02:36 < ws-client> @learath2 maps as json is called MapDir ``cargo install twmap-tools`` ``twmap-edit --mapdir dm1.map out_folder_name`` 03:10 < bridge> nameplates pr has 60 review comments :justatest: 03:26 < bridge> sounds fun 03:29 < bridge> #9284 03:29 < bridge> https://github.com/ddnet/ddnet/pull/9284 03:57 < bridge> can anyone answer while i'm making the changes. timezone and stuff 03:57 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311888739638972457/image.png?ex=674a7f19&is=67492d99&hm=d6c67f35283788c44e1217bd72e8acdd5c0856ee9571a2a4b41ad1a7d1b9dfb7& 03:57 < bridge> :greenthing: am i encouraged to just use classes instead 03:58 < bridge> i don’t understand why structs shouldn’t be used 03:58 < bridge> idiomatically even if access specifiers are different by default there is definitely a contextual difference between structs and classes imo 03:59 < bridge> structs make me think it’s fairly portable data, maybe a POD type 03:59 < bridge> class makes me think generics, methods, private machinations 03:59 < bridge> i don't either, but i tend to follow reviewers. but i'm more confused about whether i need classes or he wants me to factor out the variables into the parent class directly 04:00 < bridge> well you would need to prefix it with S if you were to keep it a struct 04:00 < bridge> but yeah 04:00 < bridge> idk the context that much 04:00 < bridge> :pepeW: maybe i'll just wait for robyt3 04:01 < bridge> gotta figure out the cursor position thing. 04:01 < bridge> well only he knows what he means 04:01 < bridge> i definitely think you should group your data tho 04:01 < bridge> ye. i'll just change it to class for now until he says other things 04:02 < bridge> i do believe double is reasonable here 04:02 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311890013763403827/image.png?ex=674a8049&is=67492ec9&hm=b5fbd5208914e14879bd49860aa1e0a767820220148e35ae69d2334c19820bbb& 04:02 < bridge> i do believe double is reasonable here tho 04:02 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311890013763403827/image.png?ex=674a8049&is=67492ec9&hm=b5fbd5208914e14879bd49860aa1e0a767820220148e35ae69d2334c19820bbb& 04:03 < bridge> weird coincidence but i guess i’m in the midst of transcribing you rn https://youtu.be/X2FWgkcTc4U?si=Hy94tq1b72q78BBn 04:03 < bridge> or maybe it isn’t weird and i’m just a nerd 04:04 < bridge> :justatest: 04:04 < bridge> i think as long as there is any reason not to use float then yeah 04:04 < bridge> use the double 04:04 < bridge> i'd rather to be freddy the fazbear 04:04 < bridge> the guy on the cover isn’t fred fwiw 04:04 < bridge> if you’re being typical chinese rn 04:04 < bridge> :justatest: 04:05 < bridge> :greenthing: well that's the only thing that is visible unless you want me to check lyrics or something 04:05 < bridge> no lyrics 04:05 < bridge> so just the vibe 04:05 < bridge> the song is just called fred 04:05 < bridge> yeah 04:05 < bridge> fair enough 04:07 < bridge> ~~oh you don't want to hear me ranting about how china is overdoing their welcome to africa countries for allying~~ 04:07 < bridge> even if i did want to 04:07 < bridge> definitely not here 04:08 < bridge> ye, still figuring out the cursor drifting 04:08 < bridge> even still no reason to hate on an american-born black guy who has nothing to do with any of that 04:08 < bridge> i don't 04:08 < bridge> :brownbear: 04:10 < bridge> what even is the pr 04:10 < bridge> cursor drifting 04:10 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311891936709185556/image.png?ex=674a8213&is=67493093&hm=e4c0a6b1bc701fa473a2cb4797ac1880fa11d59916d76a10e994d41eb21faa33& 04:10 < bridge> is this shadowing? 04:10 < bridge> is there another TargetPos? 