01:12 <+bridge> [ddnet] Skeith suggested that one could disallow people from finishing as a team if they lose one tee 01:12 <+bridge> [ddnet] to fix people skipping first parts with extra tees 01:12 <+bridge> [ddnet] I don't really like that, I think it should be okay to lose someone on the way and still get a team rank 01:12 <+bridge> [ddnet] uh no, bad idea 01:14 <+bridge> [ddnet] the problem is that I see no way to distinguish between "tee acidentally lost on the way" and "tee intentionally killed to skip early parts" 01:14 <+bridge> [ddnet] so unless we find a way to do that, I think it should stay possible to finish with an incomplete team 01:19 <+bridge> [ddnet] Playing with more than 2 tees will almost always be slower anyways. I believe people would mostly only use extra tees (dummies) to bypass certain parts, not actually speedrun with them. 01:19 <+bridge> [ddnet] it's undeniable that people play in larger teams for fun as well 01:19 <+bridge> [ddnet] you can obviously say that you don't want to support that 01:19 <+bridge> [ddnet] but I think that'd be bad 01:20 <+bridge> [ddnet] I think it's fine to do that, as long as everyone finishes. 01:20 <+bridge> [ddnet] what if it's on a faily map 01:20 <+bridge> [ddnet] yes, but that's unlikely with larger teams 01:20 <+bridge> [ddnet] i thought if someone dies in a team all die 01:20 <+bridge> [ddnet] only with /lock @Dr.Funmanteddy 01:20 <+bridge> [ddnet] oh ok 01:21 <+bridge> [ddnet] but then why would you get a team rank for an unlocked team? 01:21 <+bridge> [ddnet] Team ranks should be used for speedrunning, it sounds more like they're just playing the map... Idk, I just prefer the old idea of speedrunning, 2 tees doing the map as fast as possible 01:21 <+bridge> [ddnet] because it was deemed okay to lose tees on the way @Dr.Funmanteddy 01:22 <+bridge> [ddnet] @Skeith so you're saying that team ranks are only for speed runners. that's fine 01:22 <+bridge> [ddnet] well yeah you can keep the finish but the team rank its self should be for locked teams 01:22 <+bridge> [ddnet] I disagree with that statement, I think they're also for people playing in a team for fun 01:23 <+bridge> [ddnet] (I'm extrapolating that statement because you say that teams shouldn't be able to finish for a team rank if they lose someone. that seems unrealstic for larger teams playing for fun, hence my statement above) 01:23 <+bridge> [ddnet] It sounds like there should be a different category for speedrun ranks then, I really thought team ranks were the whole point of that 01:24 <+bridge> [ddnet] why a different category when speedrunners are always on top? 01:24 <+bridge> [ddnet] rank and teamrank points should be consolidated into time points, solo+dummy gives based on rank and 2p maps gives based on teamrank with no kills 01:24 <+bridge> [ddnet] if that was to be added 01:24 <+bridge> [ddnet] but i think there isnt much point to do that 01:24 <+bridge> [ddnet] because he's saying that speedrunning should not contain killing tees for speed 01:25 <+bridge> [ddnet] Because skipping parts with the use of extra tees that you kill off afterwards seems weird for speedrunning 01:25 <+bridge> [ddnet] it's also part of the game, i don't understand why you would want to restrict that 01:25 <+bridge> [ddnet] cutting is part of any race game 01:25 <+bridge> [ddnet] Well, it wasn't a part of the game initially, dummy didn't always exist 01:25 <+bridge> [ddnet] I guess you could have other players help you 01:25 <+bridge> [ddnet] i think its fine if its just dummies but like if you got 20 people to help you and kill would be silly 01:26 <+bridge> [ddnet] dummy didn't exist since the beginning of ddnet? 