07:14 <+bridge> [ddnet] GUYS 07:15 <+bridge> [ddnet] Hello and sorry for disturb 07:15 <+bridge> [ddnet] Why i have closed eyes for all tees on mine personal DDNET server? 07:15 <+bridge> [ddnet] what command to able use eyeemote 07:19 <+bridge> [ddnet] I mean: all tees on server have closed eyes, what command responding for it? 07:45 <+bridge> [ddnet] close eyes may be because your tees are set as afk or you have /pause 07:45 <+bridge> [ddnet] @Obiwan try to use command /emote 07:50 <+bridge> [ddnet] nono 07:50 <+bridge> [ddnet] not in afk, just ingame 07:50 <+bridge> [ddnet] kind no paused, and other 07:51 <+bridge> [ddnet] but you gave me one idea 07:55 <+bridge> [ddnet] i setted: events 0; sv_emotional_tees 1; and i trying to make eyemote, but nothing happens 07:55 <+bridge> [ddnet] but in server console writing this: 07:56 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/766902357253947412/unknown.png 08:02 <+bridge> [ddnet] have you tried other emotes than normal? 08:02 <+bridge> [ddnet] really weird thing 08:06 <+bridge> [ddnet] ye 08:06 <+bridge> [ddnet] tryed all 08:06 <+bridge> [ddnet] i searching for a problem rn 08:08 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/766905351303725076/screenshot_2020-10-17_09-07-59.png 08:16 <+bridge> [ddnet] eyeemote not works on server but saying: you can no use preset of emotes 08:16 <+bridge> [ddnet] eyeemote not works on server but saying: you can no use the preset eye emotes 08:17 <+bridge> [ddnet] @Moderator 08:17 <+bridge> [ddnet] @Moderator HALP ^^ 08:17 <+bridge> [ddnet] @Moderator HALP 09:43 <+bridge> [ddnet] @Souly Hey 09:43 <+bridge> [ddnet] I repaired 09:43 <+bridge> [ddnet] But... wtf how it works 09:44 <+bridge> [ddnet] seems ddnet doing same as discord: Getting much new features with non-normal explaining how it works (as you noticed when i always asking, the Deen answering to me all time) and still buggy 09:46 <+bridge> [ddnet] I changed value in this command: sv_max_afk_vote_time from 0 to 300 (default) , but idk how "vote_time" is refers to eyeemote ~_~ 09:46 <+bridge> [ddnet] I changed value in this command: sv_max_afk_vote_time from 0 to 300 (default) , but idk how "vote_time" is refers to eyeemote ... 09:47 <+bridge> [ddnet] because the afk state changes your eye emote 09:48 <+bridge> [ddnet] it's 300 by default, so it's your own fault 09:48 <+bridge> [ddnet] https://ddnet.tw/settingscommands/ 09:50 <+bridge> [ddnet] If you are on fng There is an extra command for wheel and it must be set in cfg before start 09:52 <+bridge> [ddnet] The problem was in non-working eye emotes, on server i cant do any eye emote, but i tryed to search for some functions to work it kind: sv_max_afk_time; sv_input_info and another, but problem was in "sv_max_afk_vote_time" - (The time in seconds a player can be afk and his _votes_ still count (0 = disabled) 300) but cant imagine how VOTE VOICE TIME refers to EYE EMOTE. If somebody will have same problem, set "sv_max_afk_vote_time" to any n 09:52 <+bridge> [ddnet] I done! 09:52 <+bridge> [ddnet] :kek: 09:52 <+bridge> [ddnet] cant understand, sorry 09:53 <+bridge> [ddnet] @Jupstar ✪ About what extra command you told? 09:53 <+bridge> [ddnet] @Jupstar ✪ About what extra command you told? And for what wheel: eye emote or just emote? 09:54 <+bridge> [ddnet] Is it even fng server else nvm 09:54 <+bridge> [ddnet] Yes, but i interested, please, can you explain more in details? 09:55 <+bridge> [ddnet] Just read fng.cfg 09:55 <+bridge> [ddnet] And the github page 09:55 <+bridge> [ddnet] It explains everything 09:56 <+bridge> [ddnet] oh, okay, thanks 10:36 <+bridge> [ddnet] Statistics don't look good on phone 10:36 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/766942733503561738/Screenshot_20201017-103558_Chrome.png 11:10 <+bridge> [ddnet] I'll try to fix 11:38 <+bridge> [ddnet] @Lady Saavik better now? 11:40 <+bridge> [ddnet] yes @deen thank you 12:18 <+ChillerDragon> oof I expected to be able to rejoin with timeout after ``pkill -f DDNet`` seems like it sends a SIGTERM and ddnet closes and disconnects cleanly :/ 12:18 <+bridge> [ddnet] yes, use kill -9 12:18 <+ChillerDragon> yea now i know 12:18 <+ChillerDragon> kill sounds so strong i should keep in mind it trys to be gentle 12:20 <+ChillerDragon> gnome bugged on me i never had something like this beofore :/ 12:20 <+ChillerDragon> time to switch to KDE @Patiga :D 12:20 <+bridge> [ddnet] ! 12:21 <+bridge> [ddnet] be sure to enable wobbly windows in the desktop effects! 