04:10 < bridge> just these two 04:11 < bridge> oh i didnt see 04:11 < bridge> no 04:11 < bridge> if the outer scoped one was defined before that branch then yes 04:11 < bridge> if the outer scoped one was declared before that branch then yes 04:11 < bridge> oh cool 04:11 < bridge> C# doesn't like that i think i'm just rusty on C++ 04:11 < bridge> C# doesn't like that, i think i'm just rusty on C++ 04:11 < bridge> lol yeah 04:12 < bridge> it was like this but robyt3 want it scoped 04:12 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311892482627469342/image.png?ex=674a8295&is=67493115&hm=c6507a6cc7f94164845877168998b145934dbcd24ae44499699879670b885585& 04:12 < bridge> i think you’d get a warning there anyway 04:12 < bridge> well MSVC doesn't really give a lot of warning and i failed too many linux builds bcs of MSVC 04:12 < bridge> :cammo: 04:13 < bridge> it's the only compiler i have rn tho 04:13 < bridge> dont use msvc then 04:13 < bridge> the first is better bcs you make one less object 04:13 < bridge> it's just habit of setting up msvc i guess 04:13 < bridge> install msys2 -> pacman -S -> win 04:14 < bridge> it really is very good 04:14 < bridge> idk why i really didn't like msys2 04:14 < bridge> but that was like 15 years ao 04:14 < bridge> but that was like 15 years ago 04:15 < bridge> idk if the project is that old 04:15 < bridge> what were ur qualms 04:15 < bridge> i forgot. hence the idk why part 04:16 < bridge> yeah 04:17 < bridge> well it’s noticeably better now than it was when i started using it in like 2019 04:17 < bridge> pacboy makes environments/packages easier 04:18 < bridge> it started that late huh 04:18 < bridge> 2019 is like after my college already. weird. 04:18 < bridge> oh wait, read that wrong 04:19 < bridge> disregard that 04:22 < bridge> i think it was from 2017 04:23 < bridge> https://web.archive.org/web/20170207113504/https://www.msys2.org/ 04:23 < bridge> nostalgic 04:23 < bridge> oh right, it's just showing cursor during spectating 04:24 < bridge> you want it to show? 04:24 < bridge> wanted it a few years back, i've implemented this three times already, never got to the place where it was useable. ye 04:25 < bridge> honestly the original thought is to catch cheaters more obviously 04:25 < bridge> and it’s drifting? 04:26 < bridge> ye, when zooming it mismatch the local cursor 04:26 < bridge> ye, when zooming + dyncam it mismatch the local cursor 04:26 < bridge> haven't dive deeper but i'm afraid that we are sending the wrong world position in the first place. 04:27 < bridge> i'm just hoping i did something wrong when displaying the cursor 04:28 < bridge> i feel like this was talked about a few months ago 04:28 < bridge> but idr 04:28 < bridge> ye 04:28 < bridge> if i know how to interpolate, this would be a much easier pr a few years ago when it was first suggested 04:28 < bridge> with all the new exprotocol and zooming stuff 04:28 < bridge> is cursor relative to player or absolute world units? i genuinely dont know 04:29 < bridge> a bit tougher 04:29 < bridge> to interpolate you’d need a second set of coords 04:29 < bridge> cursor is screenspace locally but it was scaled somewhere to worldspace and sent to the server 04:29 < bridge> the interpolation is done already. 04:29 < bridge> why should it be worldspace 04:29 < bridge> no idea. but it is what it is. not my fault 04:30 < bridge> :kekw: 04:30 < bridge> that just seems like a great way to break physics as you move further away from the origin lol 04:30 < bridge> even when i started working on it i always assumed it was screenspace, then i was corrected. 04:31 < bridge> not like that doesn’t already happen 04:31 < bridge> the sheer thought of we are having different grid precesions in different zoom levels scares me 04:31 < bridge> 04:32 < bridge> ya 04:32 < bridge> :justatest: 04:36 < bridge> need to make a map that only works if you zoom right 04:36 < bridge> ah damn. i think i run into the same issue as souly 04:36 < bridge> do i just friend it after it was changed to class? 04:36 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311898627236696218/image.png?ex=674a884e&is=674936ce&hm=6d3b1ee1ec2c1f8296499a437475dff16508e03db07ee911945a3d27268330ac& 04:37 < bridge> I guess the world space cursor makes sense for telecursor 04:38 < bridge> unless the camera position is also sent when you're in spec or pause? 