01:26 <+bridge> [ddnet] But I don't think anyone really did that back in the day? 01:26 <+bridge> [ddnet] I'm talking before ddnet 01:26 <+bridge> [ddnet] well, ddnet only has ranks since ddnet 01:26 <+bridge> [ddnet] not that i know, it was added shortly after 01:27 <+bridge> [ddnet] ``` 01:27 <+bridge> [ddnet] commit 3dce4c94899f95c8b9ca52bd25fd6a2a52e10b62 01:27 <+bridge> [ddnet] Author: def 01:27 <+bridge> [ddnet] Date: Sat Apr 26 20:29:42 2014 +0200 01:27 <+bridge> [ddnet] 01:27 <+bridge> [ddnet] Add dummy (Thanks to //toast and Monsta) 01:27 <+bridge> [ddnet] ``` 01:27 <+bridge> [ddnet] ddnet created on july 2013 01:27 <+bridge> [ddnet] But speedrunning in the past didn't make use of them, is what I'm saying. Sure it exists now and you could view it as a mechanic, but I don't think a majority of speedrunners enjoy that you need to do it 01:35 <+bridge> [ddnet] hmm, codebrowser lacks a scroll bar for history panel 01:35 <+bridge> [ddnet] I think it would at least be a quality of life improvement if future maps didn't allow it. At least for speedrunners. 01:36 <+bridge> [ddnet] do you have any example of map being cut with tees being ditched? 01:37 <+bridge> [ddnet] Water Lily 01:37 <+bridge> [ddnet] My rank was beat by Cireme, he didn't even know that certain skips existed until he watched me and Shinoa play it, he was surprised 01:37 <+bridge> [ddnet] Skipping parts gave that much of a time save 01:38 <+bridge> [ddnet] Skipping parts with the use of dummies gave that much of a time save 01:39 <+bridge> [ddnet] This is the skip he didn't do in his run, but still beat our ranks, due to using dummies 01:39 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/990038482007240744/unknown.png 01:39 <+bridge> [ddnet] It saves a lot of time 01:40 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/990038660210630676/unknown.png 01:40 <+bridge> [ddnet] You skip the long section at the bottom by putting 2 tees through the first 2 parts quickly 01:43 <+bridge> [ddnet] I'm sure there are maps that have bigger time saves using this method 01:45 <+bridge> [ddnet] Shows one for me on Firefox 01:46 <+bridge> [ddnet] doesn't for me 01:46 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/990040332660011018/unknown.png 01:47 <+bridge> [ddnet] firefox 101.0.1 01:47 <+bridge> [ddnet] ubuntu 01:47 <+bridge> [ddnet] @Cireme Do you have any input on the topic? 01:49 <+bridge> [ddnet] Respond in #general I guess 07:15 <+bridge> [ddnet] > But speedrunning in the past didn't make use of them, is what I'm saying. 07:15 <+bridge> [ddnet] You're stating this as a fact. Can you provide any reasoning or evidence? 07:16 <+bridge> [ddnet] Nope, just experience 07:16 <+bridge> [ddnet] Doesn't mean much I guess 08:57 <+bridge> [ddnet] oh i thought it was updated version 09:04 <+bridge> [ddnet] im pretty sure the android build is broken anyway 09:04 <+bridge> [ddnet] Does anyone have tried to build for android recently? 09:07 <+bridge> [ddnet] yes works for me 09:09 <+bridge> [ddnet] oh then fine πŸ™‚ 09:09 <+bridge> [ddnet] did/can you upload the apk on ddnet.tw/download ? 09:11 <+bridge> [ddnet] but why, u cannot relly use it, no controls 09:11 <+bridge> [ddnet] u can use it with keyboard and mouse xD 09:12 <+bridge> [ddnet] ah but compiling SDL is currently broken 09:12 <+bridge> [ddnet] ``` 09:12 <+bridge> [ddnet] sdl/src/audio/openslES/SDL_openslES.c:163:25: error: mixing declarations and code is incompatible with standards before C99 [-Werror,-Wdeclaration-after-statement] 09:12 <+bridge> [ddnet] const SLInterfaceID ids[1] = { SL_IID_VOLUME }; 09:12 <+bridge> [ddnet] ``` 09:12 <+bridge> [ddnet] lmao, i kinda feel like Sdl is alwys broken for me 09:15 <+bridge> [ddnet] what the hell is that error? 