00:05 < bridge> <zhn> z-team client (0.5.2) had it too xdddd
00:18 < bridge> <Ewan> finally purple
00:19 < bridge> <smlsd> Hello Devs
00:20 < bridge> <Ewan> helo Smlsd
00:25 < bridge> <smlsd> How can i create my own ddnet skin
00:25 < bridge> <Ewan> use an image editor
00:25 < bridge> <Ewan> are u familiar with photoshop/paint.net or smth like illustrator/inkscape
00:31 < bridge> <smlsd> Yes after?
00:31 < bridge> <Ewan> what
00:33 < bridge> <smlsd> Tell me how to make
00:37 < bridge> <smlsd> @mpft
00:38 < bridge> <Ewan> look at the default skin
00:38 < bridge> <Ewan> or find some template
00:38 < bridge> <Ewan> https://cdn.discordapp.com/attachments/293493549758939136/1242969865048821941/default.png?ex=664fc565&is=664e73e5&hm=d906cc69266fcdae01fd1727e17485868f5c5ae0b5b9102396f6672bd7c0b5ab&
00:38 < bridge> <Ewan> change it
00:39 < bridge> <smlsd> Ooooo
00:39 < bridge> <smlsd> Oke i understand
00:39 < bridge> <smlsd> I'm Gonna make myself
00:39 < bridge> <smlsd> Can everybody see when i send in #📬submit-skins  ?
00:39 < bridge> <Ewan> Yes
00:39 < bridge> <smlsd> Elly?
00:39 < bridge> <smlsd> Really?
00:39 < bridge> <smlsd> YEEEY
00:40 < bridge> <smlsd> I'm Gonna make myself lik
00:40 < bridge> <Ewan> it's not guaranteed lol
00:40 < bridge> <smlsd> LoL
00:40 < bridge> <Ewan> people have to vote
00:40 < bridge> <smlsd> 😭
00:40 < bridge> <Ewan> and if your skin sucks then it won't be visible to others
00:40 < bridge> <smlsd> What votes
00:40 < bridge> <smlsd> Soo hmuch likes need
00:42 < bridge> <Ewan> #📍info
00:42 < bridge> <smlsd> @mpft what blue
00:43 < bridge> <smlsd> What means i didn't understand
00:43 < bridge> <Ewan> read the messages in the channel
00:43 < bridge> <Ewan> and please stop pinging me
00:44 < bridge> <smlsd> But it's not working dude
00:44 < bridge> <smlsd> I cant see any emojis
00:44 < bridge> <smlsd> Like ✅
00:44 < bridge> <smlsd> #📬submit-skins
00:44 < bridge> <smlsd> Look
00:44 < bridge> <smlsd> Nothing have
00:46 < bridge> <smlsd> Ups sorry my bad
00:46 < bridge> <smlsd> I understand now
00:46 < bridge> <smlsd> 🙂 tu
00:46 < bridge> <smlsd> Ty
01:27 < bridge> <louis> this runs pretty well without gpu too in my experience
01:27 < bridge> <louis> but i have an 'ai power' cpu so xd
01:31 <+ws-client> <ChillerDragon> if you need a free api checkout https://trans.zillyhuhn.com/
01:43 < bridge> <Nagi01 {LAN}> mmh very interesting service
01:43 < bridge> <Nagi01 {LAN}> i think ill host one too 🙂
02:18 <+ws-client> <ChillerDragon> the translations are a bit cursed tho :D 
02:18 <+ws-client> <ChillerDragon> but ye its super simple and light weight to self host
02:29 <+ws-client> <ChillerDragon> @heinrich5991 i mean every feature I wanted is working now. But i still dont like the developer expirience. Now I have to annotate every function defintion and call with a #ifdef thingy
02:31 <+ws-client> <ChillerDragon> https://zillyhuhn.com/cs/.1716424277.png
02:32 <+ws-client> <ChillerDragon> I mean big ass functions with 10k lines are fun and then you only have to do that thing once. But I started to enjoy using more functions these days and then this would be unusabily bad to type.
