07:51 < bridge> @Robyt3 you already did so many prs on tw (which is dead :c) but hey, you make ddnet a better place aswell :) 07:51 < bridge> Thanks for all the fixes and features 08:34 < bridge> sleeping a night over the account discussion and have to say i still kinda dislike the solution in total. 08:34 < bridge> 08:34 < bridge> 1. the account lists that would need to be refreshed all the time 08:34 < bridge> 2. the advanced complexicity of how runs are considered authed 08:34 < bridge> 3. using registry to play hide and seek 08:34 < bridge> 4. trusting only the email and that alone. 08:34 < bridge> 08:34 < bridge> i'd say a fully centralized approach can do anythign you want while still being so complicated. 08:34 < bridge> 1. a session can have a key pair. so the game server can ask the auth server if the session key is connected to an account and can cache itself how long it wants to trust the session in case the auth server is down.. else it can always ask it. 08:34 < bridge> 2. if auth server is down u either once connected to the game server and it remembers you. It could still use the session id to verify if it was related to an account. I generally dunno if this can't get messy 08:34 < bridge> 3. yeah dunno, just don't save the key on the computer. problems solved 08:34 < bridge> 4. we could make steam authing and other methods a higher priority. I dunno how a forced strong password is worse than sending an email that probably also has a bad password xdd 08:34 < bridge> 08:34 < bridge> 08:34 < bridge> I feel like you want the advantages of decentralized account system. But you don't have them 08:34 < bridge> 08:34 < bridge> Teamspeak 3 has a one time key, and this is why the teamspeak 3 approach works. If we'd do that aswell it would eliminate especially point 1. & 2. which is a huge mess IMHO 08:35 < bridge> sleeping a night over the account discussion and have to say i still kinda dislike the solution in total. 08:35 < bridge> 08:35 < bridge> 1. the account lists that would need to be refreshed (and downloaded) all the time 08:35 < bridge> 2. the advanced complexicity of how runs are considered authed 08:35 < bridge> 3. using registry to play hide and seek 08:35 < bridge> 4. trusting only the email and that alone. 08:35 < bridge> 08:35 < bridge> i'd say a fully centralized approach can do anythign you want while still not being so complicated. 08:35 < bridge> 1. a session can have a key pair. so the game server can ask the auth server if the session key is connected to an account and can cache itself how long it wants to trust the session in case the auth server is down.. else it can always ask it. 08:35 < bridge> 2. if auth server is down u either once connected to the game server and it remembers you. It could still use the session id to verify if it was related to an account. I generally dunno if this can't get messy 08:35 < bridge> 3. yeah dunno, just don't save the key on the computer. problems solved 08:35 < bridge> 4. we could make steam authing and other methods a higher priority. I dunno how a forced strong password is worse than sending an email that probably also has a bad password xdd 08:35 < bridge> 08:35 < bridge> 08:35 < bridge> I feel like you want the advantages of decentralized account system. But you don't have them 08:35 < bridge> 08:35 < bridge> Teamspeak 3 has a one time key, and this is why the teamspeak 3 approach works. If we'd do that aswell it would eliminate especially point 1. & 2. which is a huge mess IMHO 08:35 < bridge> Let's play with pu$$y here🔞💦 08:35 < bridge> https://clck.ru/33rRoo 08:36 < bridge> @Discord Mod ^ 08:36 < bridge> fast, click it! 08:36 < bridge> for science 08:36 < bridge> research 08:37 < bridge> :gigachad: 08:38 < bridge> e.g. lets say what happens if u join a kog server. the auth server is down 08:38 < bridge> u finish a map. 08:38 < bridge> You change the keys 08:38 < bridge> KoG can't auth your finish 08:38 < bridge> player complains the rank is missing. 08:38 < bridge> The stupid worst case that might never happen. 