06:37 < bridge> guten morgen 07:53 < bridge> @0xdeen have you tried reverting the spectator features in 18.9? should be these commits in reverse chrono order. 07:53 < bridge> ``` 07:53 < bridge> fe3984846fbc8e0319c6a407c8f7d353ddda5d8e 07:53 < bridge> 9863c0f57e099dbcf2cba0e1f39b014975d9f367 07:53 < bridge> abbd20b07cebd8c005b25a42fd55a6655e0fa66a 07:53 < bridge> 288a7a3a19e35bde0cd61c92549d5021d15211c9 07:53 < bridge> ea9a0a3a9c7947644fd0bba950e08d174c321797 07:53 < bridge> ``` 07:53 < bridge> I tried on master they revert just fine without conflict it seems. but i imagine you need to do it in a 18.9 branch manually. 08:46 < bridge> gitignore prs ftw 09:35 < bridge> <0xdeen> perfect, thanks 09:50 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1326110668222496818/GgqKj2-WsAAXBba.png?ex=677e3c4d&is=677ceacd&hm=4dbc599db3f8e661cadfe62f2a246e89c6096bb0be86c81ad99693c05ea25f14& 09:50 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1326110720760348754/RTX-5000-1-2025-01-9ebd4a2761eacb069eb2e2f9244065cc-1024x576.png?ex=677e3c59&is=677cead9&hm=bb8ab7eba38bfebff618e985425c49ce0b7bc48737a06da49e97a8e2b8973b49& 09:50 < bridge> :justatest: 09:50 < bridge> can it run rust 10:27 < bridge> how am i meant to get the number of players in team 0? CGameTeams::Count returns 64 10:27 < bridge> finally the gradle demon was giving me nightmares 10:28 < bridge> loop thru clients 10:28 < bridge> 1. check if connected 10:28 < bridge> 2. check if in team 10:29 < bridge> sounds annoying 10:29 < bridge> yep 11:07 < bridge> from client? 11:07 < bridge> server 11:09 < bridge> then yep seems bun bun solution is only one to get it right 11:09 < bridge> i wonder in what context Count() is being used tho 11:09 < bridge> with 64 in t0 xd 11:18 < bridge> ah yes, making DLSS a necessity :kek: 11:29 < bridge> my next gpu upgrade 11:29 < bridge> https://fxtwitter.com/tsarnick/status/1876470407697772848 11:33 < bridge> they probably get the intern to develop the gaming gpus these days, they're all too busy making their 3000% margin $10,000,000 single server 12:02 < bridge> imagine he drops it 12:25 < bridge> Its definitely a dummy or those are chips that were defective off the wafer 12:32 < bridge> Wtf xddd 12:32 < bridge> Average GPU to run electron 13:36 < bridge> deen moment, ngl he is right tho 13:58 < bridge> at least change it to `2013 - ` 14:02 < ws-client> @hectavoxel oh yea including the start year sounds nice 14:02 < ws-client> go send a pr to troll deen 14:02 < bridge> cant, have to go to work 14:03 < ws-client> and next year we change it from ``2013 -`` to ``since 2013`` and then in 2027 we change it to ``2013 - today`` 14:40 < bridge> :greenthing: 14:41 < bridge> I feel a bit uncomfortable spamming this channel 14:42 < ws-client> dont worry jupstar github webhook is not bridged to irc nobody sees your spam 14:45 < bridge> atleast ur commits are not "ok" xd 14:45 < bridge> or some random word 14:45 < bridge> like i do :justatest: 14:46 < bridge> next jupstar commit: "fix" 14:47 < bridge> push directly and reduce 2/3 of the spam 14:52 < bridge> believe me the git commit msgs of the last 2 years been "u". 