mod do this 11:18 < bridge> Just use ColorHSLA::Pack, we already figured all this out for you 😄 11:18 < bridge> https://github.com/bonobo-craft/ddnet-infclass/blob/dcb706756c8e50acf22bcbd5e68a7428c24d4064/src/infcroya/skin.cpp#L154-L162 11:18 < bridge> i couldn't get rgb effect with ColorHSLA and Pack, can some one give me the exact code ? 11:19 < bridge> here is a HSLAtoInt 11:19 < bridge> Colors values are between 0.0f and 1.0f with our color functions 11:19 < bridge> Not between 0-255 11:20 < bridge> here is a HSLAtoInt (for 0.7, but also could use for 0.6) 11:20 < bridge> here is a HSLAtoInt (for 0.7, but also be able to use for 0.6) 11:21 < bridge> thank you ❤️ 11:22 < bridge> o_O 11:22 < bridge> colors looks complicated 11:22 < bridge> But these are skin colors 11:23 < bridge> For those he should use the `Pack` function 11:23 < bridge> But he likely initializes the `ColorHSLA` with 0-255 values, which are out of range 11:24 < bridge> by 0.6 you mean teeworlds and 0.7 DDNet ? 11:25 < bridge> No 11:25 < bridge> Teeworlds 0.7 11:26 < bridge> is the page you gave me works with DDNet ? 11:26 < bridge> Yes (for player skin) 11:26 < bridge> Yes (only for player skin) 11:27 < bridge> Those colors are in 0-255 values 11:29 < bridge> What's the CNetMsg_Sv_CommandInfoGroupStart use for? 11:29 < bridge> I only see it send on server 11:30 < bridge> Did it just a way to tell developer "There is a group command"? 11:31 < bridge> > This allows custom clients to understand when a list is complete, instead of relying on heuristics such as timeouts. 11:31 < bridge> #7764 11:31 < bridge> https://github.com/ddnet/ddnet/pull/7764 11:51 < bridge> But there isn't any things client to do 11:51 < bridge> i dont think so i read it somewhere 11:52 < bridge> freebsd terminal has mouse support 11:52 < bridge> epic 11:52 < bridge> hmm i read somewhere that making the motions neovide does is hard on a classic terminal 11:52 < bridge> any terminal has mouse 11:52 < bridge> oh wait 11:53 < bridge> well vim has mouse support so most terms do 11:53 < bridge> i mean the tty 11:53 < bridge> https://neovide.dev/assets/AnimatedCursor.gif 11:53 < bridge> i dont have mouse support there 11:53 < bridge> https://neovide.dev/assets/SmoothScrolling.gif 11:54 < bridge> kuhl 11:54 < bridge> welcome to 2024 11:55 < bridge> i hate smooth scrolling btw xd 11:55 < bridge> on 240hz it's useless 11:56 < bridge> smooth scrolling here means its not line based 11:56 < bridge> ah ok 11:56 < bridge> i think its better than line based 11:56 < bridge> u can follow lines easier 11:56 < bridge> i thought it scrolls smooth from line to line ^^ 11:57 < bridge> @ryozuki have u ever written a custom serde serializer? 11:57 < bridge> yeah but nothing fully 11:57 < bridge> why 11:57 < bridge> is it hard? 11:57 < bridge> shouldnt be hard as long as u can map well to the serde data model 11:57 < bridge> i kinda want to try out xd 11:57 < bridge> https://serde.rs/data-model.html 11:57 < bridge> do it 11:57 < bridge> its fun 12:09 < bridge> why im finding out you can add `u` literal to hex function on ddnet discord, i never seen it mentioned on any resource i read 😭 12:09 < bridge> why im finding out you can add `u` literal to hex on ddnet discord, i never seen it mentioned on any resource i read 😭 12:09 < bridge> https://gist.github.com/edg-l/11d096e694fd38bd71337a9abc03592d 12:09 < bridge> epic edlang ir 12:09 < bridge> ```rust 12:09 < bridge> mod Main { 12:09 < bridge> 12:09 < bridge> fn main(argc: i32) -> i32 { 12:09 < bridge> let mut x: i32 = 2; 12:09 < bridge> x = 4; 12:09 < bridge> let y: i32 = other(2); 12:09 < bridge> return x; 12:09 < bridge> } 12:09 < bridge> 12:09 < bridge> fn other(a: i32) -> i32 { 12:09 < bridge> return a; 12:09 < bridge> } 12:09 < bridge> } 12:09 < bridge> 12:09 < bridge> ``` 12:09 < bridge> this is the program 12:09 < bridge> this looks even less readable than rust:pepeW: 12:09 < bridge> dude its a IR its not mean to be readable 12:09 < bridge> its meant to be made so its easy to use on the codegen 12:09 < bridge> what IR stands for 12:09 < bridge> intermediate representation? 