12:25 <+ChillerDragon> xd 12:25 <+ChillerDragon> omg it happend again 12:25 <+ChillerDragon> i blame ddnet client 12:25 <+ChillerDragon> ddnet is virused gnome crasher 12:27 <+ChillerDragon> ok it was a joke 12:27 <+ChillerDragon> but now im serious 12:27 <+ChillerDragon> when i get pinged in ddnet chat my gnome crashed 12:28 <+ChillerDragon> 2 times in a row 12:28 <+bridge> [ddnet] I think I nkow what's causing this 12:28 <+ChillerDragon> cl_notifications 0 it is for now i guess 12:28 <+bridge> [ddnet] 8ce6de6d8e 12:28 <+bridge> [ddnet] try reverting that 12:30 <+bridge> [ddnet] does anyone know the reasoning behind map sounds being allowed to be external btw? 12:30 <+ChillerDragon> reverted can someone come ingame and ping me? :) 12:30 <+bridge> [ddnet] just because it's the same in images 12:31 <+ChillerDragon> surived ty @Patiga 12:31 <+ChillerDragon> deen too slow xd 12:31 <+bridge> [ddnet] ^^ 12:31 <+ChillerDragon> ok seems ur guess was right 12:31 <+ChillerDragon> i wasnt running master before but the ddnet-git aur package but it should be the commit anyways 12:32 <+bridge> [ddnet] ah ye i also dont really understand why the images are not automatically sorted into external/embedded 12:32 <+ChillerDragon> oh cmn i lost map progress again due to CTRL-C xd 12:32 <+ChillerDragon> i suck 12:32 <+ChillerDragon> guess now i know 12:33 <+ChillerDragon> oh wait guys the test was flawed xd 12:33 <+ChillerDragon> i had cl_notifications 0 12:34 <+ChillerDragon> rip 12:35 <+ChillerDragon> ok @deen seems like reverting 8ce6de6d8e520f96724d9e2c4c8da8f24e3e72a8. did not help 12:36 <+ChillerDragon> i can try to bisect somewhen else but im currently in the middle of something 12:36 <+ChillerDragon> ill open a issue for now i guess 12:42 <+ChillerDragon> lmao i manged to get pinged in chat in a message without my name xd https://zillyhuhn.com/cs/.1602931225.png 12:49 <+bridge> [ddnet] https://en.wikipedia.org/wiki/GNOME_Shell 12:49 <+bridge> [ddnet] Written in **JavaScript** and C 12:49 <+bridge> [ddnet] it cannot be taken seriously 😄 12:51 <+ChillerDragon> both fine and fun languages 12:54 <+bridge> [ddnet] you think using things like JS for core system components is OK? 12:55 <+bridge> [ddnet] a lot of people were making fun of Windows when it had some GUI things integrated in the kernel, and now core of main Linux desktop environment is written in JS.... 😄 13:40 <+ChillerDragon> is the core written in js? tbh idk i thought its more for the extensions and stuff 13:40 <+ChillerDragon> btw coming from the dude who runs a electron app to chat xd 13:40 <+ChillerDragon> :fire: 14:51 <+bridge> [ddnet] Discord's web version works well 14:51 <+bridge> [ddnet] no reason to use that Electron junk 😄 14:55 <+bridge> [ddnet] Not much difference between running it in a browser or in electron. Might aswell get the better integration of electron 14:57 <+bridge> [ddnet] when you run in it web browser, it can't mess with your system much 14:57 <+bridge> [ddnet] but with Electron it can 14:58 <+bridge> [ddnet] sneaking what apps are running etc 14:59 <+bridge> [ddnet] when you run it in web browser, it can't mess with your system much 16:59 <+ChillerDragon> I finished the grass_doodads svg version btw https://github.com/TeeworldsDB/mapres/blob/master/svg/grass_doodads_chiller.svg Idk how to export pngs from it properly and its untested but at least its finished xd 17:00 <+bridge> [ddnet] nice looks good 17:00 <+ChillerDragon> :) 17:11 <+bridge> [ddnet] @heinrich5991 a2b3f45f8b105ebbd6e391c48a2c529ddb1ed09c broke our configurations 17:13 <+bridge> [ddnet] `Cannot execute game command 'sv_team', put it into the map config or start the server with 'sv_test_cmds 1' to enable it` most critically `sv_solo_server` 17:14 <+bridge> [ddnet] @deen is ger4 not compiling it's own server? 17:19 <+bridge> [ddnet] How is every single one of our servers running a different version? :/ 17:21 <+bridge> [ddnet] ChillerDragon: Want to add ultra hd option in client? 17:22 <+bridge> [ddnet] @Learath2 they are all built on ddnet.tw and scped from there 17:22 <+bridge> [ddnet] @Learath2 so solo server is currently broken? Should we quickly revert and deploy? 17:23 <+bridge> [ddnet] so all servers should run the same version, depending on when they last restarted of course 17:23 <+bridge> [ddnet] I'm just going to revert `7e84e1ef43516beeab579a850d664242e91d2843` @heinrich5991 can re PR when he is around. We basically need some sort of unsafe-exec or atleast a time when we can unsafe-exec 17:23 <+bridge> [ddnet] ok, thanks 17:23 <+bridge> [ddnet] How do I get it to deploy to all servers btw? 17:23 <+bridge> [ddnet] yeah, before map start would be great 17:23 <+bridge> [ddnet] `ssh ddnet.tw "cd src; ./build.sh` 17:23 <+bridge> [ddnet] " 17:24 <+bridge> [ddnet] `~/servers/all-restart.sh` afterwards to start running a bit earlier 17:24 <+bridge> [ddnet] also on ddnet.tw, no need to connect to the other servers manually (they are too many for that now anyway) 17:26 <+ChillerDragon> @deen is this a question or a request? 