08:00 < bridge> @sollybunny: was rejected upstream :/ 08:01 < bridge> hi chillerdragon 08:01 < bridge> how you doing 08:02 < bridge> Hungover 08:02 < bridge> U? 08:03 < bridge> morning 08:04 < bridge> morning hungover 08:08 < bridge> doing great 08:08 < chillerbot12> Elo @Assa I just wanted to remind you that: test if long map setup times cause client getting stuck in loading screen waiting for server response 08:08 < bridge> messing around with ddnet-insta 08:13 < bridge> !remind 40 fix ddrace teams 08:13 < chillerbot12> Helo chillerdragon I will remind you at Thu Sep 11 2025 08:53:39 GMT+0200 (Central European Summer Time) about your matter again. 08:53 < chillerbot12> Elo @chillerdragon I just wanted to remind you that: fix ddrace teams 09:32 < bridge> good bot 🙂 09:32 < bridge> :heartw: 09:47 < bridge> 20 seconds for map loading, server still responds, can't confirm that this is an issue 09:48 < bridge> I don't know if the client sends keep alives or something 10:12 < bridge> damn 10:12 < bridge> getting SMP is haaard 10:13 < bridge> Survival Multiplayer? :justatest: 10:21 < bridge> somehow the scheduler from core 1 dies after a few runs 10:21 < bridge> 20 seconds isn't very long, may add delays manually and try multiple minutes :justatest: I tested using very long map map download and couldn't reproduce the issue 10:26 < bridge> I wonder if the callback does anything else except swap in the graphic backend and updating the UI. Maybe it breaks if a single layer takes to long? 🤔 10:27 < bridge> The RenderLoading doesn't do networking I think 10:42 < bridge> okay after 100 seconds you fly out for too weak connection xD 10:46 < bridge> :brownbear: 10:57 < bridge> @chillerdragon wwhhyyy 10:58 < bridge> germans 10:58 < bridge> check ur phones in 2 mins 10:58 < bridge> im gonna hack em and play a loud sound 11:03 < bridge> you hacked my BOS Melder as well I assume 11:11 < ws-client> **** @ryozuki i saw a coffee machine being affected that is not connected to the internet <:tee_thinking:478629518358085653> 11:11 < ws-client> **** how did you hack into an airgapped device watfak 11:12 < ws-client> **** @Solly (bun bun) wat why? you mean my commit msg nitpick? 11:13 < ws-client> **** I would like to strictly enforce conventional commits for change log layout consistency and filtering. Also helps to differentiate between ddnet and ddnet-insta commits in the log. And helps reads to ignore things they dont care about. 11:13 < bridge> Haha wrong reply 11:13 < bridge> Why was it rejected upstream 11:14 < ws-client> **** shiet failed reply guess, i fell off 11:14 < ws-client> **** because heinrich 11:14 < ws-client> **** https://github.com/ddnet/ddnet/issues/9877 11:36 < bridge> Who knows how to make sure that in the game, if the player has registered /spec, then he disappears from the map while he is watching , it is in the ddnet or opengores code 11:43 < bridge> `sv_pauseable 1` 11:44 < ws-client> **** ew spec 11:45 < bridge> bro im fucking blind 11:45 < bridge> it says "char" nor "chat" 11:54 < bridge> @chillerdragon btw we are repeating work by kaizo-insta 11:54 < bridge> 11:54 < bridge> it woulda been easier to steal from them then redo the entire thing and they will have 2 of the same gamemode now so weve wasted their work 11:54 < bridge> i dont care if the one we have is comparably good 11:55 < bridge> credit is not required as they didnt create foot 11:55 < bridge> credit for who originally did maybes 11:55 < bridge> like ryo for tsmash 11:55 < bridge> also in ddnet-insta i cant /spec 11:55 < ws-client> **** ah so you just talk about wasted time hm yes 11:56 < ws-client> **** but the damage is already done we cant revert time now, can we? 11:56 < bridge> nop! 11:56 < ws-client> **** so what are you suggesting we do now? 11:56 < bridge> replying to: "i dont care if the one we have is comparably good" 11:56 < bridge> which is why . 11:56 < bridge> my msg on the github was just informing as i didnt previously know kaizo had a foot 11:56 < ws-client> **** well we all knew xd 11:57 < bridge> i didnt! 