08:38 < bridge> 08:38 < bridge> The player won't understand why it cannot work. 08:38 < bridge> 08:38 < bridge> Or creating an account while the auth server is down.. kinda sounds scary too 09:10 < bridge> > Or creating an account while the auth server is down.. kinda sounds scary too 09:10 < bridge> 09:10 < bridge> Then don't allow account creation in that case? Could be as simple as pinging the server and or providing an api to check if it is running. 09:11 < bridge> Adding a messagebox to the client that tells the user than the account server is currently offline makes sense too 09:11 < bridge> Adding a messagebox to the client that tells the user that the account server is currently offline and that they can't make accounts at the moment makes sense too 09:13 < bridge> Yeah obviously. But this was something Heinrich didn't seem to want 09:13 < bridge> Probably BCS it's again less decentralized 09:13 < bridge> So understandable 09:16 < bridge> I'd suggest using an already existing protocol like OAuth2, so you can for example use Keycloak and federate User Accounts between multiple instances 09:52 < bridge> Let's play with pu$$y here🔞💦 09:52 < bridge> https://clck.ru/33rRoo 09:52 < bridge> Zwelf\: ban ^ 09:53 < bridge> he can still see it on matrix if i deleted on discord right? 09:53 < bridge> no 09:53 < bridge> its double sided bridge 09:53 < bridge> but it says that u deleted the message 09:53 < bridge> User was "iriska" 09:54 < bridge> done 09:54 < bridge> Then why do old edited messages stay 09:54 < bridge> It was a bot 09:54 < bridge> Someone invited them 09:54 < bridge> Ban the inviter too 09:54 < bridge> he knows the name 09:55 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1088747812168147004/image.png 09:55 < bridge> Also can you take away permissions to invite bots? 09:55 < bridge> Or is that Discord being trash again 09:55 < bridge> they arent bot accounts i think 09:55 < bridge> all users on matrix appear as bot here 09:55 < bridge> they are normal matrix users 09:56 < bridge> just like me 09:56 < bridge> Oh so they did post it on matrix 09:56 < bridge> its the bridge that flags it as bot 09:56 < bridge> (or better the bot itself is the bridge ^^) 10:49 < bridge> the bot does nothing but generate webhooks and read messages 10:49 < bridge> the webhooks are what show aa ‘bot’ 10:49 < bridge> the webhooks are what show as ‘bot’ 10:49 < bridge> the bot does nothing but generate webhooks and read messages to send in matrix using a similar mechanism 10:52 < bridge> we really need them for all channels 10:52 < bridge> matrix is so cool 10:55 < bridge> yeah 11:16 < bridge> *DefaultO editing their message 50+ times intensifies :mhmokay:* 12:28 < bridge> The finish would never be lost though, finishes would belong to keys until the account is resolved. 12:40 < bridge> yeah thats pretty cool, but my point is not that its not cool 12:40 < bridge> I just wanted to say it won't be lost. That's all 13:43 < bridge> @Ryozuki when a client is 0.7 client it will not send the skin json stuff at all? 13:44 < bridge> some ppl are still missing skin data in server info, but its only individuals 13:44 < bridge> it should send 0.7 skins 13:44 < bridge> ok 13:44 < bridge> so they are simply not parsed as default 13:44 < bridge> good to know 13:45 < bridge> but good way to know which nobos play 0.7 xdd 14:53 < bridge> what's your ingame name 14:53 < bridge> He hides it from everyone, mr anonymous 15:32 < bridge> i dont hide it 15:33 < bridge> i mostly played as nameless tee. but actually its keks 15:33 < bridge> or Glückskeks 15:33 < bridge> or Detektiv Keks 15:33 < bridge> i simply dislike to play in big groups esp. t0 15:34 < bridge> and playing as nameless tee gives some degree of annonymity 15:35 < bridge> I played with a guy called Butterkeks and thought it was u :justatest: he told me he wasn't 15:35 < bridge> Unless you lied to me :D 15:35 < bridge> yeah thats not me 15:35 < bridge> i usually have my clan tag in 15:36 < bridge> ΛLIVΣ 15:40 < bridge> u are a 0.7 enjoyer 15:40 < bridge> 15:41 < bridge> But let me clearly tell you something. 