09:15 <+bridge> [ddnet] they use local variables but dont delcare them at the start of the function i guess 09:15 <+bridge> [ddnet] in c99 u put your locals at the start 09:16 <+bridge> [ddnet] lol, but why c99 anyway? can't you change standard? 09:17 <+bridge> [ddnet] yes i'll add it i guess. maybe cmake is doing this, i read that arch also had problems with that πŸ˜„ 09:18 <+bridge> [ddnet] -std=c11 i think 09:26 <+bridge> [ddnet] hmm, maybe indeed cmake adds its own set of flags 09:26 <+bridge> [ddnet] 09:28 <+bridge> [ddnet] https://www.phoronix.com/scan.php?page=news_item&px=Arch-Linux-Bizarre-Zstd 09:28 <+bridge> [ddnet] btw this was not a performance hit actually 09:28 <+bridge> [ddnet] it was just that c99 has no timespec structure or smth like that 09:31 <+bridge> [ddnet] already posted by ryozuki i think 09:31 <+bridge> [ddnet] but yeah flags can have weird interaction with optimization pass order as well 09:32 <+bridge> [ddnet] ah no, posted by you last week πŸ˜„ 09:32 <+bridge> [ddnet] https://discord.com/channels/252358080522747904/293493549758939136/987940120818749460 09:32 <+bridge> [ddnet] yes but its funny that its kinda related 09:49 <+bridge> [ddnet] ``` 09:49 <+bridge> [ddnet] sdl/src/render/opengles/SDL_render_gles.c:26:10: fatal error: 'SDL_opengles.h' file not found 09:49 <+bridge> [ddnet] #include "SDL_opengles.h" 09:49 <+bridge> [ddnet] ``` 09:49 <+bridge> [ddnet] I dunno, smth is really broken in SDL repo xd 09:51 <+bridge> [ddnet] how big is the android studio dependency? 09:52 <+bridge> [ddnet] 1-2gb 09:52 <+bridge> [ddnet] i remember for a school project having to download like 20GB of their ide+compiler 09:52 <+bridge> [ddnet] maybe if u compile urself xD 09:52 <+bridge> [ddnet] it was on windows though 09:53 <+bridge> [ddnet] maybe i choosed default installation which came with simulator/emulator as well 09:55 <+bridge> [ddnet] even then it should not be 20gb xD 09:55 <+bridge> [ddnet] maybe the emulator img is that huge 09:58 <+bridge> [ddnet] well, i ended up deleting it as soon as the project was over 09:58 <+bridge> [ddnet] android studio was struggling to run on my poor pc with 4gb ram 09:59 <+bridge> [ddnet] i had to buy another stick 09:59 <+bridge> [ddnet] which was actually the only good thing to come out of that project πŸ˜„ 10:12 <+bridge> [ddnet] @Chairn gradle only support java 17 yet 10:12 <+bridge> [ddnet] maybe thats a problem for u 10:12 <+bridge> [ddnet] https://github.com/gradle/gradle/issues/20372 10:12 <+bridge> [ddnet] 10:12 <+bridge> [ddnet] if u tried with java 19 10:16 <+bridge> [ddnet] why DEV=ON ? 10:17 <+bridge> [ddnet] what 10:18 <+bridge> [ddnet] in the cmake call in cmake_android.sh 10:18 <+bridge> [ddnet] i dunno 10:19 <+bridge> [ddnet] back then it probs was not related to checksum 10:20 <+bridge> [ddnet] and then its easier to use debug & release client 10:20 <+bridge> [ddnet] does it overwrite other stuff too? 10:20 <+bridge> [ddnet] i kinda dislike all this complexity anyway 10:21 <+bridge> [ddnet] it keeps debug symbol, i don't think we need them for android 10:23 <+bridge> [ddnet] they are stripped by gradle on release 10:24 <+bridge> [ddnet] oh 11:47 <+bridge> [ddnet] that's c89, c99 is the one that allowed mixing 11:47 <+bridge> [ddnet] more annoying is that it uses a border instead of an outline and changes the line breaks all the time when i hover over one element πŸ˜„ 11:47 <+bridge> [ddnet] yeah