11:57 < ws-client> **** u must be new here 11:57 < bridge> i am! 11:57 < ws-client> **** ok so what now 11:57 < bridge> continue as previosuly 11:57 < ws-client> **** ok 11:57 < bridge> annoiy kaizo person a bit 11:57 < ws-client> **** not sure if it annoys him 11:57 < ws-client> **** i think he might have hard forked already 11:58 < bridge> oh true 11:58 < bridge> did pull from ddnet instead of ddnet-insta 11:58 < ws-client> **** if he cares he can let me know and i am ready to bend my self in any direction to make his fork life easier 11:58 < bridge> 👻 11:58 < bridge> spec broken in insta 11:58 < ws-client> **** whats broken about it 11:58 < bridge> 🥺 11:59 < bridge> command not found 11:59 < ws-client> **** in which mode? 11:59 < bridge> ddrace 11:59 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415637822412689519/image.png?ex=68c3eef6&is=68c29d76&hm=e879c9bc49eaef312073ad03c31b1004e0639d3e2ee9f2184f71b5e1eb872e7a& 11:59 < ws-client> **** seems like a bug indeed 11:59 < ws-client> **** could you open an issue? 11:59 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415637918231433289/image.png?ex=68c3ef0c&is=68c29d8c&hm=bf79587caa8edad05dc7988d09a6c90b4817407f2b27dc90b755c45cfd4f77a5& 11:59 < bridge> urmm 11:59 < bridge> did i ef something 11:59 < bridge> or is that just cuz i have test cmds on 12:00 < ws-client> **** i assume `aGameType[16]` is your problem 12:00 < bridge> oh i meant the fact it is test 12:00 < bridge> not the being cutoff 12:00 < ws-client> **** a 12:00 < bridge> i think everyone deals with that 12:00 < ws-client> **** do you have testings cmds 12:01 < bridge> yis 12:01 < bridge> with another server 12:01 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415638388328890409/image.png?ex=68c3ef7d&is=68c29dfd&hm=679cb87fa1eac855b54349391ab38ad09779b3fc7e6da2f09a58b4867a692e3f& 12:01 < bridge> and /spec works 12:05 < bridge> @chillerdragon free merge right 12:05 < bridge> https://github.com/ddnet/ddnet/pull/10885 12:05 < bridge> once ci clears 12:05 < bridge> these are the important ones tho 12:05 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415639468680937566/image.png?ex=68c3f07e&is=68c29efe&hm=38d15e91a3b3a724743edb4301014fe009a29b6540281f6130882c8a4a427436& 12:05 < bridge> so idk 12:11 < bridge> What parameter is needed when starting the DDNet Server to load the server.cfg config In the directory with the DDNet server itself 12:12 < bridge> -f or exec 12:12 < bridge> I also recommend deleting system wide configs to avoid confusion 12:13 < bridge> @sollybunny: spec availability is not related to test or not. This is a bug introduced by PvP modes. Go send issue 12:13 < bridge> i did 12:14 < bridge> Thank 12:14 < bridge> s 12:14 < bridge> native btw 12:14 < bridge> idk what this means, me being native shouldnt imply that i have perfect spelling grammer and punctuation 12:14 < bridge> I think I found the worst performing map :pepeW: 12:14 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415641758293033022/screenshot_2025-09-11_12-14-32.png?ex=68c3f2a0&is=68c2a120&hm=6143e16abad148c27ad00ff019298ea76668dcb5db83240a425402722aed31c3& 12:15 < bridge> snap limit reached 12:15 < bridge> it just means all my life ive heard and spoke english how other people have spoken english 12:22 < bridge> @kebscs this is caused by dragger prediction, not even by the laser head 12:23 < bridge> it directly points me to `int CCollision::IntersectNoLaser(vec2 Pos0, vec2 Pos1, vec2 *pOutCollision, vec2 *pOutBeforeCollision)` :pepeW: 12:25 < bridge> i can check why but i didnt add dragger prediction 12:26 < bridge> i can check why, but i didnt add dragger prediction 12:28 < bridge> collision code is shared and prediction looks same as on server 12:28 < bridge> why is clamp an offender here oO 12:28 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415645246175576064/Unbenannt.png?ex=68c3f5e0&is=68c2a460&hm=68247ffe0c98b87312448d26db0500b1fbb8ca9b546e208750febd8d21285063& 12:28 < bridge> so idk without possibly changing server physics 12:30 < bridge> well if this