12:10 < bridge> will read :okSanya: 12:10 < bridge> yes 12:10 < bridge> https://en.wikipedia.org/wiki/Intermediate_representation 12:10 < bridge> a complex compiler goes like 12:11 < bridge> source code -> lex / parse -> AST -> IR (there can be like diferent levels of ir, for instance, rust has THIR, HIR, MIR) -> codegen 12:12 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1203297037743955998/image.png?ex=65d094b1&is=65be1fb1&hm=d20a3590a206c0842d3532ae23e5ea0e8610b518d7d95bea2148e7f3adc249c2& 12:13 < bridge> Rust but not rust 12:13 < bridge> @learath2 yeah xd 12:13 < bridge> when i got smth more solid ill deviate from this 12:13 < bridge> Bro is so addicted to rust his own language looks like rust 12:14 < bridge> dis website is so cool https://astexplorer.net/ 12:14 < bridge> I use Kitty and if I go to insert mode, the cursor switches from a block to a line 12:15 < bridge> It should be part of the xterm "standard" 12:15 < bridge> https://github.com/edg-l/edlang/blob/master/lib/edlang_ir/src/lib.rs 12:15 < bridge> all my code is free software u can find it here 12:15 < bridge> agpl 12:16 < bridge> the repo is a bit mess cuz im transitioning from ast -> codegen to ast -> ir -> codegen 12:16 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1203298022071476264/image.png?ex=65d0959c&is=65be209c&hm=c209954bd5e134a0abe262117ffd7e30ef86ca6060285903b69165294ef1206a& 12:16 < bridge> totally not copying rustc structure 12:17 < bridge> Kitty neovim cursor stuff 12:17 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1203298168633040907/6HPRZN2.mp4?ex=65d095bf&is=65be20bf&hm=09cdd6abaf7d14ad80c4ebe1d127b9deeb140a41d08abf8d4aea86a755b4437b& 12:17 < bridge> when someone asks for ur homework and you tell them make it different so it doesnt look the same 12:17 < bridge> Also I just remembered that it also uses underline in replace mode 12:17 < bridge> I just remembered that it also uses underline in replace mode 12:18 < bridge> i use alacritty 12:18 < bridge> is kitty made with rust? no 12:19 < bridge> thats the answer 12:19 < bridge> :gigachad: 12:19 < bridge> Kitty has some nice features that Alacritty still doesnt have 12:19 < bridge> i hope alacritty uses vulkan someday 12:19 < bridge> like a proper image protocol and ligatures 12:19 < bridge> nice 12:19 < bridge> i dont use em 12:19 < bridge> xD 12:19 < bridge> i dont even use alacritty vim mode 12:20 < bridge> I love ligatures though :O 12:20 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1203299030151602176/2iDDAcU.png?ex=65d0968d&is=65be218d&hm=071b5849eae7ae6e9d27a2a30539e92880f4cea69260169b19c7a8dc87dd5580& 12:20 < bridge> i dont 12:21 < bridge> i like seeing exactly 12:21 < bridge> what i type 12:21 < bridge> its confusing af xd 12:21 < bridge> I'm not a ligature enjoyer either 12:21 < bridge> for rust they should add a fish 12:21 < bridge> for ::<> 12:22 < bridge> @learath2 btw rust turbofish is made to avoid c++ mistake 12:23 < bridge> Who is a turbofish? 12:23 < bridge> https://scrumplex.rocks/img/1706959325_Uoj1Ne.png 12:23 < bridge> c++ uses Vec::xx 12:23 < bridge> ^^ 12:23 < bridge> rust requires 12:23 < bridge> Vec::::new() 12:24 < bridge> > The Vec:: syntax is what C++ uses. why doesn't that work with Rust and requires such odd syntax (Vec::::) ? 12:24 < bridge> Images in terminal? What kinda blasphemy? 