02:33 <+ws-client> <ChillerDragon> there has to be a way to find a function in an shared object that is also already defined in the main executable
03:00 <+ws-client> <ChillerDragon> https://zillyhuhn.com/cs/.1716426013.png
03:00 <+ws-client> <ChillerDragon> so this works ....
03:00 <+ws-client> <ChillerDragon> but even i am scared of those macros
03:02 <+ws-client> <ChillerDragon> I could replace function names in the final binary using some hexediting step at the end of the cmake pipeline xd
03:03 <+ws-client> <ChillerDragon> that should work too. And then writing the code feels super smooht. No macros.
03:14 < bridge> <learath2> What are you even making?
03:21 < bridge> <learath2> With ifuncs anything is possible
04:30 <+ws-client> <ChillerDragon> watafak is ifuncs
04:31 <+ws-client> <ChillerDragon> i tried messing with linker scripts to rename my functions but that was no fun :( 
04:33 <+ws-client> <ChillerDragon> What i am working on is my developer expirience. I want to control tees on the server side. Like let them walk left or right. And those controls should be written in C. And i want them to get applied on compile without server reload.
04:33 <+ws-client> <ChillerDragon> it all works now but only if i have all function names twice. because fakin dlsym does not look in my shared object if they are also statically linked into the server executable already
05:59 <+ws-client> <ChillerDragon> Hmm interesting i managed to get dlsym work as i wanted it to. But outside of ddnet https://github.com/ChillerDragon/CompileTests/tree/6cb5015eeb0df128510c8c186491f6629816b800 but this one has all the features i want. Not sure how the ddnet code base is different from that. I mean ddnets cmakelists is huge who knowns which flags break rdynamic
06:37 < bridge> <0xdeen> @alex0006 Thank you!
06:41 < bridge> <Bota> 🙃
08:22 < bridge> <ryozuki> morning
08:27 < bridge> <Bota> 🚂
08:30 < bridge> <Ewan> gm
09:03 < bridge> <sedonya> morning
09:13 < bridge> <meloƞ> Morning Windows enjoyers and others
09:15 < bridge> <Bota> Im a Windown Addict :monkalaugh:
09:19 < bridge> <meloƞ> @ryozuki its my Birthday, am i now officially older than you? I forgot
09:21 < bridge> <ryozuki> ye
09:21 < bridge> <Bota> Happy Birthday! 🎂🎉🎁🎈🎊
10:20 <+ws-client> <JStar> morning may kids and others
10:20 <+ws-client> <JStar> @melon happy birthday. may kids are best, i also got older this month sadness
10:44 < bridge> <ryozuki> https://discourse.llvm.org/t/rfc-building-llvm-for-webassembly/79073
10:45 < bridge> <ryozuki> mine is on 29th
10:46 <+ws-client> <JStar> @ryozuki +rep
10:47 < bridge> <ryozuki> why u chat on irc now
10:47 < bridge> <ryozuki> btw llvm on wasm can be pog
10:47 < bridge> <ryozuki> better godbolt, rust wasm compiler, etc
10:47 <+ws-client> <JStar> i always forget to refresh my login on mobile xd
10:48 <+ws-client> <JStar> i cant access the pw from my mobile lmao
10:48 < bridge> <ryozuki> and u can compile wasm on wasm host
10:48 <+ws-client> <JStar> yeah would be funny
10:48 <+ws-client> <JStar> wasm-ception
10:56 < bridge> <AssassinTee> Good morning, does anyone know, if the ddnet client has true command line arguments? I want to start the client in the editor directly with a map, something like
10:56 < bridge> <AssassinTee> `ddnet-client --map something.map`
10:56 <+ws-client> <JStar> ./DDNet "mymap.map" should work or not?