04:38 < bridge> wDyM tElEcUrSoR 04:38 < bridge> i'm so outdated on ddnet features 04:39 < bridge> it's a practice mode thing 04:39 < bridge> thank god 04:40 < bridge> cuz i just found out we are not sending cursor positions accurately unless you click or move. if we have gameplay features relying only on cursor positions i'm gonna scream 04:50 < bridge> Why friend when it's already in the class 04:50 < bridge> CGameClient doesn't have access to private members of nested class 04:51 < bridge> im great at this 04:51 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311902356866338856/image.png?ex=674a8bc7&is=67493a47&hm=fa0b1f56b5efdc0576bd8e1ac93e9b1bba72711c40aeac4b7e414296e0f0bea3& 04:52 < bridge> Could just make them public if that's necessary i guess 04:53 < bridge> You still need to use m_cursorinfo.variable to access them 04:53 < bridge> is there a verdict on that? 04:53 < bridge> no ones said they dont want this 04:53 < bridge> apart from myr 04:53 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311902836036075520/image.png?ex=674a8c3a&is=67493aba&hm=42b3b498742a1559e229bf5cc13c22351495fede65204552c95a28fcd663e804& 04:53 < bridge> i'm doing this currently 04:53 < bridge> its just that i fail to make good code 04:53 < bridge> whats this 04:54 < bridge> :nouis: is it ugly or you are just asking what pr i'm working on 04:54 < bridge> doesnt look ugly? 04:54 < bridge> i cant say i dont know 04:54 < bridge> what is it are u working on 04:54 < bridge> just making sure 04:54 < bridge> #9292 04:54 < bridge> https://github.com/ddnet/ddnet/pull/9292 04:54 < bridge> I would just make the variables public rather than getters 04:55 < bridge> also you arent meant to use "Get" for getters 04:55 < bridge> should just be Weapon() 04:55 < bridge> etc etc 04:55 < bridge> was wondering about that thx 04:55 < bridge> i will as soon as robyt3 say just make it class 04:56 < bridge> rn i'm trying to implement a dirty server side cursor position lol 04:56 < bridge> I mean public inside this class om screenshot 04:56 < bridge> just to see what is happening with that drifting 04:57 < bridge> o thats cool 04:57 < bridge> why do you need a seperate class? 04:57 < bridge> i know. i can do that later and remove the getters 04:57 < bridge> robyt3 says not to use struct 04:57 < bridge> why do you need a seperate struct 04:57 < bridge> and it is too many things to just dump into cgameclient i feel like 04:58 < bridge> it is nested 04:58 < bridge> oke ¯\_(ツ)_/¯ 04:58 < bridge> i was just following other codes tbh 04:58 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311904102086869013/image.png?ex=674a8d67&is=67493be7&hm=abcec2c7a0c468e585ee3812635b1b5f5c01490cdf0108a77d02fe4fff977840& 05:02 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311905093829922866/image.png?ex=674a8e54&is=67493cd4&hm=fab06285917cdc17fd7e5f81e10bdfa71c5f9edc7d3597e081b8ca06bfc997f9& 05:02 < bridge> 😢 05:03 < bridge> oh dear i have to rename like 20 symbols 05:03 < bridge> poor vscode is gonna struggle 05:03 < bridge> who was the guy who said this was a crutch 05:03 < bridge> yeah im not doing this manually x-x 05:04 < bridge> @kebscs bad news 05:04 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311905636258283530/image.png?ex=674a8ed5&is=67493d55&hm=f3f0228c11175af030e667a283e59179fdc77950b82f3bc50cccaffe0cd2e198& 05:04 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311905698241708073/image.png?ex=674a8ee4&is=67493d64&hm=c86ffb0280d4becfd74681272c3ec6c10b5ad67a3ac47c11daee19a092551ac2& 05:05 < bridge> spec cursor is correct, the position sent to server is wrong :justatest: 05:06 < bridge> spec cursor is correctly displaying server cursor position, the position sent to server is wrong :justatest: 05:06 < bridge> im curious how this interacts with `tc_improve_mouse_precision` 05:07 < bridge> tf is that 05:07 < bridge> tc... tater client? 05:07 < bridge> it normalizes your mouse cursor to 1000 distance 05:07 < bridge> specificially for me because i play on low sense 05:07 < bridge> specificially for me because i play on low max distance 05:07 < bridge> damn it 05:07 < bridge> i think it will work fine 05:07 < bridge> well i guess it'll just get clamped 05:07 < bridge> mite look a bit weird 05:07 < bridge> where does it read the data for local clients from? 05:08 < bridge> local cursor 05:08 < bridge> do you read from input? 