12:24 < bridge> https://www.reddit.com/r/rust/comments/v4rir6/turbofish_why/ 12:24 < bridge> > Because it means that the meaning of < depends on whether the preceding symbol refers to a template, or not, which requires either: 12:24 < bridge> > 12:24 < bridge> > A complex parsing algorithm, such as GLR. 12:24 < bridge> > 12:24 < bridge> > A parsing algorithm which mixes parsing and symbol look-up. 12:24 < bridge> https://sw.kovidgoyal.net/kitty/graphics-protocol/ :D 12:24 < bridge> https://github.com/rust-lang/rust/blob/master/tests/ui/parser/bastion-of-the-turbofish.rs 12:25 < bridge> This isn't Kitty exclusive either. WezTerm and I think KDE's Konsole also support the protocol 12:25 < bridge> ```rust 12:25 < bridge> // check-pass 12:25 < bridge> 12:25 < bridge> // Bastion of the Turbofish 12:25 < bridge> // ------------------------ 12:25 < bridge> // Beware travellers, lest you venture into waters callous and unforgiving, 12:25 < bridge> // where hope must be abandoned, ere it is cruelly torn from you. For here 12:25 < bridge> // stands the bastion of the Turbofish: an impenetrable fortress holding 12:25 < bridge> // unshaking against those who would dare suggest the supererogation of the 12:25 < bridge> // Turbofish. 12:25 < bridge> // 12:25 < bridge> // Once I was young and foolish and had the impudence to imagine that I could 12:25 < bridge> // shake free from the coils by which that creature had us tightly bound. I 12:25 < bridge> // dared to suggest that there was a better way: a brighter future, in which 12:25 < bridge> // Rustaceans both new and old could be rid of that vile beast. But alas! In 12:25 < bridge> // my foolhardiness my ignorance was unveiled and my dreams were dashed 12:25 < bridge> // unforgivingly against the rock of syntactic ambiguity. 12:25 < bridge> // 12:25 < bridge> // This humble program, small and insignificant though it might seem, 12:25 < bridge> // demonstrates that to which we had previously cast a blind eye: an ambiguity 12:25 < bridge> // in permitting generic arguments to be provided without the consent of the 12:26 < bridge> // Great Turbofish. Should you be so naïve as to try to revolt against its 12:26 < bridge> // mighty clutches, here shall its wrath be indomitably displayed. This 12:26 < bridge> // program must pass for all eternity: forever watched by the guardian angel 12:26 < bridge> // which gave this beast its name, and stands fundamentally at odds with the 12:26 < bridge> // impetuous rebellion against the Turbofish. 12:26 < bridge> // 12:26 < bridge> // My heart aches in sorrow, for I know I am defeated. Let this be a warning 12:26 < bridge> // to all those who come after: for they too must overcome the impassible 12:26 < bridge> // hurdle of defeating the great beast, championed by a resolute winged 12:26 < bridge> // guardian. 12:26 < bridge> this is a rustc test 12:26 < bridge> Hm, what is the other meaning of in C++? 12:27 < bridge> any type 12:28 < bridge> its using a templated type 12:28 < bridge> or wahtever 12:28 < bridge> like u got a generic class 12:28 < bridge> Oh, I just saw it in the reddit post, it can be interpreted as less than T greater than ... 12:28 < bridge> yeah 12:28 < bridge> I was wondering what the parsing ambiguity was 12:28 < bridge> in c++ to solve this 12:28 < bridge> they do like complex af parsing 12:28 < bridge> and symbol lookup 12:28 < bridge> iirc 12:29 < bridge> https://en.wikipedia.org/wiki/GLR_parser 12:29 < bridge> > A complex parsing algorithm, such as GLR. 12:29 < bridge> > 12:29 < bridge> > A parsing algorithm which mixes parsing and symbol look-up. 12:29 < bridge> Is rust even parsable by just an LR parser? 