15:41 < bridge> We'd need to fix the map format. We'd need to improve network security etc. 15:41 < bridge> And the 0.7 client is an annoying useless problem. Sadly some devs are so obsessed to help the 5 users with their 0.7 instead of making ddnet better for all other 99% of the players, like every other game in extistence would do. 15:41 < bridge> 15:41 < bridge> So yes, please stop using it you are part of the problem with that client 15:41 < bridge> same applies to old ddnet clients btw 15:42 < bridge> Bro I'll be happy if you make your servers not accessible with 0.7 clients, then I'll make my own ddrace community like in 2019 and it'll be great 15:42 < bridge> Pls do it faster 15:42 < bridge> mh ok gl 16:35 < bridge> edlang incoming 16:35 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1088848681266073610/image.png 16:37 < bridge> u meant rused-lang 16:37 < bridge> kek 16:37 < bridge> i have 0 originality 16:37 < bridge> imagine inventing syntax 16:52 < bridge> ``` 16:52 < bridge> input: 16:52 < bridge> 16:52 < bridge> fn add(a: i64, b: i64) -> i64 { 16:52 < bridge> return a + b; 16:52 < bridge> } 16:52 < bridge> 16:52 < bridge> fn main() -> i64 { 16:52 < bridge> let x = 2 + 3; 16:52 < bridge> let y = add(x, 4); 16:52 < bridge> return y * 2; 16:52 < bridge> } 16:52 < bridge> 16:52 < bridge> output: 16:52 < bridge> ; ModuleID = 'source' 16:52 < bridge> source_filename = "source" 16:52 < bridge> 16:52 < bridge> define i64 @add(i64 %0, i64 %1) { 16:53 < bridge> entry: 16:53 < bridge> %add = add i64 %0, %1 16:53 < bridge> ret i64 %add 16:53 < bridge> } 16:53 < bridge> 16:53 < bridge> define i64 @main() { 16:53 < bridge> entry: 16:53 < bridge> %add_call = call i64 @add(i64 5, i64 4) 16:53 < bridge> %mul = mul i64 %add_call, 2 16:53 < bridge> ret i64 %mul 16:53 < bridge> } 16:53 < bridge> ``` 16:53 < bridge> function calls pog 16:54 < bridge> epic 16:56 < bridge> tbh its not rly good rn xd 16:56 < bridge> i probs need some type inference 16:56 < bridge> for constants 16:56 < bridge> rn it assumets its i64 16:56 < bridge> assumes 16:57 < bridge> fix all rust annoyances 16:57 < bridge> i would need to know how to make rust itself first 16:57 < bridge> xD 16:57 < bridge> and thats impossible for me 16:57 < bridge> xd 16:58 < bridge> https://gist.github.com/edg-l/06a442118a996bae5218b24d134173d3 16:58 < bridge> this is the grammar rn 16:59 < bridge> r u professionally working on compiler now btw or what 16:59 < bridge> ye 16:59 < bridge> cool 17:00 < bridge> who'd have thought its u doin it and not learath 17:00 < bridge> well 17:00 < bridge> learath is in the endless loop of university 17:00 < bridge> he just needs to go and work 17:00 < bridge> xD 17:01 < bridge> things get more complex when i get to if else 17:01 < bridge> u need to do some control flow analisys 17:01 < bridge> and build phi nodes 17:02 < bridge> https://mapping-high-level-constructs-to-llvm-ir.readthedocs.io/en/latest/control-structures/ssa-phi.html 18:08 < bridge> ENVÍEME EL NOMBRE DE SU PAÍS ESTAMOS BENDICIENDO A ALGUIEN ANTES DE LA MEDIANOCHE $300 $500 - $800 Y $1200 18:08 < bridge> 👇👇👇👇 18:08 < bridge> https://t.me/+aDWFgb_tv_NmOTVk 18:56 < bridge> Why were you reinventing rust btw? Just for the lulz? 19:10 < bridge> im not reinventing it 19:10 < bridge> im learning more llvm ir 19:10 < bridge> making a language 19:10 < bridge> idk 19:10 < bridge> xD 19:10 < bridge> i wanted maybe to do a teeworlds oriented language, whathever that means 19:10 < bridge> but i have no idea how it would be "teeworlds oriented" 19:11 < bridge> anyways expect nothing useful out of this 20:09 < bridge> matricks was working on a non turing complete language for teeworlds so servers could push game code to clients 20:25 < bridge> <судный день.> the assembler 20:27 < bridge> matricks inventing shellcode 20:27 < bridge> omg 20:28 < bridge> <<~{Barsik}~>> edlang = rust 1.67? 20:28 < bridge> its a safe lang too cuz there are no allocations! kek 20:29 < bridge> <<~{Barsik}~>> lulz 20:29 < bridge> <<~{Barsik}~>> no code is the best code 20:29 < bridge> <<~{Barsik}~>> not that "no code" zoomer thingie but 20:29 < bridge> <<~{Barsik}~>> okey nvm 20:33 < bridge> wait, does this mean we can have modded servers be real? 20:34 < bridge> they are already real lol 20:34 < bridge> fair 20:34 < bridge> ddnet is a modded server 20:34 < bridge> no i mean- why the turing thing 20:34 < bridge> then 20:35 < bridge> non turing complete so no halting problem 20:35 < bridge> ah 20:35 < bridge> the server could then tell the client a moditication 20:35 < bridge> but this cant be good enough anyway 20:37 < bridge> wasm ez