05:08 < bridge> yes 05:08 < bridge> oke 05:08 < bridge> local cursor is not touched 05:08 < bridge> then thatll show wrong 05:08 < bridge> ;--; 05:08 < bridge> from mouse input 05:08 < bridge> also tater client will deal with it. this is purely client side anyway 05:08 < bridge> `m_aMousePos[g_Config.m_ClDummy]` 05:09 < bridge> it's just spectating a tater client will be weird 05:09 < bridge> or `m_aInputData[g_Config.m_ClDummy]` 05:09 < bridge> im not worried about how other people see taters 05:09 < bridge> its just yourself 05:09 < bridge> then tater will deal with it 05:09 < bridge> like the tater client will implement it their way if they were to implement it 05:09 < bridge> this is ddnet client 05:10 < bridge> (it should be this one) 05:10 < bridge> Also are targetxy still scaled by zoom? Someone fix that 05:10 < bridge> fine i'll check what ddnet is doing 05:10 < bridge> yes and it is scaled wrong 05:10 < bridge> (it should be this one) 05:10 < bridge> (or just read it from mpclients[clientid].mousepos then that works eveywhere) 05:10 < bridge> o yeah how u gonna deal with that 05:10 < bridge> i don't 05:10 < bridge> nice 05:11 < bridge> yes, make a new net msg for mouse position pls that has the mouse position and the zoom 05:11 < bridge> let's just say out of scope for this pr 05:11 < bridge> i feel like the zoom thing should be fixed first x-x 05:11 < bridge> if that was fixed the spec cursor will just show whatever is sent by the server 05:11 < bridge> Gn now tho 05:11 < bridge> night night 05:12 < bridge> we can make it screenspace (which i was doing until i found out we are using world space cursor) 05:12 < bridge> you could find the max distance and normalize that to normal zoom 05:12 < bridge> server doesn't have max distance yet 05:12 < bridge> server does? 05:12 < bridge> clients dont get them 05:12 < bridge> but u can work it out 05:12 < bridge> it does? 05:12 < bridge> like mouse distance or view distance 05:13 < bridge> someone in an fng server commented about my insanley low max mouse distance 05:13 < bridge> they just saw your cursor 05:13 < bridge> ¯\_(ツ)_/¯ 05:13 < bridge> as in they did work it out 05:13 < bridge> :justatest: 05:13 < bridge> i guess so 05:14 < bridge> i like how vscode asks me if it hasnt accidently tried to replace more things than it has 05:14 < bridge> but i dont even check and just press yes cuz its always right x-x 05:15 < bridge> @tsfreddie would cl_show_hook_coll_other 2 not do what u want 05:15 < bridge> actually 05:15 < bridge> for cheat detection at least 05:16 < bridge> cheat detection wasn't really possible anyway, so i shifted to make it at least look good for normal players 05:18 < bridge> I guess dyncam extra range isn't scaled by zoom 05:19 < bridge> it really shouldn't be scaled imho 05:19 < bridge> it really shouldn't be scaled in the first place imho 05:59 < bridge> :brownbear: i hate timezone 05:59 < bridge> now i need to wait a day for comments on pr and stuff 06:03 < bridge> mood 06:03 < bridge> european morning 06:04 < bridge> @tsfreddie `tc_improve_mouse_precision` scales mouse position such that the max mouse distance looks like 1000 if it is not already greater 06:04 < bridge> i used to use a fixed factor but when coming to put it into tc we came up with a better... way 06:05 < bridge> it's probably a interpretation thing 06:05 < bridge> by fixed scale i mean scaling it not just normalize it 06:05 < bridge> well it does normalize it 06:05 < bridge> before then multiplying it by a facotr based on your max mouse distance 06:05 < bridge> ah 06:05 < bridge> before then multiplying it by a factor based on your max mouse distance 06:06 < bridge> before then multiplying it by a factor based on your max mouse distance (such that it goes from 0 to 1000) 06:06 < bridge> either way it doesn't really matter how tater does it tho 06:06 < bridge> before