10:58 < bridge> <AssassinTee> does it? If so thank you 👍
12:57 < bridge> <Peakies> HBD Melon senpai 🎉
15:06 < bridge> <learath2> My favourite type of website. Just `error_generic` and `not_valid` doesn't say what the generic error is nor what is not valid on the form
15:06 < bridge> <learath2> https://cdn.discordapp.com/attachments/293493549758939136/1243188398001225759/image.png?ex=665090eb&is=664f3f6b&hm=44760a3aa4c742de8029dbb91c9b5255b56facedac731d13978d2c82b4789ecb&
15:07 < bridge> <learath2> Submitting the form again with no change immediately worked
15:10 < bridge> <learath2> Which actually didn't work, but added my tax code to the db so now I can't try again as it shows me as registered but I didn't even finish registering
15:10 < bridge> <heinrich5991> thanks software
15:12 < bridge> <learath2> but atleast the website looks cool with react and implemented using a flux pattern for state consistency
15:13 < bridge> <learath2> turns out if you don't think out your logic properly, no amount of fancy patterns will save you
15:33 < bridge> <fokkonaut> thanks a lot
15:46 < bridge> <ryozuki> :gigachad:
15:48 < bridge> <learath2> I'm really enjoying the builtin testing framework in golang. It's just so easy to use, no thinking about what to use, how to configure it
15:49 < bridge> <learath2> Just put tests in a file with suffix `_test.go` and name the functions with prefix `Test`, it even has builtin coverage, so you can just add `-cover` and it'll calculate coverage too while testing
15:50 < bridge> <learath2> Also builtin support for benchmarks, and not the stupid kind, it modifies the iteration count until the result is statistically significant like `hyperfine`
15:51 < bridge> <learath2> Parallel testing is also something you don't have to think about at all, if the test can run in parallel to others you just call `t.Parallel()`
15:52 < bridge> <heinrich5991> rust also has this kind of test support 🙂
15:52 < bridge> <egyt> Is this Esperanto?
15:52 < bridge> <heinrich5991> no built-in test coverage though
15:52 <+ws-client> <ChillerDragon> @AssassinTee ye the editor can open demo files and map names when you provide it as argument. It looks at the file extension. But breaks on spaces because it could also be a console command xd
15:53 < bridge> <learath2> Rust was also quite enjoyable after C++ tbf 😄
15:53 <+ws-client> <ChillerDragon> ``./DDNet "my map.map"`` runs the console command "my" and passes "map.map" as argument.
15:53 < bridge> <ryozuki> https://en.cppreference.com/w/cpp/utility/tuple/tie
15:53 < bridge> <ryozuki> c++ tuples
15:53 < bridge> <ryozuki> xdd
15:53 < bridge> <ryozuki> they cant add syntax sugar right
15:54 < bridge> <ryozuki> its so ugly
15:56 < bridge> <learath2> Mh, tuples are the least offensive part of modern C++ to me
15:56 <+ws-client> <ChillerDragon> @Jupstar ✪ having an insecure password for chat.zillyhuhn.com is so pog! makes it such an accessible chat platform.
15:56 < bridge> <learath2> Especially not tie. Look at how they interact with `std::get` though
15:56 < bridge> <heinrich5991> sum types are annoying in C++
15:57 < bridge> <heinrich5991> and go, I assume
15:59 < bridge> <learath2> Yeah not the most enjoyable things, in go for sum types you have to forgo type safety
16:00 < bridge> <learath2> well mostly, sometimes you can get away with an interface that atleast keeps some information
16:03 < bridge> <ryozuki> xd
16:03 < bridge> <ryozuki> rust sum types are its power
16:03 < bridge> <ryozuki> ADT
16:06 < bridge> <learath2> Your best bet in C++ would be  a tagged union or a std::variant, both have huge downsides as you know
16:11 <+ws-client> <JStar> @chillerdragon true xDDD, i mean it's super secure, ngl
16:40 < bridge> <thenofis> How to reset the running time in the server code?