17:27 <+bridge> [ddnet] A joke/request/question combo 17:27 <+bridge> [ddnet] mostly just an idea of what we could do with the svgs 17:27 <+bridge> [ddnet] Did you know that it's not trivial to revert merge commits? 17:27 <+bridge> [ddnet] yes 17:28 <+bridge> [ddnet] revert all changes in the PR in reverse order 17:28 <+bridge> [ddnet] and squash that with `git rebase -i master` 17:28 <+bridge> [ddnet] is what I would do 17:29 <+bridge> [ddnet] Odd that git doesn't have that as default 17:29 <+ChillerDragon> @deen well i just made the individual doodads for my movie and website and did the whole tileset on request on Jupstar. I play in high res I would love to see better graphics but I currently do not plan to thinker with client code reagarding this 17:30 <+bridge> [ddnet] I assume git revert -m with the correct parent would do that 17:32 <+bridge> [ddnet] but apparently that has the weird behaviour of still leaving A and B in the repository, so after @heinrich5991 fixes it, re merging it would apparently only merge the new commits he makes 17:32 <+bridge> [ddnet] https://github.com/git/git/blob/master/Documentation/howto/revert-a-faulty-merge.txt 17:32 <+bridge> [ddnet] ok, never used that 17:33 <+bridge> [ddnet] @deen are you perhaps editing build.sh on ddnet.tw? 17:35 <+bridge> [ddnet] I opened it in vim, but only read 17:35 <+bridge> [ddnet] because you asked me about it 😄 17:38 <+bridge> [ddnet] I still think we should add the git hash of the server on the finishes 17:38 <+bridge> [ddnet] yes 17:38 <+bridge> [ddnet] yes 17:38 <+bridge> [ddnet] would help with bugs 17:38 <+bridge> [ddnet] is it in the teehistorian files? 17:39 <+bridge> [ddnet] then we could extract it retroactively from all existing teehistorian files 17:41 <+bridge> [ddnet] Yep it is 17:42 <+bridge> [ddnet] `"server_version":"0.6.4, 15.0.5 (d3691bf221c2d6d2)"` 17:42 <+bridge> [ddnet] so basically we have the info already, it's just not so easy to access 17:42 <+bridge> [ddnet] @deen Copying to servers seems to hang 17:42 <+bridge> [ddnet] @Learath2 it's done slowly and some servers (CHN) can do 5 KB/s 17:42 <+bridge> [ddnet] Ah 17:43 <+bridge> [ddnet] yeah 17:43 <+bridge> [ddnet] Hm, that might be a good reason to compile on those servers 17:43 <+bridge> [ddnet] ps aux|grep scp shows the CHN servers are the ones 17:43 <+bridge> [ddnet] but `git pull` can also take 5-10 min or just fail 17:43 <+bridge> [ddnet] Amazing 😄 17:43 <+bridge> [ddnet] no one plays solo on CHN anyway, you could still restart the rest already 17:44 <+bridge> [ddnet] is CHN last on the list though? 17:44 <+bridge> [ddnet] they are copied at the same time 17:44 <+bridge> [ddnet] Oh, great 17:44 <+bridge> [ddnet] all others are done 17:44 <+bridge> [ddnet] :thonk: hi 17:45 <+bridge> [ddnet] @deen btw `cat: all-locations: No such file or directory` we should probably make all our scripts cd to `~/servers` 17:45 <+bridge> [ddnet] oops, yeah 17:46 <+bridge> [ddnet] hi @TsFreddie 🌈 17:46 <+bridge> [ddnet] Btw you could have put `all-restart.sh` in terms of `execute-all.sh` instead of `restart-ddnet-on-empty.sh` 17:47 <+bridge> [ddnet] I didn't write that script, but feel free to rename 17:47 <+bridge> [ddnet] saw you guys talking about CHN servers, and I realize I have no idea what you guys are doing 17:47 <+bridge> [ddnet] so carry on 17:48 <+bridge> [ddnet] Solo servers are broken, Learath2 is reverting the change and deploying to serv ers 17:48 <+bridge> [ddnet] and CHN has slow internet to ddnet.tw main server 17:48 <+bridge> [ddnet] oof. ye, you can probably close CHN solo servers, nobody would even notice 17:49 <+bridge> [ddnet] There are no dedicated solo servers anyway 17:49 <+bridge> [ddnet] Okay that should be all the servers 17:49 <+bridge> [ddnet] now I'll join the servers they are cheating on and kill those by hand 17:49 <+bridge> [ddnet] :thonk: wait so what is broken 17:51 <+bridge> [ddnet] see #bugs, solo mode 17:51 <+bridge> [ddnet] so you can touch others 17:52 <+bridge> [ddnet] @deen since we'll have to delete the solo ranks from today anyway, I'll just restart the solo servers that are running the old version even if people are playing on them 17:54 <+bridge> [ddnet] how will you enforce that people don't change server type to solo on other running servers? 17:54 <+bridge> [ddnet] I think you have to pw-protect all, or alternatively remove solo option from votes (but that is super hacky) 17:54 <+bridge> [ddnet] (and disable /map) 17:57 <+bridge> [ddnet] @deen hm, ok Ill pw-protect old servers 17:57 <+bridge> [ddnet] Or remove all votes? 