then multiplying it by a factor based on your max mouse distance (such that it goes from 0 to 1000) (TODO: make it work with min mouse distance) 06:06 < bridge> ddnet need to decide on the route to go for 06:06 < bridge> cuz if we normalize it, we then need another way to send cursor position for spec cursor to work properly 06:07 < bridge> or we can decide spec cursor is not worth the precision trade off and normalize it and scape the spec cursor pr 06:07 < bridge> any way is okay but we (as in everyone here) gotta settle on one idea 06:08 < bridge> :kekw: 06:08 < bridge> i was earlier concerned about if you spectate yourself and then move your mouse 06:08 < bridge> you would see your normalized mouse 06:08 < bridge> which means this probably will go unfixed for quite a while 06:08 < bridge> again, i didn't change anything about local cursor and tater client does not affect ddnet client 06:08 < bridge> the spec cursor pr is purely client-side 06:09 < bridge> no but you may pull data for local clients from a place where its already been normalized 06:09 < bridge> or you think i stored local cursor positions in demo or something 06:10 < bridge> i did not touch local cursor 06:10 < bridge> 🤦‍♂️ 06:10 < bridge> what ddnet used to look like after that pr it will still look like that 06:10 < bridge> ill just look at what u did 06:10 < bridge> 06:10 < bridge> there is no "spectate yourself" you know 06:11 < bridge> press spectate then click 06:11 < bridge> spectate yourself is just spectating the server tee of you 06:11 < bridge> which is not local 06:11 < bridge> it is yourself but it is not local 06:11 < bridge> atleast for the angle is uses local 06:11 < bridge> atleast for the mouse position is uses local 06:11 < bridge> ug syfrkg lrf akdif 06:11 < bridge> ye 06:11 < bridge> atleast for the mouse position it uses local 06:11 < bridge> that means it doesn't use angle in qwertymak 06:12 < bridge> :nouis: 06:12 < bridge> (angle is just a property of a mouse position, its never stored anywhere) 06:12 < bridge> the angle u see on urself when spectating is random shit 06:12 < bridge> once in specview the data fully comes from server 06:12 < bridge> so in spec cursor there is no spectating yourself i guess 06:13 < bridge> you cant spectate yourself? 06:13 < bridge> thats weird 06:13 < bridge> you can 06:13 < bridge> but it is not local 06:13 < bridge> it just renders the server cursor position 06:13 < bridge> when you are in spectate mode 06:13 < bridge> (you cant click onto yourself when spectating) 06:14 < bridge> just pause and use right shift 06:14 < bridge> 06:14 < bridge> o that works? 06:14 < bridge> it does :D 06:14 < bridge> it does use local mouse position tho 06:14 < bridge> im on a server with 300 ping 06:14 < bridge> fine let me check 06:15 < bridge> I've been sitting in freeze for two hours 06:15 < bridge> have u tried space heater 06:15 < bridge> and fikmesån joined 06:15 < bridge> did not expect that 06:15 < bridge> stop worry about weird stuff 06:15 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311923565993398302/2024-11-29_13-15-21.mp4?ex=674a9f88&is=67494e08&hm=1f7695bebc308e6f7ef119dfd142cd438337aeb37e8ccbd6c5fdd22684579a2a& 06:15 < bridge> wait why are you doing inerpolation yourself 06:16 < bridge> doesnt prediction do that already? 06:16 < bridge> for testimg 06:16 < bridge> ? 06:16 < bridge> extrapolation on cursor? no thanks 06:16 < bridge> even linear interpolation looks bad 06:17 < bridge> also sample points doesn't tell you whether the cursor moves. server can say cursor position x is 2 2 5 5 9 9, and linear interp will just jitter each step. 06:17 < bridge> and it does say that because we don't send cursor position every tick at 50hz 06:18 < bridge> we get jittery mouse by default 06:18 < bridge> wait mouse isnt interpolated at all 06:19 < bridge> would it be possible to shove what you have into prediction 06:19 < bridge> of course it isn't... mouse cursor position isn't even a teeworlds thing 06:19 < bridge> we added it in ex protocol 06:19 < bridge> no 06:19 < bridge> extrapolation on mouse cursor will just make it fly out of screen most of the time 06:20 < bridge> is it not? 06:20 < bridge> i thought angle was never sent 06:20 < bridge> it was always mouse position 06:20 < bridge> the infclass selection wheel works in 0.7 06:20 < bridge> it's the other way around i think 06:20 < bridge> ? 