14:52 < bridge> 14:52 < bridge> And I almost created prs with that here too xDD 14:52 < bridge> muscle memory 14:52 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1326186809373364286/image.png?ex=677e8336&is=677d31b6&hm=1838223b086d1e40837f464084f1da4ff05ba0633bb35278e85771e1a7ab009f& 14:55 < bridge> xd 14:55 < bridge> u 14:55 < bridge> no u 15:30 < bridge> https://en.wikipedia.org/wiki/Uniform_Function_Call_Syntax 17:10 < bridge> https://docs.rs/dispose/latest/dispose/index.html 17:10 < bridge> @jupeyy_keks check this xd 17:13 < bridge> I wish rust would support that instead of the current Drop trait 😄 17:13 < bridge> ofc without a wrapper ^^ 17:15 < bridge> `#[non_exhaustive]` is also such a design flaw.. it forces me to ignore future cases 17:15 < bridge> I cannot understand libraries use it 17:15 < bridge> I'd prefer a compile time error xd 17:20 < bridge> @heinrich5991 hi, is there any issue with master4 ? 17:20 < bridge> ``` 17:20 < bridge> v3royal | [677d541a][register]: WARNING: Master server is not responding, switching master 17:20 < bridge> v3royal | [677d541a][engine/mastersrv]: refreshing master server addresses 17:20 < bridge> v3royal | [677d541a][register]: refreshing ip addresses 17:20 < bridge> v3royal | [677d541a][engine/mastersrv]: saving addresses 17:21 < bridge> v3royal | [677d541a][register]: fetching server counts 17:21 < bridge> v3royal | [677d541d][register]: chose 'master4.teeworlds.com' as master, sending heartbeats 17:21 < bridge> ``` 17:21 < bridge> No changes on my side 17:21 < bridge> <0xdeen> It has reduced traffic for 1 month now, let me check 17:22 < bridge> alright thank you 17:24 < bridge> note that this is legacy registering, not https (forgot to mention) 17:25 < bridge> <0xdeen> yes, that was clear from the message 17:25 < bridge> <0xdeen> working now? 17:25 < bridge> <0xdeen> I just restarted it, didn't check where the mastersrv got stuck 17:38 < bridge> no clue what is happening, from the tcpdump my server sends a query but no answer: 17:38 < bridge> ``` 17:38 < bridge> 16:38:06.483390 IP (tos 0x10, ttl 63, id 2707, offset 0, flags [DF], proto UDP (17), length 42) 17:38 < bridge> 57.129.64.12.8303 > 49.12.97.180.8300: [bad udp cksum 0x0c75 -> 0xd9f7!] UDP, length 14 17:38 < bridge> 0x0000: 4510 002a 0a93 4000 3f11 24d3 3981 400c E..*..@.?.$.9.@. 17:38 < bridge> 0x0010: 310c 61b4 206f 206c 0016 0c75 ffff ffff 1.a..o.l...u.... 17:38 < bridge> 0x0020: ffff ffff ffff 636f 7532 ......cou2 17:38 < bridge> 16:38:07.483396 IP (tos 0x10, ttl 63, id 2904, offset 0, flags [DF], proto UDP (17), length 42) 17:38 < bridge> 57.129.64.12.8303 > 49.12.97.180.8300: [bad udp cksum 0x0c75 -> 0xd9f7!] UDP, length 14 17:38 < bridge> 0x0000: 4510 002a 0b58 4000 3f11 240e 3981 400c E..*.X@.?.$.9.@. 17:38 < bridge> 0x0010: 310c 61b4 206f 206c 0016 0c75 ffff ffff 1.a..o.l...u.... 17:38 < bridge> 0x0020: ffff ffff ffff 636f 7532 ......cou2 17:38 < bridge> 16:38:08.483403 IP (tos 0x10, ttl 63, id 2993, offset 0, flags [DF], proto UDP (17), length 42) 17:38 < bridge> 57.129.64.12.8303 > 49.12.97.180.8300: [bad udp cksum 0x0c75 -> 0xd9f7!] UDP, length 14 17:38 < bridge> 0x0000: 4510 002a 0bb1 4000 3f11 23b5 3981 400c E..*..@.?.#.9.@. 17:38 < bridge> 0x0010: 310c 61b4 206f 206c 0016 0c75 ffff ffff 1.a..o.l...u.... 17:38 < bridge> 0x0020: ffff ffff ffff 636f 7532 ......cou2 17:38 < bridge> 16:38:09.483377 IP (tos 0x10, ttl 63, id 3233, offset 0, flags [DF], proto UDP (17), length 44) 17:38 < bridge> 57.129.64.12.8303 > 49.12.97.180.8300: [bad udp cksum 0x0c77 -> 0xce8e!] UDP, length 16 17:38 < bridge> 0x0000: 4510 002c 0ca1 4000 3f11 22c3 3981 400c E..,..