17:57 <+bridge> [ddnet] and disable sv_map 17:57 <+bridge> [ddnet] I mean most people never change map anyway 17:57 <+bridge> [ddnet] would be less painful 17:58 <+bridge> [ddnet] so sv_map 0, clearvotes 17:58 <+bridge> [ddnet] yes 17:58 <+bridge> [ddnet] what about servers that are already solo? 17:59 <+bridge> [ddnet] hard restart 17:59 <+bridge> [ddnet] or get in with gdb and set them to solo 😄 17:59 <+bridge> [ddnet] I feel bad for the people playing 😄 17:59 <+bridge> [ddnet] I guess it has to be done 17:59 <+bridge> [ddnet] Tell them you're sorry? 18:06 <+bridge> [ddnet] Oh, version.h is broken again 18:07 <+bridge> [ddnet] @deen somehow servers running 282508 show up as 15.0.5. 282508 is 15.1.3 :/ 18:10 <+bridge> [ddnet] Well apparently all servers show up as 15.0.5 bow 18:10 <+bridge> [ddnet] now* 18:10 <+bridge> [ddnet] Ah, I know why 18:10 <+bridge> [ddnet] that's my fault 18:11 <+bridge> [ddnet] cmake caches the version and overwrites the one in version.h with a define 18:11 <+bridge> [ddnet] I have the same problem locally. A clean build should help. @heinrich5991 as the cmake expert might have a better idea 18:11 <+bridge> [ddnet] Uh cmake gets it from version.h but caches it? 18:11 <+ChillerDragon> i have no eye on the ddnet proj so idk baut current issues so maybe this is known already see kill feed https://zillyhuhn.com/cs/.1602950996.png 18:11 <+bridge> [ddnet] yes 18:11 <+bridge> [ddnet] so that I can overwrite it for RCs and nightlies 18:12 <+bridge> [ddnet] theoretically cmake automatically detects when the version changes by looking at the git dir 18:12 <+bridge> [ddnet] but we have consistently managed to break it 18:12 <+bridge> [ddnet] ChillerDragon: how to reproduce? Not known to me 18:12 <+bridge> [ddnet] as a precatuion, maybe it would be good to just delete version.h (or wherever that thing is) before the build on ddnet.tw 18:12 <+bridge> [ddnet] that information is kind of crucial and hard to see when it is wrong 18:12 <+ChillerDragon> idk saw it first time @deen its probably the ppl reconnecting I can also provide the demo if it helps 18:13 <+bridge> [ddnet] ChillerDragon: if the demo reproduces it, would be great 18:13 <+ChillerDragon> k hope i dont forget after the run 18:13 <+bridge> [ddnet] @heinrich5991 but version.h is not auto-generated, it just reads the define. I guess a `touch version.h` should work 18:14 <+bridge> [ddnet] As a precaution we should always rm -rf the build dir and create a new one 18:14 <+bridge> [ddnet] I do that for the releases 18:14 <+bridge> [ddnet] I mean in build.sh 18:15 <+bridge> [ddnet] I was thinking maybe we should get rid of version.h 18:16 <+bridge> [ddnet] We could add a game/generated/version.h instead which cmake can generate 18:17 <+bridge> [ddnet] This reading from version.h sounds redundant to me 18:25 <+bridge> [ddnet] true 18:36 <+ChillerDragon> timeout code 123 last stage at Patchwork 6h time tho xd timeout any minute 18:44 <+bridge> [ddnet] Okay, cmake seems to have no way to distinguish between a variable set in the cache and a variable set on the command line 19:19 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/767074237025746994/2020-10-18_00-30-59.mp4 19:19 <+bridge> [ddnet] :nervous: really can't beat real dyncam 19:19 <+bridge> [ddnet] me playing fng with my dyncam without changing habits. still suck at aim with that thing on. 19:20 <+bridge> [ddnet] but at least i won't feel sick i guess? 19:31 <+bridge> [ddnet] @Learath2 did you delete ranks and teamranks on solo maps already? 19:32 <+bridge> [ddnet] Not yet 20:12 <+bridge> [ddnet] ^ can i somehow view the 4 surpressed clang-tidy errors? 20:12 <+bridge> [ddnet] the first one is easy to fix, are the 4 surpressed ones irrelevant or do i first get to see them when i fixed the first one? 20:15 <+bridge> [ddnet] ah 20:15 <+bridge> [ddnet] '1 warning treated as error' 20:15 <+bridge> [ddnet] maybe shouldve read the whole message 21:37 <+bridge> [ddnet] ChillerDragon: you around? 21:41 <+bridge> [ddnet] <☠BlackWolf☠> whats dis 21:41 <+bridge> [ddnet] <☠BlackWolf☠> oh 21:41 <+bridge> [ddnet] <☠BlackWolf☠> dev 21:47 <+bridge> [ddnet] @Дядя Женя what gpu u have btw 21:51 <+bridge> [ddnet] did you know that ssh reads from standard input?.... 21:56 <+bridge> [ddnet] CHL really is one potato server... 21:56 <+bridge> [ddnet] what's wrong with Chile? 