true misread it 11:47 <+bridge> [ddnet] before c99 πŸ˜„ 13:05 <+bridge> [ddnet] @heinrich5991 do you still scrape for new maps? 15:46 <+bridge> [ddnet] ``` 15:46 <+bridge> [ddnet] [2022-06-25 16:43:24][opengl]: Version string: OpenGL ES 3.2 Mesa 22.1.1 15:46 <+bridge> [ddnet] /home/arda/ddnet/src/engine/client/backend_sdl.h:234:60: runtime error: load of value 190, which is not a valid value for type 'bool' 15:46 <+bridge> [ddnet] SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior /home/arda/ddnet/src/engine/client/backend_sdl.h:234:60 in 15:46 <+bridge> [ddnet] ``` 15:46 <+bridge> [ddnet] @Not Keks i get this when i use gles backend 15:49 <+bridge> [ddnet] uninitialized data 15:50 <+bridge> [ddnet] what's funny with uninitialized bool is that they can be both true and false πŸ˜„ 15:51 <+bridge> [ddnet] schrΓΆdinger's boolean? 15:51 <+ChillerDragon> xd 16:09 <+bridge> [ddnet] ty 16:11 <+bridge> [ddnet] bruh you are fast :D 17:02 <+bridge> [ddnet] @ChillerDragon why you didn't use timeout utility rather than wait_for_fifo in integration_test.sh ? 17:02 <+bridge> [ddnet] currently, script lives forever if server crash 17:09 <+bridge> [ddnet] I wonder how @ ChillerDragon is displayed in IRC 17:17 <+bridge> [ddnet] he's using matricks 17:20 <+bridge> [ddnet] If he could he would use pen and paper πŸ˜› 17:20 <+bridge> [ddnet] he uses both matrix and IRC 17:32 <+bridge> [ddnet] @deen shouldn't this assert rather than message report ? 17:32 <+bridge> [ddnet] https://github.com/ddnet/ddnet/blob/master/src/engine/server/databases/sqlite.cpp#L127 17:32 <+bridge> [ddnet] 17:36 <+bridge> [ddnet] shouldnt it be compile time instead 17:36 <+bridge> [ddnet] ok dunno if we ship sqlite ourself 17:38 <+bridge> [ddnet] well, it does pass integration test with my outdated version 18:28 <+ChillerDragon> @Chairn you mean `expect` or what? No particular reason i tend to just crack down stuff my self instead of working with other tools. It just easier for me to understand 18:28 <+ChillerDragon> and it just works :D i have horrible expirience with `expect` and see no reason to add another dependency if bashs native fifo check works so well 18:29 <+ChillerDragon> https://zillyhuhn.com/cs/.1656174542.png 18:29 <+ChillerDragon> ^ this is how the ping looks like on irc @c0d3d3v 18:30 <+bridge> [ddnet] and this is how it looks like in matrix https://zillyhuhn.com/cs/.1656174581.png 18:34 <+bridge> [ddnet] no i mean timeout https://man7.org/linux/man-pages/man1/timeout.1.html 18:35 <+bridge> [ddnet] how do i use valgrind.supp file? 19:32 <+bridge> [ddnet] i cant see my mouse cursor when teeworlds is open unless i hide the application (osx 10.15.4, ddnet 16.1) 19:35 <+bridge> [ddnet] it wasnt like this in ddnet 15.5 19:55 <+bridge> [ddnet] did u try changing display mode or renderer 19:55 <+bridge> [ddnet] in graphics tab 19:56 <+bridge> [ddnet] update OS 20:15 <+bridge> [ddnet] When noby has tech issues 20:20 <+bridge> [ddnet] oh changing gl version fixed it also giving 3-4k fps ty 20:37 <+bridge> [ddnet] @Not Keks can you cancel #5504 ? 20:40 <+bridge> [ddnet] thx 21:15 <+ChillerDragon> @Chairn ah nice im def gonna use that in some of my projects but dont see how its useful in the CI 21:19 <+bridge> [ddnet] script can lives forever if server crashes 21:19 <+bridge> [ddnet] with timeout 20 it will last 20 seconds max 21:22 <+bridge> [ddnet] oO 21:23 <+bridge> [ddnet] is that safe ? 