is the only thing in a tight loop ofc it is 12:32 < bridge> you may be able to take out the clamp and have special behaviour for if the sides are oob 12:37 < ws-client> **** std bloat hitting again 12:37 < ws-client> **** i rather have an std in my pants than in my c++ 12:39 < bridge> the previous clamp we had was an std::clamp wrapper 12:39 < bridge> given std::.. is just wrappers for what we already have it only effects compile time not performance 12:39 < bridge> and insentivies using expensive constructs but eh 12:43 < bridge> it's actually a good idea to move part of this out of this loop 12:43 < bridge> multiplying a fixed NX/NY with a might be much faster 13:06 < bridge> I guess if I fix this, I am changing physics again ._. 13:07 < bridge> if u just take the clamp out without accounting for this you do change fixes 13:07 < bridge> if u just take the clamp out without accounting for this you do change physics 13:07 < bridge> stop 13:07 < bridge> but u dont have to 13:07 < bridge> the extra behaviour for oob is cold path anyway 13:08 < bridge> why gh trolling 13:08 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415655243731046491/image.png?ex=68c3ff2f&is=68c2adaf&hm=35b8b0eb60a73d191b0ad81ad45c1cea441316f439b6e698dc50d920cf11050f& 13:14 < bridge> technically this would need a proper grid traversal algorithm and not this bullshit, but changing this may cause other physics, because tiles may not be skipped that previously have been 13:14 < bridge> yeye I stop Teero don't worry 13:15 < bridge> you can even be lazy about it, if oob do old code, if not oob do new code 13:16 < bridge> yea if we actually decide to change the shit line intersection and movebox i'd like to be part of it xD 13:16 < bridge> oh yeah teero upstream tas code B) 13:16 < bridge> those are the main bottlenecks for the physics (not including multi-tee) 13:16 < bridge> :feelsbadman: 13:17 < bridge> maybe i should break physics as much as possible to achieve max performance without any notable visual/gameplay changes 13:17 < bridge> that would be nice 13:18 < bridge> do everything in fixed so theres no weird float shenangns (apart from prediction) 13:18 < bridge> okay, what about quantization? 13:19 < bridge> what does thine mean 13:19 < bridge> velocity and position can only be units of 1/32 already right? 13:19 < bridge> velocity and position can only be units of 1/32 of a block already right? 13:19 < bridge> https://github.com/ddnet/ddnet/blob/c3b10b56df901707034db6f42739bc63a21134d2/src/game/gamecore.cpp#L695-L700 13:20 < bridge> this is hard to do while keeping corner edges 13:20 < bridge> if you ignore than 1 thing then you can optimize fully 13:21 < bridge> yea they're not needed, fuck them. there will be more major changes lmao xd 13:21 < bridge> well 13:21 < bridge> why couldnt you keep them 13:21 < bridge> you could actually do them on purpose 13:21 < bridge> also it's mostly used by bots/tas anyways so its just for cheaters imo 13:21 < bridge> you can do fast block collision for seeing if you are inside them at all then do a slow manual check 13:21 < bridge> actually yea wait 13:21 < bridge> just say that the tiniest 1% of the corner is allowed to pass through 13:21 < bridge> (i dont think its that slow) 13:22 < bridge> (corner edges are going slightly into freeze?) 13:22 < bridge> yea just return if you find 2 freeze px xd 13:23 < bridge> @totar do i keep the global distance field? XD its kinda very expensive rn 13:23 < bridge> for what 13:23 < bridge> for line intersection 13:23 < bridge> right 13:23 < bridge> the precompute 13:23 < bridge> Teero care this isn't safe space™! 13:23 < bridge> for hypothetical upstream physics or in your library 13:24 < bridge> also what melon said 13:24 < bridge> but I don't think your library is supposed to be secret? 