06:21 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311924898767310859/image.png?ex=674aa0c6&is=67494f46&hm=cfaa675db33960f4e8c5b464ae4f9ad333cdeab0ca153ca4707f4947ea292506& 06:21 < bridge> https://chillerdragon.github.io/teeworlds-protocol/06/snap_items.html#obj_character_core 06:21 < bridge> Freddie was faster but yes 06:21 < bridge> ghasp! 06:21 < bridge> wait why 06:21 < bridge> i man 06:21 < bridge> i mean 06:21 < bridge> even in 0.7 06:21 < bridge> it is still angle 06:22 < bridge> why not? cursor position is not used for anything on client-side 06:22 < bridge> ive never seen anywhere angle used its always a mouse position turned into an angle where needed 06:22 < bridge> literally nothing in vanilla teeworlds calls for a target position client side 06:23 < bridge> it sends a single low precision angle just to show the angle player is pointing at 06:23 < bridge> saves a varint i guess 06:23 < bridge> but 0.7 infclass wheel works 06:23 < bridge> that's server side? 06:23 < bridge> Based on the angle I guess 06:24 < bridge> So even if your cursor is not far enough distance wise it should still select 06:24 < bridge> but you cant select it unless your ein dyn 06:24 < bridge> Yes the wheel is server side 06:24 < bridge> but you cant select it unless your mouse is far enough 06:24 < bridge> idk how client side map reads angle 06:24 < bridge> Weird 06:24 < bridge> server does have player mouse position 06:24 < bridge> client only sees other players angle 06:24 < bridge> idk what's bothering you lol 06:25 < bridge> https://chillerdragon.github.io/teeworlds-protocol/06/snap_items.html#obj_character 06:25 < bridge> i thought it was angles all the way down 06:25 < bridge> There is HookX 06:25 < bridge> i thought it was mouse positions all the way down 06:25 < bridge> But that’s only on hook I guess 06:25 < bridge> that's kinda needed for pred 06:25 < bridge> i think 06:25 < bridge> and animation probably 06:25 < bridge> pick an angle or mouse position dont use both D: 06:25 < bridge> i would do the same to save bandwidth tbh 06:26 < bridge> Wait I got lost 06:26 < bridge> vanilla teeworlds seems to care a lot about each byte. until 0.7 06:26 < bridge> The client sends its mouse pos as coordinates 06:26 < bridge> But the server does not 06:26 < bridge> So the server knows the client taget x and y https://chillerdragon.github.io/teeworlds-protocol/06/snap_items.html#obj_player_input 06:26 < bridge> how did chiller got lost 06:26 < bridge> then how does mr tsfreddies thing work 06:26 < bridge> But the client does not know the other players target x and y 06:26 < bridge> bruh 06:26 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1311926321374560297/image.png?ex=674aa219&is=67495099&hm=429043c66d4468efcaa3024b25f42b4e4998a2a534c23b7f5c0f6706a4328100& 06:26 < bridge> 😢 06:27 < bridge> it's a ddnet thing 06:27 < bridge> i kinda hate this tbh. too many bandwidth used for not a lot 06:27 < bridge> i kinda hate this tbh. too much bandwidth used for not a lot for regular player 06:27 < bridge> i guess hook collision for other player is better? 06:28 < bridge> my first attempt at spec cursor is to just send target position per player. not per player^2. then this was added when i tried it the second time. 06:28 < bridge> this is my third time implementing spec cursor. now we have a weird cursor drifting bug 06:28 < bridge> :kekw: 06:29 < bridge> #3262 this feature started in 2020 bruh 06:29 < bridge> https://github.com/ddnet/ddnet/pull/3262 06:33 < bridge> 👀 06:34 < bridge> ```cpp 06:34 < bridge> int StrongWeakSpriteId; 06:34 < bridge> switch(Data.m_HookWeakStrong) 06:34 < bridge> { 06:34 < bridge> case TRISTATE::ALL: 06:34 < bridge> HookWeakStrongColor = color_cast(ColorHSLA(6401973)); 06:34 < bridge>