16:40 < bridge> <thenofis> https://cdn.discordapp.com/attachments/293493549758939136/1243211925513437225/image.png?ex=6650a6d4&is=664f5554&hm=ad2d47c99f32a819eaeaf771ebd52e457d2282977c291e1586679f7e0abde9df&
16:44 < bridge> <ryozuki> https://dl.acm.org/doi/pdf/10.1145/356580.356581
16:44 < bridge> <ryozuki> Von Neumann's First Computer Program
16:45 < bridge> <ryozuki> ```
16:45 < bridge> <ryozuki> A third precaution: The notation used in
16:45 < bridge> <ryozuki> this paper differs considerably from that
16:45 < bridge> <ryozuki> used by von Neumann, so that modern
16:45 < bridge> <ryozuki> readers can more easily understand what he
16:45 < bridge> <ryozuki> did. Where he would write, for instance,
16:45 < bridge> <ryozuki> 15) c -}- (m' -- 1)(p ~- 1) ~-~ 14 [ p ~- 2,
16:45 < bridge> <ryozuki> we will use an equivalent assembly-like
16:45 < bridge> <ryozuki> language form,
16:45 < bridge> <ryozuki> MOVEIN PIK p~- 1, BUFFER, [YPTR].
16:45 < bridge> <ryozuki> ```
16:45 < bridge> <ryozuki> xd
16:47 < bridge> <ƤØŢĀŤ0Co> Isn't it in sql db?
16:50 < bridge> <learath2> I have question, do you have a limitation on what can be passed to C? You don't seem to handle any of the complexity that comes from passing mixed class parameters
16:51 < bridge> <ryozuki> can you give a example of mixed class parameters
16:52 < bridge> <learath2> e.g. argument 1 is MEMORY class and argument 2 is INTEGER class
16:52 < bridge> <robyt3> Works for me on Windows, e.g. with `DDNet "data/maps/Sunny Side Up.map"`
16:52 < bridge> <learath2> argument 1 should end up on stack, argument 2 should end up in rdi
16:54 < bridge> <heinrich5991> interesting. so if 1 is a large struct, then the first integer parameter will be in rdi?
16:54 < bridge> <learath2> Yes, larger than 8 eightbytes to be exact for amd64 sysv
16:55 < bridge> <learath2> Floats also break this btw, those need to go to xmm0 thru xmm7
16:55 < bridge> <ryozuki> we have no floats
16:55 < bridge> <ryozuki> i guess yeah we have a limitation
16:55 < bridge> <learath2> class SSE to be more exact
16:57 < bridge> <ryozuki> some stuff is passed by allocating it on the heap
16:58 < bridge> <ryozuki> i think thats a way to avoid the complexity xd
16:58 < bridge> <learath2> This way of passing basically only works as long as everything belongs to class integer
16:58 < bridge> <ryozuki> yeah you could say the arguments that would be memory class are passed as a pointer
16:58 < bridge> <learath2> So only pointers to  aggregates
16:58 < bridge> <ryozuki> ye
16:59 < bridge> <learath2> Also be very careful with ints with bitsize larger than 64
17:00 < bridge> <ryozuki> ```rust
17:00 < bridge> <ryozuki> pub fn push_aligned(&mut self, align: usize, mut values: &[u64]) {
17:00 < bridge> <ryozuki>         assert!(align.is_power_of_two());
17:00 < bridge> <ryozuki>         assert!(align <= 16);
17:00 < bridge> <ryozuki> 
17:00 < bridge> <ryozuki>         #[cfg(target_arch = "x86_64")]
17:00 < bridge> <ryozuki>         const NUM_REGISTER_ARGS: usize = 6;
17:00 < bridge> <ryozuki>         #[cfg(target_arch = "aarch64")]
17:00 < bridge> <ryozuki>         const NUM_REGISTER_ARGS: usize = 8;
17:00 < bridge> <ryozuki> 
17:00 < bridge> <ryozuki>         if align == 16 {
17:00 < bridge> <ryozuki>             // This works because on both aarch64 and x86_64 the stack is already aligned to
17:00 < bridge> <ryozuki>             // 16 bytes when the trampoline starts pushing values.
17:00 < bridge> <ryozuki> 
17:00 < bridge> <ryozuki>             // Whenever a value spans across multiple registers, if it's in a position where it would be split between
17:00 < bridge> <ryozuki>             // registers and the stack it must be padded so that the entire value is stored within the stack.