13:24 < bridge> I don't think ddnet should calculate and store a distance field lol, not worth 13:24 < bridge> oh is this a DDNet physics in c99 talk? I didn't pay attention 13:25 < bridge> making the line intersection formula correct is already like 10 times fastert 13:25 < bridge> :justatest: 13:25 < bridge> making the line intersection formula correct is already like 10 times faster 13:25 < bridge> making the line intersection formula correct is already like 10 times faster, if you're accepting physics changes 13:25 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415659455596724304/image.png?ex=68c4031b&is=68c2b19b&hm=68ae9e81da9bcb5f49c59e300b562c324542a11775f3205f2eebff36c9340d27& 13:25 < bridge> AI? 13:25 < bridge> sus 13:25 < bridge> wheres libteero 13:25 < bridge> Botters hijacked \#developer again 13:26 < bridge> libteero is sharp and difficult to handle, be careful 13:26 < bridge> libteero isnt public 13:26 < bridge> idk I don't control that 13:27 < bridge> what is this lmao 13:27 < bridge> if it was me I would make a much less optimized library for general use 13:27 < bridge> your very official tiktok content 13:27 < bridge> @totar but adding back the clamps and max velocity and moving pickups would be bad 13:27 < bridge> but if i can just mess up the ordering 13:27 < bridge> then i dont car 13:27 < bridge> then i dont care 13:28 < bridge> you can keep your library cutting edge, it's ok 13:28 < bridge> I didn't say YOU should do it 13:28 < bridge> i wanna tho 13:28 < bridge> still not at 10mil tbh xD 13:28 < bridge> 9.89 or smth 13:28 < bridge> bot clients will need the best performance possible, a noble cause 13:28 < bridge> xd 13:29 < bridge> dont worry no one cares about tas runs anymore 13:29 < bridge> its killing the game, but their ability to make good tasses doesnt effect how fast it kills the game 13:29 < bridge> also if we break the physics to make them more optimized then you need 2 versions of prediction for all vanilla and all the old ddnet servers 13:29 < bridge> the music XDDDDDDDDDD 13:30 < bridge> just update everything to ddnet-insta 13:30 < bridge> #UpstreamDDNet-insta 13:30 < bridge> (I have no idea what's in it) 13:30 < bridge> (everything i think) 13:30 < bridge> all the vanilla gamemodes and the common mod gamemodes 13:31 < bridge> i added tsmash 13:31 < bridge> is the code bad tho 13:31 < bridge> in my honest opinion 13:31 < bridge> a bit 13:31 < bridge> theres some weird inheritance stuff going on 13:32 < bridge> I'm not opposed to upstreaming ddnet-insta in theory and we completely reverse #7777, but the issue is that then you need to PR against ddnet instead of ddnet-insta which is probably much harder 13:32 < bridge> https://github.com/ddnet/ddnet/issues/7777 13:32 < bridge> fuck I didn't mean to do that 13:32 < bridge> do what? 13:32 < bridge> huh? 13:32 < bridge> huh 13:32 < bridge> mustve been the wind 13:33 < bridge> adding stuff to ddnet-insta is much easier than ddnet 13:33 < bridge> probloom is that you cant do split maintaining because vulnrabilities exist and i dont think other maintainers would be happy handing over ddnet-insta chunk just to chiller 13:34 < bridge> get their grimey hands all over the vanilla they threw away 13:34 < bridge> i dont think having ddnet-insta seperate is that bad 13:34 < bridge> if only the server and client physics were factored into a single library, it would be so good for moddability and maintability. utopia for all 13:34 < bridge> ik 13:34 < bridge> thats what ddnet-rs does right 13:34 < bridge> it's even better 13:34 < bridge> the physics is a module that the server sends to the client 13:34 < bridge> excoose me 13:34 < bridge> isnt that 13:34 < bridge> bad 13:34 < bridge> WASM 13:34 < bridge> o 13:34 < bridge> nuh uh 13:35 < bridge> but one step at a time 13:35 < bridge> we don't need that 13:36 < bridge> (also wasm is slow, so we can pretend this is important) 13:36 < bridge> (confirmation bias) 13:36 < bridge> man 13:37 < bridge> fzcj i need to complete skin rendering for tas tool finally so i can have my own client to test the physics directly xd 13:37 < bridge> ^ hacker 13:37 < bridge> [ban](https://cdn.discordapp.com/emojis/892423037046128702.webp?size=48&name=ban) 13:37 < bridge> since if i am not doing proper physics anyways it will be hard to