@.?.".9.@. 17:38 < bridge> 0x0010: 310c 61b4 206f 206c 0018 0c77 ffff ffff 1.a..o.l...w.... 17:39 < bridge> 0x0020: ffff ffff ffff 6265 6132 206f ......bea2.o 17:39 < bridge> ``` 17:39 < bridge> 10sec after it retries 17:40 < bridge> with an mtr I can reach to the end (tho I don't have an answer from the master server I guess as it doesn't know how to answer) 18:05 < bridge> <0xdeen> damn, then we have to check closer. iptables was clean so my only other guess is that the hoster might be blocking traffic 18:10 < bridge> whats SpecChar 18:11 < bridge> seem to always be 0 18:12 < bridge> i think on a server with spec enabled 18:12 < bridge> its the circle char 18:12 < bridge> but im probs wrong 18:12 < bridge> back then i think i implemented the spec char :justatest: 18:13 < bridge> ah yea the outline 18:52 < bridge> <0xdeen> I checked tcpdump and the packets come in and the responses go out: 18:52 < bridge> <0xdeen> ``` 18:52 < bridge> <0xdeen> 18:51:40.729306 eth0 In IP bw.reitw.fr.8305 > master.ddnet.org.8300: UDP, length 16 18:52 < bridge> <0xdeen> 18:51:40.744488 eth0 In IP bw.reitw.fr.8306 > master.ddnet.org.8300: UDP, length 16 18:52 < bridge> <0xdeen> 18:51:40.750681 eth0 In IP bw.reitw.fr.8304 > master.ddnet.org.8300: UDP, length 16 18:52 < bridge> <0xdeen> 18:51:41.540856 eth0 Out IP master.ddnet.org.8300 > bw.reitw.fr.8307: UDP, length 14 18:52 < bridge> <0xdeen> 18:51:41.550004 eth0 Out IP master.ddnet.org.8300 > bw.reitw.fr.8305: UDP, length 14 18:52 < bridge> <0xdeen> 18:51:41.550048 eth0 Out IP master.ddnet.org.8300 > bw.reitw.fr.8304: UDP, length 14 18:52 < bridge> <0xdeen> 18:51:41.550108 eth0 Out IP master.ddnet.org.8300 > bw.reitw.fr.8306: UDP, length 14 18:52 < bridge> <0xdeen> ``` 18:52 < bridge> <0xdeen> Maybe something on OVH side blocking it? 18:53 < bridge> <0xdeen> Oooh, 8303 is missing! 18:53 < bridge> <0xdeen> I'll ask our hoster 18:54 < bridge> why my 1 day ddnet rs pr has merge conflicts already xd 18:58 < bridge> <0xdeen> @reitw is the server even running? I can't connect to it on 8303, so seems like a problem on your side 19:01 < bridge> he cooks 19:02 < bridge> yes that's ovh 19:02 < bridge> forgot to notify you sorry 19:02 < bridge> someone is having fun since yesterday apparently, whenever i change my port to fix the issue he targets it 19:49 < bridge> wait theres a difference ebtween ddrace and ddnet 19:50 < bridge> i thought ddrace was just the gamemode and ddnet a thing that hosts alot of it which happens to have a client 19:50 < bridge> or is 2011 when the client became a thing 20:05 < ChillerDragon> ``time ./DDNet quit`` 20:05 < ChillerDragon> ``real 0m40.904s`` 20:05 < ChillerDragon> i still miss the times quitting ddnet was instant tbh 20:06 < ChillerDragon> my internet bugged and i couldn't close the client anymore -.