17:00 < bridge> <ryozuki>             if self.invoke_data.len() >= NUM_REGISTER_ARGS {
17:00 < bridge> <ryozuki>                 if self.invoke_data.len() & 1 != 0 {
17:00 < bridge> <ryozuki>                     self.invoke_data.push(0);
17:00 < bridge> <ryozuki>                 }
17:00 < bridge> <ryozuki>             } else if self.invoke_data.len() + 1 >= NUM_REGISTER_ARGS {
17:00 < bridge> <ryozuki>                 self.invoke_data.push(0);
17:00 < bridge> <ryozuki>             } else {
17:00 < bridge> <ryozuki>                 let new_len = self.invoke_data.len() + values.len();
17:00 < bridge> <ryozuki>                 if new_len >= NUM_REGISTER_ARGS && new_len % 2 != 0 {
17:00 < bridge> <ryozuki>                     let chunk;
17:00 < bridge> <ryozuki>                     (chunk, values) = if values.len() >= 4 {
17:00 < bridge> <ryozuki>                         values.split_at(4)
17:00 < bridge> <ryozuki>                     } else {
17:00 < bridge> <ryozuki>                         (values, [].as_slice())
17:00 < bridge> <ryozuki>                     };
17:00 < bridge> <devinci24> you can access the time with m_Score on the player class. It's an std::optional so you can reset it with .reset() method. That may be wrong but at least it's what I did :p.
17:01 < bridge> <devinci24> It wont delete the time in the db though.
17:01 < bridge> <devinci24> 
17:01 < bridge> <devinci24> You could also just delete the times directly in the db if you don't want them ig.
17:01 < bridge> <learath2> I would honestly just ban anything larger than a register and only special case i128 if you need those
17:01 < bridge> <ryozuki> thats what we do tbh
17:01 < bridge> <ryozuki> the align 16 if branch is for u128 mainly
17:02 < bridge> <heinrich5991> a 64-byte struct is passed via registers?
17:02 < bridge> <heinrich5991> do you mean larger than 2 eightbytes?
17:03 < bridge> <learath2> Unless I'm forgetting something, it should end up in registers
17:04 < bridge> <learath2> Let me check again
17:05 < bridge> <thenofis> Is there any way to get the time of the race?
17:07 < bridge> <learath2> Oh yeah, I forgot about the cleanup step, yes it's 2
20:00 < bridge> <AssassinTee> Thank you for the explanation! I have never seen it anywhere. I plan to integrate a button in the software I am currently creating, where you can start the ddnet client as check method
20:03 < bridge> <robyt3> Depends on your shell/OS I guess, it works on Windows with paths containing spaces. This is basically what happens when you double-click a .map file in the window explorer.
20:03 < bridge> <robyt3> Depends on your shell/OS I guess, it works on Windows with paths containing spaces. This is basically what happens when you double-click a .map file in the windows explorer.
20:03 < bridge> <AssassinTee> oh, the map suffix has other meanings, this doesn't work on my windows machine xD
20:03 < bridge> <milkeeycat> chillerdragon: https://www.twitch.tv/videos/2153217989 xd
20:04 < bridge> <robyt3> DDNet client should change the default editor though, unless you changed it back ^^
20:05 < bridge> <AssassinTee> tested this, it asks me now which software I want to use
20:05 < bridge> <AssassinTee> ddnet ist the first choice
20:06 < bridge> <robyt3> yeah, that's Windows default behavior since 10, the user always needs to confirm the choice
20:06 < bridge> <AssassinTee> this has some nice implications, if you work on a map and want to start the editor quick, just link the map
20:11 < bridge> <daniok_78680> hello everyone! I wanted to clarify why it says "the server has not responded yet" on most of my servers
20:12 < bridge> <Jupstar ✪> bcs u are from ukraine and want to connec to russian servers
20:12 < bridge> <Jupstar ✪> bcs u are from ukraine and want to connect to russian servers
20:12 < bridge> <daniok_78680> thank you!