12:29 < bridge> iirc yes 12:30 < bridge> https://github.com/chuckcscccl/rustlr/ 12:30 < bridge> hm i wonder xd 12:31 < bridge> > ust-analyzer uses a hand-written recursive descent / Pratt parser 12:31 < bridge> > rust-analyzer uses a hand-written recursive descent / Pratt parser 12:32 < bridge> So it is LR huh cool 12:32 < bridge> > An operator-precedence parser is a simple shift-reduce parser that is capable of parsing a subset of LR(1) grammars. More precisely, the operator-precedence parser can parse all LR(1) grammars where two consecutive nonterminals and epsilon never appear in the right-hand side of any rule. 12:40 < bridge> Actually I remember rust-analyzer having some trouble with some rust. I wonder if they use an LR parser but the language as defined isn't LR 12:41 < bridge> In other news I'll be trying out C++20 coroutines today 12:53 < bridge> xD 13:31 < Konsti> i just updated client to latest and instantly regretted 13:31 < bridge> rip Konsti 13:31 < Konsti> how do i find servers that are neither kog or ddnet now? 13:32 < bridge> its all under the globe, you disable the filters on the left tab 13:32 < bridge> its all under the globe, you disable the filters on the right, at the left tab 13:32 < bridge> its all under the globe, you disable the filters on the right, at the left tab | on update, the DDRaceNetwork community filter is on by default 13:32 < bridge> Include "None" in your community filter 13:33 < Konsti> xd thanks i found it 13:33 < Konsti> unfair to check that by default 13:36 < bridge> well it IS the ddracenetwork client, so i find it okay either way to have it automatically checked tbh 13:36 < bridge> but it can be confusing as to why you dont see other servers on startup - true 13:38 < bridge> "None" is disabled by default? 13:53 < bridge> @archimede67 Thanks! ^ 13:59 < bridge> :poggers2: 14:08 < bridge> AAAAAA coroutines are soo cool 14:08 < bridge> They just need some STL support for them to be very usable, but even by hand they are not too bad, and so flexible too 14:08 < bridge> I hope the performance isn't god awful 14:12 < bridge> ugh 14:12 < bridge> should i implement short circuit for ifs etc 14:12 < bridge> sounds pain 14:14 < bridge> i see it puts and and or in a diferent way at the ast level 14:31 < bridge> you have to, it would be very meh without 14:32 < ws-client> Konsti still here? 14:32 < ChillerDragon> ah rip rat gone 14:35 < bridge> unfortunately the serde data model lacks fixed-size byte arrays: https://github.com/serde-rs/serde/issues/2120 14:46 < bridge> @archimede67 WTF 14:50 < bridge> https://paste.pr0.tips/SuA look at this 14:54 < bridge> the meat is the `gen_steps` function? 14:54 < bridge> the rest is helper code? 14:54 < bridge> does the compiler to a transform into a state machine? 15:03 < bridge> Yes, the rest is just boilerplate only because the stl doesn't come with helpers yet, easily can be made generic 15:04 < bridge> Hm, I don't think that's the case with this kind of stackless coroutine 15:04 < bridge> yea, seems like a lot of overhead right now 15:04 < bridge> is there a different option? 15:04 < bridge> I think rust/C# compile stuff like that into state machines 15:05 < bridge> Oh, I just took a look at what's generated and yeah it does seem to be a state machine 15:10 < bridge> C++23 is bringing `std::generator` the first stl template that wraps coroutines, with that you'll just have to do 15:10 < bridge> ```cpp 15:10 < bridge> std::generator gen_steps(int start, int end, int stride = 1) { 15:10 < bridge> for(int i = start; i < end; i += stride) co_yield i; 15:10 < bridge> } 15:10 < bridge> ``` 15:11 < bridge> that looks like dynamic dispatch 15:11 < bridge> is it? 15:11 < bridge> Huh, why? 15:11 < bridge> > `std::generator` 15:11 < bridge> there's nothing unique about this type 15:12 < bridge> like `std::function` is also dynamic dispatch 15:12 < bridge> There is no type erasure here though, why would there be dynamic dispatch? 15:13 < bridge> Specialized on int it'd look about exactly like what I wrote with an implementation for `std::generator::iterator` so you can use range based for loops 15:13 < bridge> how can there be no type erasure? 15:14 < bridge> the type `std::generator` mentions nothing about this particular function 15:14 < bridge> i.e. it looks like it's the same type for all generators returning `int` 15:14 < bridge> am I missing something? 