21:25 <+bridge> [ddnet] no wtf 21:25 <+bridge> [ddnet] guess he wanted to write if 21:25 <+bridge> [ddnet] ah 21:25 <+bridge> [ddnet] no 21:25 <+bridge> [ddnet] no just no xd 21:26 <+bridge> [ddnet] but maybe the remove entity makes problems 21:26 <+bridge> [ddnet] maybe its safer to copy all objects to a temporary list first 21:26 <+bridge> [ddnet] this code is such a mess xD 21:27 <+bridge> [ddnet] yup 21:27 <+bridge> [ddnet] im still trying to attach gdb to one of the client 21:27 <+bridge> [ddnet] both crash 21:27 <+bridge> [ddnet] i cant fucking have interactivity with gdb through the script... 21:28 <+bridge> [ddnet] I always get MemoryLeak messages from UBSan whenever i close the client or crash :d 21:28 <+bridge> [ddnet] i got them too, you can ignore them 21:29 <+bridge> [ddnet] best part is that these leaks are not from the client but from other libs 21:29 <+bridge> [ddnet] This is why source is not visible like others? 21:30 <+bridge> [ddnet] Or says Direct/Indirect leaks 21:31 <+bridge> [ddnet] yeah i'd say it is safe 21:31 <+bridge> [ddnet] just really unintuitive code 21:40 <+bridge> [ddnet] @Chairn yeah i'm also able to reproduce on integration test, but this time not wayland related 21:40 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/990340644511813673/log 21:42 <+bridge> [ddnet] this fucking bash is creating a subprocess of itself and i get pid of bash and not ddnet... 21:42 <+bridge> [ddnet] check the integration test script 21:42 <+bridge> [ddnet] i solved that part πŸ˜„ 21:43 <+bridge> [ddnet] im trying to attach debugger to running instance of ddnet in the script 21:43 <+bridge> [ddnet] but i got only bash subprocess' pid with $! 21:44 <+bridge> [ddnet] thatβ€˜s weird, sounds like it should work 21:44 <+bridge> [ddnet] if we invoke the process with & and no nohup or other stuff 21:44 <+bridge> [ddnet] nah, it uses `||` and i realised i modified the wrong line 21:45 <+bridge> [ddnet] and obviously it doesnt crash anymore 21:45 <+bridge> [ddnet] looks like its very timing dependent 21:49 <+bridge> [ddnet] nice heisenbug we got here 21:51 <+bridge> [ddnet] where should i set the breakpoint? 21:51 <+bridge> [ddnet] destructors aren't called when it doesn't crash... 21:55 <+bridge> [ddnet] looks like stack corruption of vtable corruption 21:56 <+bridge> [ddnet] address executed is on 2 instructions 22:13 <+bridge> [ddnet] i think i get libpng related messages this time :d 22:13 <+bridge> [ddnet] `[2022-06-25 23:12:38][libpng]: warning for file "downloadedskins/Shadow.21178.tmp": iCCP: known incorrect sRGB profile` this when i use Shadow skin 22:15 <+bridge> [ddnet] if libpng says so, then it is so 22:15 <+bridge> [ddnet] when i close the client: 22:15 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/990349499157319740/message 22:15 <+bridge> [ddnet] when i close the client: 22:15 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/990349628513874000/message.txt 22:16 <+bridge> [ddnet] nah this one 22:16 <+bridge> [ddnet] i guess graphics is deallocated already 22:19 <+bridge> [ddnet] why does that code even run in a different thread? 22:19 <+bridge> [ddnet] when was this added 22:19 <+bridge> [ddnet] our storage is not thread safe 22:35 <+bridge> [ddnet] create an issue on github pls and ping heinrich or smth that the jobs should be cleaned before any kind of destruction happens 22:35 <+bridge> [ddnet] 22:35 <+bridge> [ddnet] I dunno if storage is save for saving but since the code exists for ever i guess its fine^^ 22:35 <+bridge> [ddnet] ah no its actually loading data 22:36 <+bridge> [ddnet] then it probs not fine, dunno, maybe for just opening files it is 22:36 <+bridge> [ddnet] guess aslong readfile from storage isnt used or snmth like that 22:44 <+bridge> [ddnet] i give up 22:44 <+bridge> [ddnet] its beyond my capacities