21:57 <+bridge> [ddnet] it's having trouble connecting to the debian mirror took like a minute 21:57 <+bridge> [ddnet] oh well, not too bad 21:57 <+bridge> [ddnet] 😄 21:57 <+bridge> [ddnet] recently they shut down the server every day for 4 days because they thought I didn't pay, bu tI did 21:58 <+bridge> [ddnet] and then they broke the system clock 21:58 <+bridge> [ddnet] and we noticed because people had finishes before maps were released 21:58 <+bridge> [ddnet] anyway, I set up a little script to get the versions of finishes from teehistorian files on the servers 21:59 <+bridge> [ddnet] <Дядя Женя> @Jupstar ✪ gigabyte gtx 1660 ti 21:59 <+bridge> [ddnet] ty 21:59 <+bridge> [ddnet] it doesn't have a fallback to our actual teehistorian database so it can only check however much we have on the servers, but I think that should be enough 22:00 <+bridge> [ddnet] @Learath2 not sure I would do that straight on servers, might cause lags 22:00 <+bridge> [ddnet] maybe do it from our teehistorian mirrors 22:00 <+bridge> [ddnet] it's rather trivial to parse the header 22:00 <+bridge> [ddnet] it's in json 22:00 <+bridge> [ddnet] I'm still syncing back my teehistorian backup, so my server is not complete yet 😄 22:05 <+bridge> [ddnet] @deen hm, /root/.ssh/authorized_keys doesn't seem to be the same on all servers 22:05 <+bridge> [ddnet] rus's seem to be missing ddnet.tw's 22:06 <+bridge> [ddnet] chl too 22:08 <+bridge> [ddnet] rus, chl, usa, jap, sgp, zaf all seem to be missing root@ddnet.tw's key on their root so the main server can't ssh to them 22:09 <+bridge> [ddnet] With this many servers we really should look into something like ansible to deploy these so they end up identical 22:18 <+bridge> [ddnet] CHN3 can't even resolve it's debian mirror 22:18 <+bridge> [ddnet] even though it's mirror is located in china 22:22 <+bridge> [ddnet] idk who hosts `http://mirrors.cloud.aliyuncs.com` but it seems to be out of commision and chn6 and chn3 use it 22:23 <+bridge> [ddnet] oof 22:23 <+bridge> [ddnet] fix it 22:24 <+bridge> [ddnet] looks like Alibaba's 22:25 <+bridge> [ddnet] The apt settings are generated with some cloud magic I'm not familiar with, not sure if I'm allowed to even touch it 22:26 <+bridge> [ddnet] There should be other mirrors. I remember both Peking university tsinghua university host mirrors for a lot of things 22:26 <+bridge> [ddnet] There should be other mirrors. I remember both Peking university and tsinghua university host mirrors for a lot of things 22:27 <+bridge> [ddnet] https://mirrors.tuna.tsinghua.edu.cn/help/debian/ 22:27 <+bridge> [ddnet] They do 22:27 <+bridge> [ddnet] debian has a load balancer mirror in cn too, the problem is idk where to change it 😛 22:28 <+bridge> [ddnet] usually the mirror would just be used in `/etc/apt/sources.list` 22:29 <+bridge> [ddnet] learath using a linux with systemd :monkalaugh: 22:29 <+bridge> [ddnet] apt is not part of systemd yet thank god 22:29 <+bridge> [ddnet] :dabbing: 22:29 <+bridge> [ddnet] though it's the first time I've touched a system using cloud-init] 22:30 <+bridge> [ddnet] Windows guy here, I'm out 22:30 <+bridge> [ddnet] Oh right can I get a role 22:31 <+bridge> [ddnet] Which role would you like? 22:32 <+bridge> [ddnet] You are the music bot now 22:32 <+bridge> [ddnet] @Learath2 https://guix.gnu.org/ 22:32 <+bridge> [ddnet] stallman approved distro 22:32 <+bridge> [ddnet] :monkalaugh: 22:33 <+bridge> [ddnet] @Learath2 t it uses GNU Shepherd instead of the now ubiquitous systemd. 22:33 <+bridge> [ddnet] https://www.gnu.org/software/shepherd/ 22:35 <+bridge> [ddnet] Coded in scheme, lmao 22:37 <+bridge> [ddnet] made with https://www.gnu.org/software/guile/ 22:37 <+bridge> [ddnet] This is why we can't have nice things, people who can't stop creating their own languages 22:38 <+bridge> [ddnet] ??? 22:38 <+bridge> [ddnet] i dont see how this prevent nice things 22:39 <+bridge> [ddnet] Person A wants to create nice thing A-thing, A instead creates awesome-A-lang then codes A-thing in said lang. No one can read thus contribute to A-thing. So A-thing will never be nice 22:40 <+bridge> [ddnet] wrong 22:40 <+bridge> [ddnet] u assume people wont learn A lang to do said thing 22:40 <+bridge> [ddnet] That's a rather fair assumption. When is the last time you learned an entire language to contribute to something or use something? 