- 20:06 < bridge> Do same test with rs 20:06 < bridge> DDNet is also the server software, the most relevant implementation of DDRace 20:08 < ChillerDragon> my internet is now back so its ddnet ``real 0m2.060s`` vs ddnet-rs ``real 0m1.569s`` 20:08 < bridge> too slow. close the window and do all the calculation afterwards 20:18 < bridge> waat, is that a debug build? xd 20:19 < ChillerDragon> yes debug 20:26 < bridge> i remember instant closing then the next time i updated it was a couple of seconds 20:26 < bridge> why is ... that 20:26 < bridge> very simple, we cannot cancel http requests 20:27 < bridge> and since file writing and requests use the same tasks handler 20:27 < bridge> they must be handled, bcs file writing must succeed 20:28 < bridge> "must succeed" :justatest: 20:37 < bridge> why can’t you fork there and at least close the window? 20:38 < bridge> Some jobs still use the graphics 20:39 < bridge> And Windows can't fork 20:41 < bridge> just CHttpJobs or whatever? 20:41 < bridge> those seem to be the culprits 20:42 < bridge> they can also be changed to be conscious of the game state so if it’s been forked while blocked by http request then it will just quit immediately without graphics calls 20:42 < bridge> but yea idk about windows 20:42 < bridge> there must be a similar mechanism 20:42 < bridge> it would work on mingw but that isn’t a real solution 20:43 < bridge> those seem to be the culprit 20:52 < bridge> All image loading jobs, like those for skins, need to stop before the graphics can be shutdown 20:53 < bridge> Cygwin implements fork on Windows but it's a lot slower, that's also why `make` is slower on Windows IIRC 20:53 < bridge> `http: https://master3.ddnet.org/ddnet/15/servers.json failed. libcurl error (42): Callback aborted` 20:54 < bridge> yeah, but it's up to curl to cancel it 20:54 < bridge> i see, but it's weird that it doesnt work for chiller 20:54 < bridge> cancelation will only go though when there is progress AFAIK, so if there is zero progress than the cancelation doesn't work 20:54 < bridge> 😮 20:54 < bridge> rip 20:54 < bridge> cancelation will only go through when there is progress AFAIK, so if there is zero progress than the cancelation doesn't work 20:55 < bridge> what would ddnet be like if it just killed all windows support 20:55 < bridge> a 4D game 20:55 < bridge> Probably less alive, since more than 90% of players use windows 20:56 < bridge> (probably, I don't have any stats) 21:27 < bridge> ChillerDragon: i think i'm gonna stick to writing a teeworlds protocol library in rust, a whole server is too much lmao 21:28 < bridge> i think i will still continue making the server but not in bevy 21:28 < bridge> and it will not be a priority 21:29 < bridge> You can wait for those, or use an alternative solution 21:30 < bridge> i speak of libwinpthread which is independent of cygwin, not sure about its performance but i dont think it should matter too much in this instance 21:35 < bridge> @hectavoxel this guarantees the skin sounds are loaded btw 21:36 < bridge> usually the map loading is the slowest anyway 21:36 < bridge> for ctf1 maybe not xdd 21:43 < bridge> 🔥 22:19 < ws-client> arent you using libtw2? so the lib is done or what is missing @kollpotato 22:27 < bridge> but that would only increase stop time when there are http jobs and most of the time when im closing its after the windows been idle for a while or in game the time is also pretty consistent 22:27 < bridge> and why cant you cancel them, curl surely lets you do that, you can just delete the file it was writing to 22:28 < bridge> ah 22:29 < bridge> i mean its better than no cancellation