20:12 < bridge> <zhn> yo jupjopjapstar
20:12 < bridge> <Jupstar ✪> yo
20:13 < bridge> <zhn> jopjapjupstar
20:13 < bridge> <Jupstar ✪> clever
20:13 < bridge> <zhn> chillerdragon is chillentano
20:13 < bridge> <daniok_78680> Oooooooh I'm Russian
20:13 < bridge> <Jupstar ✪> ok then russia seems to block traffic to outside of russia now
20:13 < bridge> <daniok_78680> my English is bad
20:13 < bridge> <Jupstar ✪> at least for your ISP
20:13 < bridge> <zhn> which servers are inaccessible
20:14 < bridge> <zhn> ips or names
20:14 < bridge> <zhn> how are you jupstar
20:14 < bridge> <daniok_78680> all servers are from ddnet
20:15 < bridge> <Jupstar ✪> i'm fine, how are you zohan the best moderator that is 100% still moderator?
20:15 < bridge> <zhn> are you sure?
20:15 < bridge> <Jupstar ✪> even RUS servers?
20:15 < bridge> <zhn> :justatest:
20:15 < bridge> <zhn> i have an access to all ddnet servers fifo but sshhh
20:15 < bridge> <daniok_78680> Yes
20:15 < bridge> <Jupstar ✪> mhh, u can try to restart your router
20:16 < bridge> <Jupstar ✪> has it ever worked for you?
20:16 < bridge> <Jupstar ✪> or are you a new player?
20:16 < bridge> <daniok_78680> I've already done it
20:16 < bridge> <AssassinTee> you can try a vpn
20:16 < bridge> <AssassinTee> then you know if it's related to this issue
20:17 < bridge> <daniok_78680> I do not know how to download vpn on pc😅
20:17 < bridge> <AssassinTee> I don't know of you'll get "sanctioned" if you try to do that
20:18 < bridge> <daniok_78680> because of the VPN?
20:18 < bridge> <zhn> can you try to ping any server from your pc?
20:19 < bridge> <daniok_78680> And how is it?
20:19 < bridge> <zhn> press win+R and write cmd.exe
20:20 < bridge> <zhn> press enter and write "ping <ip_of_any_ddnet_server_here>" in console
20:20 < bridge> <zhn> and press enter one more time
20:20 < bridge> <AssassinTee> shouldn't also the domain work? master1.ddnet.org?
20:20 < bridge> <AssassinTee> or something like that
20:21 < bridge> <Jupstar ✪> i mean he sees the servers, so server list works
20:21 < bridge> <Jupstar ✪> maybe UDP is blocked
20:22 < bridge> <daniok_78680> issues
20:26 < bridge> <zhn> jupsti was right
21:16 < bridge> <ryozuki> @learath2 do u think its possible to do the assembly i gave u in C without inline asm?
21:19 < bridge> <learath2> The trampoline thing? Surely it should be possible
21:19 < bridge> <learath2> (With the restrictions we talked about)
21:25 < bridge> <learath2> Actually no, what I was thinking of doesn't quite work
21:27 < bridge> <ryozuki> remember it has a arbitrary num of args
21:27 < bridge> <ryozuki> i would like to see it in pure c or rust
21:37 < bridge> <learath2> I was thinking `__builtin_apply` but it's just too weak
21:49 < bridge> <learath2> Btw something like libffi already has this stuff properly implemented if you’d like to take a look there
21:50 < bridge> <ryozuki> oh
21:51 < bridge> <learath2> (in assembly ofc, that’s the only proper way to do it, I was thinking of a way to trick the compiler into generating the assembly for you, which didn’t quite pan out)
21:57 < bridge> <chairn> do the arguments are all the same size or can they be of different types ?
21:58 < bridge> <learath2> He constrained it to the situation where we only have to handle amd64 sysv abi with only INTEGER class arguments
21:59 < bridge> <learath2> Well aarch64 too but I think if you can get C or Rust code to generate correct assembly for one the other should also be free anyway
22:01 < bridge> <chairn> then just a function with first parameter as int, then varargs should work
22:03 < bridge> <learath2> Oh, you also start with just a pointer to the function, a pointer to an array of the args and the amount of args
22:04 < bridge> <ryozuki> ye
22:04 < bridge> <ryozuki> the trampoline builds the call
22:08 < bridge> <learath2> There is a very jank way to do it, a switch statement over the arg count, and each branch has a call with 0 thru MAX_ARGCOUNT arguments 😄
22:10 < bridge> <learath2> It won’t optimize properly at all I bet