15:15 < bridge> Oh, yep it's exactly the same for every generator that returns int 15:15 < bridge> but why does that imply dynamic dispatch anywhere? 15:15 < bridge> sooo, it needs to have a function pointer to advance to the next element? 15:15 < bridge> rust async can do static dispatch right? 15:15 < bridge> making it dynamic dispatch 15:15 < bridge> yes 15:16 < bridge> iirc it requires object safe 15:16 < bridge> whivh is why u need to box some stugg 15:16 < bridge> stuff 15:16 < bridge> omg mobile 15:16 < bridge> dynamic dispatch needs object safety 15:16 < bridge> ah ye 15:17 < bridge> good morning 15:17 < bridge> try cppcoro 15:17 < bridge> try edlang 15:17 < bridge> available soon 15:18 < bridge> Would that even really be considered dynamic dispatch? Which implementation of resume() gets called is completely determined at compile time 15:18 < bridge> with short circuit 15:18 < bridge> Yeah already saw it, I wanted to see what Core is capable of for now 15:19 < bridge> i feel sick but idk why 15:19 < bridge> every time a function pointer is involved, it's dynamic dispatch, no? 15:20 < bridge> for performance considerations 15:20 < bridge> I am unsure, idk if there is an official definition we can refer to, but for me picking which function to call at runtime is the characteristic of dynamic dispatch 15:21 < bridge> I see 15:21 < bridge> I wanted to say that this is as much dynamic dispatch as passing an `std::function`, even if it's a constant 15:21 < bridge> but I was wrong about there being no type erasure to begin with, generators hold a `std::coroutine_handle<>` erasing the promise type 15:25 < bridge> I wonder how well current compilers can see through this to optimize, what I wrote should be optimized out to just a for loop 15:30 < bridge> https://godbolt.org/z/ef8xj87cb atleast clang is pretty good at it 😄 15:30 < bridge> Though I did have to suggest inline to it 15:31 < bridge> my oficial def is if the llvm call is indirect or not 15:31 < bridge> gcc is godawful at it 15:31 < bridge> llvm ir 15:35 < bridge> hungry 15:35 < bridge> It has been a while since I've seen gcc fumble so badly against clang 15:35 < bridge> im at gym :owo: 15:36 < bridge> this basement looking thing has no Internet and no wifi 15:36 < bridge> u gotta go to a spot to get better Signal 15:36 < bridge> xd 15:38 < bridge> MSVC is even worse, I guess until they both improve using coroutines in anything serious is out of question 15:40 < bridge> rip 15:40 < bridge> tldr: use rust 15:43 < bridge> one day 15:43 < bridge> but 15:43 < bridge> if u wanna do async u almodt 15:43 < bridge> almost always 15:44 < bridge> wanna use erlang or elixir 15:44 < bridge> Anyway, enough messing around, it is time to learn some vulkan 15:44 < bridge> or gleam 15:44 < bridge> gleam uses BEAM too but is rust like 15:44 < bridge> i need to learn it 15:44 < bridge> I wonder what performance and the ecosystem is like for these more exotic languages 15:45 < bridge> they literally made for async 15:45 < bridge> the lang itself is a actor system 15:45 < bridge> One of the biggest charms of using C++, Rust, Java, JS, Python is that the performance characteristics are very well known 15:45 < bridge> its fault tolerant 15:45 < bridge> whatsapp uses erlang 15:46 < bridge> u should learn it 15:46 < bridge> i need to too 15:46 < bridge> It at the very least sounds cool 15:46 < bridge> my company uses erlang elixir for distributed, and to call cpu bound programs in rust 15:46 < bridge> and rust for making VMs and cpu bound stuff 15:47 < bridge> my company uses erlang elixir for distributed, and to call cpu bound programs made in rust 15:47 < bridge> also BEAM allows full introspection 15:47 < bridge> to memory, tasks etc 15:47 < bridge> I wonder if gleam has erlang ffi seeing as they run on the same vm 15:47 < bridge> and u can kill stugg 15:47 < bridge> stuff 15:47 < bridge>