22:41 <+bridge> [ddnet] i love learning new languages 22:41 <+bridge> [ddnet] but i have yet to find a project in a language unknown to me 22:41 <+bridge> [ddnet] <Дядя Женя> C++ for ddnet 22:41 <+bridge> [ddnet] I personally just use an alternative that is in/uses the 7 or so languages that I (and most of the programmers out there) are familiar with 22:42 <+bridge> [ddnet] atleast it looks cute 22:42 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/767125321279340574/kid-programming-a-robot.png 22:42 <+bridge> [ddnet] I wouldn't learn Guile to contribute to Shepherd, I'd use/contribute to OpenRC, minix, runit, busybox 22:42 <+bridge> [ddnet] wow discord ruins it 22:42 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/767125372890120222/unknown.png 22:42 <+bridge> [ddnet] Heh, transparencygang 22:43 <+bridge> [ddnet] @Learath2 i would learn guile to contribute to GNU shepherd so gnu stallman recognizes me as a gnu propheyt 22:43 <+bridge> [ddnet] gnu guile* 22:43 <+bridge> [ddnet] How can alibabas debian mirror die? :/ 22:54 <+bridge> [ddnet] Okay, tsinghua uni's mirror works fine 22:54 <+bridge> [ddnet] hope I didn't break anything 😛 22:56 <+bridge> [ddnet] ``` 22:56 <+bridge> [ddnet] ../src/engine/external/json-parser/json.c:402:65: runtime error: applying non-zero offset 5 to null pointer 22:56 <+bridge> [ddnet] SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior ../src/engine/external/json-parser/json.c:402:65 in 22:56 <+bridge> [ddnet] ``` 22:57 <+bridge> [ddnet] @Jupstar ✪ Should we exclude external? 22:57 <+bridge> [ddnet] but still sounds bad? 22:57 <+bridge> [ddnet] (we should probably update json.c) 22:58 <+bridge> [ddnet] i just ran it with all 22:58 <+bridge> [ddnet] maybe we pass in some bad json object? 22:58 <+bridge> [ddnet] oh or that 22:58 <+bridge> [ddnet] i dunno, but doesnt sound healthy xd, but could ofc be false positiv 22:58 <+bridge> [ddnet] i dunno, but doesnt sound healthy xd, but could ofc be false positive 22:59 <+bridge> [ddnet] I wonder what is null at L402 22:59 <+bridge> [ddnet] CServerBrowser::LoadDDNetInfoJson() 22:59 <+bridge> [ddnet] that is the caller 22:59 <+bridge> [ddnet] `(*(json_char **) &top->u.object.values) += string_length + 1;` 22:59 <+bridge> [ddnet] @Jupstar ✪ how did you figure out the caller btw? 23:00 <+bridge> [ddnet] call stack from ubsan 23:00 <+bridge> [ddnet] UBSAN_OPTIONS=print_stacktrace=1 23:02 <+bridge> [ddnet] fwiw east added the json parser 6 years ago, last commit on the json parser is from a year ago, an update might not be the worst idea 23:03 <+bridge> [ddnet] I updated it to 1.1.0 3 years ago 23:04 <+bridge> [ddnet] And there has been no new release since then 23:04 <+bridge> [ddnet] https://github.com/udp/json-parser/tags 23:05 <+bridge> [ddnet] @Jupstar ✪ can you print out the json we are passing json_parse_ex, see if it's malformed maybe? 23:05 <+bridge> [ddnet] @Jupstar ✪ https://forum.ddnet.tw/viewtopic.php?t=7003&p=68629#p68629 23:06 <+bridge> [ddnet] (but even if a string is malformatted json the parser should not run into ub) 23:06 <+bridge> [ddnet] @deen #3108 23:06 <+bridge> [ddnet] https://github.com/ddnet/ddnet/pull/3108 23:06 <+bridge> [ddnet] thats the fix 23:06 <+bridge> [ddnet] Yes, let's merge that? 23:06 <+bridge> [ddnet] Should we look into a more robust json parser? 23:06 <+bridge> [ddnet] yes, then i am finally free again 23:07 <+bridge> [ddnet] and should I downport it into a new instant release? 23:07 <+bridge> [ddnet] because next release will take a few weeks at least 23:07 <+bridge> [ddnet] idc, just do what u want 23:07 <+bridge> [ddnet] i'll remove the gl code that is not 3.3 some day 23:08 <+bridge> [ddnet] and in 17 years we just default to 3.3 23:08 <+bridge> [ddnet] when teeworlds gets 30 years old 23:09 <+bridge> [ddnet] I'd guess teeworlds will be staying with gl1 for a loong while 23:09 <+bridge> [ddnet] @Jupstar ✪ u think anyone would be around in 17 years 23:09 <+bridge> [ddnet] will* 23:09 <+bridge> [ddnet] i rq before 23:09 <+bridge> [ddnet] Of course we'll all be around and discussing about whether we should finally add accounts 23:09 <+bridge> [ddnet] btw @deen someone with intel had a crash even in gl 1.x ;D 23:09 <+bridge> [ddnet] 😄 23:09 <+bridge> [ddnet] but dont tag me then 23:10 <+bridge> [ddnet] 😄 23:10 <+bridge> [ddnet] 😦 23:10 <+bridge> [ddnet] > btw @deen someone with intel had a crash even in gl 1.x ;D 23:10 <+bridge> [ddnet] @Jupstar ✪ then whats the point 23:10 <+bridge> [ddnet] of the downgrade 23:10 <+bridge> [ddnet] How come you tell me this after I prepare the release 23:10 <+bridge> [ddnet] i dont get tagged or blamed xD 23:10 <+bridge> [ddnet] But you changed most of the graphics code 23:10 <+bridge> [ddnet] @deen ???? i tagged u 23:10 <+bridge> [ddnet] I wonder if an older client helps them not crash 23:10 <+bridge> [ddnet] yes 23:10 <+bridge> [ddnet] https://discord.com/channels/252358080522747904/757720336274948198/766714731188191282 23:11 <+bridge> [ddnet] so there is something wrong in our code, but we can't reproduce it 23:11 <+bridge> [ddnet] there is something wrong in intel xD 23:11 <+bridge> [ddnet] but we had no such crash reports for years before 23:11 <+bridge> [ddnet] i had no crash ever 23:11 <+bridge> [ddnet] atleast not one i provocated 23:11 <+bridge> [ddnet] That's the "we can't reproduce it" part 23:12 <+bridge> [ddnet] I'd kill to see a crashdump :/ 23:12 <+bridge> [ddnet] @deen maybe the issue exist since forever but we didnt have enough players with x setups to get that 23:12 <+bridge> [ddnet] Does it happen to anyone tech savvy enough to run it in a debugger? 