test if things work 13:38 < bridge> unless i use the tool i already have in my cleint 13:38 < bridge> but thats annoying 13:38 < bridge> but you don't need skin rendering, manually create a circle, and some ovals for feat/eyes. 13:38 < bridge> cuz rendering is only half complete there 13:38 < bridge> good enough 13:38 < bridge> ok 13:38 < bridge> doing right now 13:39 < bridge> build failed ): 13:39 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415663151592116266/image.png?ex=68c4068d&is=68c2b50d&hm=2d6d694e38d453bb6a457058a9692536926bad7424a4d9d157e4af9e5fdaf299& 13:40 < bridge> omg you have that turned on 13:40 < bridge> it just goes into a label i ignore 13:40 < bridge> fair 13:40 < bridge> [cat_uwu](https://cdn.discordapp.com/emojis/1230626195649990696.webp?size=48&name=cat_uwu) 13:40 < bridge> oh yeah i need to do 13:40 < bridge> tater stuff 13:40 < bridge> oh yeah i need to do tater 13:40 < bridge> :justatest: 13:40 < bridge> so u can release 13:41 < bridge> yes 13:41 < bridge> hehe 13:41 < bridge> wait 13:41 < bridge> thats so funny 13:41 < bridge> you edited that message 13:41 < bridge> to make it more funny 13:41 < bridge> what is ur time 13:41 < bridge> we don't talk about it 13:41 < bridge> no 13:41 < bridge> as in what time is it 13:41 < bridge> refer 13:41 < bridge> er 13:42 < bridge> 742 am? 13:42 < bridge> 0742 am? 13:42 < bridge> (everyone lives in new york) 13:42 < bridge> ```cpp 13:42 < bridge> constexpr std::string VersionNumberToString(int Version) 13:42 < bridge> { 13:42 < bridge> return std::format("{}.{}.{}", Version / 1000, (Version / 100) % 10, Version % 10); 13:42 < bridge> } 13:42 < bridge> 13:42 < bridge> constexpr std::pair[] DOMAINS = { 13:42 < bridge> {"DDNet", VersionNumberToString(DDNET_VERSION_NUMBER)}, 13:42 < bridge> {"TClient", TCLIENT_VERSION}, 13:42 < bridge> }; 13:42 < bridge> ``` 13:42 < bridge> no constexpr std::format 13:43 < bridge> that is wack 13:43 < bridge> lol 13:43 < bridge> is cpp version not high enough 13:43 < bridge> its not constexpr ever 13:43 < bridge> https://en.cppreference.com/w/cpp/utility/format/format.html 13:44 < bridge> are there any contexpr string manipulation things 13:44 < bridge> idk 13:44 < bridge> are there any constexpr string manipulation things 13:44 < bridge> i think i might be overengineering 13:44 < bridge> checking a file for a string 13:44 < bridge> that's scaring me 13:45 < bridge> ```cpp 13:45 < bridge> DDNet 19.5.0 13:45 < bridge> TClient 10.4.0 13:45 < bridge> ``` 13:45 < bridge> am looking for ^ 13:45 < bridge> oh are you fixing the cmake version mess 13:45 < bridge> mm no 13:45 < bridge> ok 13:45 < bridge> im trying to make it so if u update only DDNet.exe it tells u u are an idiot 13:45 < bridge> and making it so you could theoretically upstream this 13:46 < bridge> I regret bumping tclient major version so many times, if I keep doing it then it will catch up to ddnet version. Which would be confusing unless we fly past it and keep bumping. 13:47 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415664954551566347/image.png?ex=68c4083a&is=68c2b6ba&hm=84874be8a6d5fb8acc3b55b6799f4a76bd52134f7b216d48bba600cd8968da2b& 13:47 < bridge> why do u major version bump? 13:47 < bridge> idk, I was doing it for major features, but I never expected to reach 10 13:47 < bridge> we could keep up with ddnet versions 13:47 < bridge> and just use patch 13:47 < bridge> since ddnet rarely uses it 13:47 < bridge> and we rarely patch 13:47 < bridge> it make sense when the major version doubled the number of custom features 13:47 < bridge> it made sense when the major version doubled the number of custom features 13:48 < bridge> idk what you mean 13:48 < bridge> ddnet 19.5.0 = tclient 19.5.0 13:48 < bridge> eh 13:48 < bridge> that's how it was for a while 13:48 < bridge> its a bit of a pain due to nightly version bump 13:49 < bridge> it's good enuf, we're probably not going to reach 19 13:49 < bridge> we would have to do the deen make a branch for the version 13:49 < bridge> if we get close then we can just change the format so it doesn't look so similar 13:49 < bridge> 0.4.10 13:49 < bridge> reverse semvar 13:49 < bridge> do lik rust software 13:49 < bridge> 0.0.2 13:49 < bridge> 0.0.56 13:49 < bridge> lol 13:49 < bridge> xd 13:50 < bridge> https://0ver.org/ 13:50 < bridge> In short, software versioning best practice is like the modern list/array: 0-based. 