23:12 <+bridge> [ddnet] no 23:12 <+bridge> [ddnet] its all normies 23:12 <+bridge> [ddnet] @Ryozuki no chance 23:12 <+bridge> [ddnet] @Learath2 enough to integrate the crash reporter? 23:13 <+bridge> [ddnet] @deen heh, I wish I had the time, maybe I can slowly work on it at nights 23:13 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/295908390956433410/767003311290843136/unknown.png 23:13 <+bridge> [ddnet] 23:13 <+bridge> [ddnet] this for example would be a prove for a bug OUTSIDE the gl code, 23:13 <+bridge> [ddnet] 23:13 <+bridge> [ddnet] bcs i checked and we never render such a huge quad, nor are we buffering chat text in not gl 3.3 in VRAM 23:13 <+bridge> [ddnet] Maybe we should ask someone who had the crash to give us teamviewer access? 23:13 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/295908390956433410/767003311290843136/unknown.png 23:13 <+bridge> [ddnet] 23:13 <+bridge> [ddnet] this for example would be a proof for a bug OUTSIDE the gl code, 23:13 <+bridge> [ddnet] 23:13 <+bridge> [ddnet] bcs i checked and we never render such a huge quad, nor are we buffering chat text in not gl 3.3 in VRAM 23:13 <+bridge> [ddnet] and we install a debugger? 23:13 <+bridge> [ddnet] lmao 23:13 <+bridge> [ddnet] @deen I'd be willing to help someone with it but it needs to be someone patient 23:13 <+bridge> [ddnet] the worst is most of times they dont even speak english 23:14 <+bridge> [ddnet] Oh and someone that has good internet because Visual Studio is like 20G now 23:14 <+bridge> [ddnet] :justatest: 23:14 <+bridge> [ddnet] Gotta love windows 23:14 <+bridge> [ddnet] isnt mingw smaller 23:14 <+bridge> [ddnet] cant u debug there 23:14 <+bridge> [ddnet] Does gdb work on Windows? 23:14 <+bridge> [ddnet] Oh we actually cross-compile with mingw 23:14 <+bridge> [ddnet] I should be able to use gdb yeah 23:14 <+bridge> [ddnet] yes 23:15 <+bridge> [ddnet] but we don't have symbols, so I'll have to pray the crash still happens when I build on the guys computer :d 23:15 <+bridge> [ddnet] Anyway, find me someone and I'll make it happen 23:15 <+bridge> [ddnet] Then let's make a special build with symbols for that 23:16 <+bridge> [ddnet] If we get 1 guy reporting this issue per day, there must 100 more who don't report it 23:16 <+bridge> [ddnet] Oh, btw it's probably not the driver itself crashing, driver crashes lead to bsods on windows. It's probably either our code or the userspace part of the driver 23:17 <+bridge> [ddnet] it should be evident on the crash screen btw, if the userspace part of the driver crashes windows will tell them 23:17 <+bridge> [ddnet] btw we did debug the crash 23:17 <+bridge> [ddnet] the one guy with the intel gpu and me 23:17 <+bridge> [ddnet] for like 1h 23:17 <+bridge> [ddnet] and found nothing 23:17 <+bridge> [ddnet] and when we updated his driver it instant worked 23:17 <+bridge> [ddnet] It doesn't crash under debugging? 23:18 <+bridge> [ddnet] so do you have some backtrace, where did it crash exactly? 23:18 <+bridge> [ddnet] it random places 23:18 <+bridge> [ddnet] well it just didnt continue processing 23:18 <+bridge> [ddnet] so not a real crash, true 23:18 <+bridge> [ddnet] it freezes? 23:18 <+bridge> [ddnet] it doesnt show anything anymore 23:18 <+bridge> [ddnet] but that was different from intel to intel 23:19 <+bridge> [ddnet] if you interrupt the program with a signal, where do you land? does it appear the program is still running? 23:19 <+bridge> [ddnet] yes all threads were working 23:19 <+bridge> [ddnet] no infinite loops 23:19 <+bridge> [ddnet] nothing 23:19 <+bridge> [ddnet] all threads are working as expected but nothing on the screen? 23:20 <+bridge> [ddnet] yes 23:20 <+bridge> [ddnet] it internally died 23:20 <+bridge> [ddnet] and on other intel platform probs the same behaviour caused crashes 23:20 <+bridge> [ddnet] instead of, whatever u want to call this 23:20 <+bridge> [ddnet] mh, graphics debugging I'm not really good at 23:20 <+bridge> [ddnet] its not really possible 23:20 <+bridge> [ddnet] on linux maybe with open source drivers 23:21 <+bridge> [ddnet] but never heard of any bug or crash in linux the last 2-3 years 23:21 <+bridge> [ddnet] Is there a way to make opengl verbose? 23:21 <+bridge> [ddnet] yes 23:21 <+bridge> [ddnet] already did that, but nothing really interesting 23:22 <+bridge> [ddnet] It'd be nice to have as a setting btw, something like `gfx_debug 1` we can tell people to enable so we can see it in the logs 23:22 <+bridge> [ddnet] it requires opengl 4.3 tho 23:22 <+bridge> [ddnet] xd 23:23 <+bridge> [ddnet] well atleast the driver needs to support it 23:23 <+bridge> [ddnet] I seriously hate all the graphics people 23:23 <+bridge> [ddnet] <☠BlackWolf☠> I F3 23:23 <+bridge> [ddnet] How did people debug their damn programs before 4.3 then??? 23:23 <+bridge> [ddnet] well amd and nvidia probs have tools 23:24 <+bridge> [ddnet] but no standard 23:24 <+bridge> [ddnet] @Jupstar ✪ there are extensions for opengl 1.0 tho 23:24 <+bridge> [ddnet] <☠BlackWolf☠> U fix my problem @Jupstar ✪ 23:24 <+bridge> [ddnet] <☠BlackWolf☠> :happy: 23:24 <+bridge> [ddnet] i fixed nothing XD 23:24 <+bridge> [ddnet] lmao 23:24 <+bridge> [ddnet] <☠BlackWolf☠> u did 23:24 <+bridge> [ddnet] <☠BlackWolf☠> :c 23:25 <+bridge> [ddnet] intel fixed it for u 23:25 <+bridge> [ddnet] <☠BlackWolf☠> D: 23:25 <+bridge> [ddnet] they successfully support gl 1.x 23:25 <+bridge> [ddnet] *clap* 23:25 <+bridge> [ddnet] <☠BlackWolf☠> :brownbear: 23:25 <+bridge> [ddnet] @Learath2 btw its very hard to crash opengl 23:25 <+bridge> [ddnet] bascailly only when uploading or deleting buffers 23:26 <+bridge> [ddnet] everything else is defined somewhere in the specs 23:26 <+bridge> [ddnet] all drawing and shit is not allowed to crash 23:26 <+bridge> [ddnet] how would you crash opengl anyway? it's just an api 23:26 <+bridge> [ddnet] but it says what the impl has todo 23:26 <+bridge> [ddnet] so its more a standard 23:27 <+bridge> [ddnet] anyway, i gave tw a try for modernization, failed, doesn't matter whom to blame, me or intel 23:27 <+bridge> [ddnet] we'll stick with borderfix for the next 10 years 😄 23:28 <+bridge> [ddnet] Wait, that means we need borderfix again? 23:28 <+bridge> [ddnet] well we need all gl 1.x stuff again 23:28 <+bridge> [ddnet] oh no 23:28 <+bridge> [ddnet] that was the idea of the new renderer xD 23:28 <+bridge> [ddnet] so I need to also revert all the fixed graphics? 23:29 <+bridge> [ddnet] nvidia and amd have like 90% marketshare.. so they all could easily get opengl 4.5 running 23:29 <+bridge> [ddnet] just for 0.0001% of ppl 23:29 <+bridge> [ddnet] :monkalaugh: 23:29 <+bridge> [ddnet] @Jupstar ✪ that is only for dgpu 23:29 <+bridge> [ddnet] Intel has 60-70% total gpu market share 23:29 <+bridge> [ddnet] steam charts 23:29 <+bridge> [ddnet] i doubt 23:30 <+bridge> [ddnet] everyone has intel cpus and they come with integrated gpus 23:30 <+bridge> [ddnet] yes 23:30 <+bridge> [ddnet] wow 23:30 <+bridge> [ddnet] doesnt mean its used 23:30 <+bridge> [ddnet] I have 4 computers in my household, all only with Intel 23:30 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/767137599672156170/unknown.png 23:31 <+bridge> [ddnet] @deen let's revert #3108 for now and try to debug instead? 23:31 <+bridge> [ddnet] https://github.com/ddnet/ddnet/pull/3108 23:31 <+bridge> [ddnet] Reverting all the graphics changes will be a whole mess 23:31 <+bridge> [ddnet] yeah 23:31 <+bridge> [ddnet] yeah but if it doesnt work, it doesnt work xD 23:32 <+bridge> [ddnet] and borderfix being necessary again is another mess 😛 23:32 <+bridge> [ddnet] Is the 0.7 client crashing for these people too? 23:32 <+bridge> [ddnet] I thought borderfix was independent since Teeworlds 0.7 also doesn't need borderfix and has opengl 1? 23:32 <+bridge> [ddnet] the problem is 0.7 still uses mostly 1.x code 23:32 <+bridge> [ddnet] @deen i'm aware, they use 3d textures without mipmapping 23:33 <+bridge> [ddnet] @deen go in 0.7 load multimap into editor 23:33 <+bridge> [ddnet] zoom out 23:33 <+bridge> [ddnet] 0.7 try is completly useless 23:33 <+bridge> [ddnet] if you dont create mipmaps you can also NOT use 3d textures 23:37 <+bridge> [ddnet] Are we linking to GLEW correctly? 23:37 <+bridge> [ddnet] r we using glew? 23:37 <+bridge> [ddnet] yes 23:38 <+bridge> [ddnet] @Learath2 there are actually long error messages for glew linking specifically: https://ddnet.tw/downloads/win32.log 23:45 <+bridge> [ddnet] @Jupstar ✪ are there things like validation layers for opengl? xd 23:45 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/767141152272416818/unknown.png 23:45 <+bridge> [ddnet] there is validateprogram 23:45 <+bridge> [ddnet] u can basically only crash opengl when creating binary data 23:46 <+bridge> [ddnet] u may not crash opengl but u may be doing things wrong 23:46 <+bridge> [ddnet] and not know it 23:46 <+bridge> [ddnet] or drawing too much 23:46 <+bridge> [ddnet] but we need a crash xd 23:49 <+bridge> [ddnet] eh? what is even compiling glew.c?? 23:57 <+bridge> [ddnet] @Jupstar ✪ does checking in order for `KHR_debug`, `ARB_debug_output` and `AMD_debug_output` sound about right?