13:50 < bridge> sometimes I look at browsers and think "wow their version number is so high they must be irresponsible with it" but actually they're just 20 years old and millions of features 13:51 < bridge> its just a number 13:51 < bridge> yeah 13:51 < bridge> (:justatest: ) 13:52 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1415666209696907404/image.png?ex=68c40966&is=68c2b7e6&hm=5719a26766ba05e935ce8566f4c5d75e31e4babbb9aa1e6e84c837146fad3ce1& 13:52 < bridge> https://tenor.com/view/caught-in-4k-you-exposed-gif-21379673 13:52 < bridge> https://tenor.com/view/rip-bozo-gif-22294771 13:52 < bridge> I'm in another project that somehow does 4 decimal semvar, 1.4.1.3 13:52 < bridge> thats common in java world 13:52 < bridge> yet Linux is version 6.16 13:52 < bridge> and c# 13:52 < bridge> ok I guess that explains it 13:53 < bridge> btw terraria is like that 13:53 < bridge> its c# 13:53 < bridge> I've seen x.x.x.x versions in a few backends before 13:53 < bridge> 13:53 < bridge> Major, minor, Bugfix, hotfix 13:53 < bridge> yeah it's C# 13:53 < bridge> i think last is revision 13:53 < bridge> Oh? 13:53 < bridge> we use it like how melon said 13:53 < bridge> the besty obiously is date based versioning 13:53 < bridge> 4th is for hotfix 13:53 < bridge> imho the one conveying most info 13:53 < bridge> vs 2069 13:54 < bridge> just add -beta or -alpha if ur not prod rdy 13:54 < bridge> do minecraft and have all 3 13:54 < bridge> java has snapshots too 13:54 < bridge> The "I give up" versioning scheme: 13:54 < bridge> just one number that gets incremented with every release 13:55 < bridge> I think it's good to have an easy way of knowing if the new version break the API 13:55 < bridge> if there is an API yes 13:56 < bridge> it's very easy to version your API, you just follow the rules 13:56 < bridge> I think it's good to have an easy way of knowing if the new version breaks the API 13:56 < bridge> Clearly DDNet should only increment major version when the protocol breaks backwards compat 13:56 < bridge> we still in 1.x 13:56 < bridge> versioning api depends heavily on the language package manager and rules it uses to select compaatible packages 13:57 < bridge> semantic versioning basically 13:57 < bridge> Clearly DDNet should only increment major version when the protocol breaks backwards compat 13:57 < bridge> we still in 1.x (or maybe a couple versions more idk my history) 13:57 < bridge> the 1. is implied. Actual version is secretly 1.19.4 13:57 < bridge> It's like floating point! 13:58 < bridge> Or maybe it's 0.19.4? 13:58 < bridge> And 0. Is implied 14:00 < bridge> how would breaking the ddnet protocol even work. If the server and client support 0.6 then they can always connect. but I guess the extensions would be disabled? 14:01 < bridge> We remove vanilla support on DDRace servers 14:01 < bridge> :gigachad: 14:01 < bridge> well that would certainly be breakings 14:01 < bridge> well that would certainly be breaking 14:02 < bridge> Btw it seems like making map settings only writable during a mao load is complicated 14:03 < bridge> I looked at the code a bit and there isn't an obvious solution 14:03 < bridge> honestly I'm surprised how easily jupstar did the protocol translation for ddnet-rs. We could probably redesign the entire protocol from scratch and still have a working translation for connecting to old servers. 14:04 < bridge> What about breaking config files? Like what if a command was removed? 14:05 < bridge> I think it saves unknown commands now 14:05 < bridge> so it doesn't break 14:05 < bridge> idk tho 14:05 < bridge> haven't tested it 14:05 < bridge> I meant breaking in the sense that those configs would no longer work 14:06 < bridge>