12:31 < ddnet-commits> [ddnet] Ryozuki opened pull request #666: Fix some warnings (master...pr_fix_warnings) https://git.io/vyK8v 12:33 < Ryozuki> 666 12:33 < Ryozuki> my commit is hell 13:45 <@heinrich5991> Ryozuki: can you add what warnings you fixed? e.g. print out what it said before your commits 13:46 <@heinrich5991> Ryozuki: maybe also say what build system you used (bam or cmake) 14:03 < Ryozuki> done i guess 14:25 <@heinrich5991> Ryozuki: could you try whether moving the #define up to the very top also fixes the issue? 14:25 <@heinrich5991> https://stackoverflow.com/questions/12871513/how-to-get-rid-of-win32-winnt-not-defined-warning 14:30 < Learath2> heinrich5991: could you give chat.cpp:L288 a look? I don't quite understand SearchType but it feels like the if and the else if there need to be reversed 14:30 < Learath2> atleast for a single letter autocomplete 14:31 < Learath2> actually even if more then that if its at the beginning it should be prioritized 14:34 <@heinrich5991> wonderful code w/o comments 14:35 < Learath2> yeah and its not even descriptive 14:39 <@heinrich5991> Learath2: not really motivated to understand that code :/ 14:39 < Learath2> exactly :P I just hoped you knew how it worked 14:41 <@heinrich5991> git blame blames oy 14:42 < Ryozuki> i move #define _WIN32_WINNT 0x0501 to the top? heinrich5991 14:42 <@heinrich5991> yes 14:42 <@heinrich5991> Learath2: https://github.com/ddnet/ddnet/commit/e1ef01f 14:43 < Learath2> look at that madness 14:43 < Ryozuki> yep it fixes it 14:43 < Learath2> i guess the report on the autocomplete issue is invalid 14:44 <@heinrich5991> or it is indeed the wrong order 14:44 <@heinrich5991> you could try it out :) 14:45 < Ryozuki> mm how should i name this commit? xD 14:46 <@heinrich5991> move #define _WIN32_WINNT to the top of the source file 14:46 < Ryozuki> ok done 14:46 <@heinrich5991> Ryozuki: "top" and you missed server.cpp 14:46 <@heinrich5991> (it's still below #include ) 14:47 < Ryozuki> i think it makes no diference 14:47 < Ryozuki> anyway ill push that to 14:48 <@heinrich5991> yea, but if mingw decides to include windows.h in some day, we're covered :) 15:16 < ddnet-commits> [ddnet] heinrich5991 closed pull request #666: Fix some warnings (Maybe fixes #643) (master...pr_fix_warnings) https://git.io/vyK8v 15:18 < ddnet-commits> [ddnet] heinrich5991 closed pull request #664: Just color the heart preserving team information. Fixes #606 (master...dd_pr_friends2) https://git.io/vyrOA 15:18 <@heinrich5991> Ryozuki: could you try out https://github.com/ddnet/ddnet/pull/652? (just whether it compiles) 15:18 < Ryozuki> later, im working on the telegun right now 15:18 <@heinrich5991> kk 15:21 < Ryozuki> btw u know 15:21 < Ryozuki> travisci? 15:21 < Ryozuki> https://travis-ci.org/ 15:23 <@heinrich5991> I think deen set up circleci before 15:24 < Learath2> then apparently east broke it 15:24 <@heinrich5991> someoneā„¢ should fix it 15:25 < redix> heinrich5991: cmake uses the dynamic runtime library on windows while bam uses the dynamic one. Is this intended? 15:26 <@heinrich5991> of which library? 15:26 <@heinrich5991> and are you sure? :) 15:26 <@heinrich5991> and no, it wouldn't be intended if that is indeed the case 15:28 < Learath2> hmm the circleci hook seems to be working fine, should check on that side deen 15:30 < redix> the c runtime library. msvcrXXX.dll and so on. bam adds the /MT or /MTd flag by default (https://github.com/matricks/bam/blob/master/src/driver_cl.lua#L25) for whatever reason 15:33 <@heinrich5991> wait 15:33 <@heinrich5991> redix: you wrote that both cmake and bam use the dynamic runtime library. which one doesn't? 15:34 < redix> bam uses the static one due to the /MT flag. cmake uses the dynamic one 15:35 <@heinrich5991> ah 15:36 < redix> -static-libgcc and -static-libstdc++ might also be added for mingw 15:45 <@heinrich5991> I guess we can add MT or MTd on windows. you can make a PR if you want, otherwise I'll look into that later 15:47 <@heinrich5991> redix: ^ 16:01 < Ryozuki> someone here is good drawing? 16:17 <@heinrich5991> not me 16:18 < Ryozuki> xD 16:33 < ddnet-commits> [ddnet] Redix opened pull request #667: CMake improvements (master...pc_cmake) https://git.io/vyK2M 16:48 < ddnet-commits> [ddnet] heinrich5991 pushed 1 new commit to master: https://git.io/vyKa3 16:48 < ddnet-commits> ddnet/master d91cca7 heinrich5991: Partially revert warning-fixing, it caused warnings on other systems 16:54 < Ryozuki> :o 16:54 < Ryozuki> what warnings caused? heinrich5991 16:54 <@heinrich5991> the same 16:55 < Ryozuki> ?? 16:55 <@heinrich5991> it seems width and height are unsigned on some platforms 16:55 < Ryozuki> ah 16:55 < Ryozuki> mm 16:55 < Ryozuki> u can put if windows blah 16:55 <@heinrich5991> nah, that's even uglier 16:55 < Ryozuki> xD 16:55 <@heinrich5991> you can do casts maybe 16:55 <@heinrich5991> that should fix it as well 17:02 < necropotame> Ryozuki, you need a graphic for the portal gun, or you need something else? 17:03 < Ryozuki> i need "teleport gun on/off" and a graphic for untelegun 17:03 < Ryozuki> untelegun it's like unhookable 17:03 < Ryozuki> u can't teleport there 17:03 < Ryozuki> and it's placed inside the unhookable/hookable 17:03 < Ryozuki> the on and of should be 2 tiles 17:04 < Ryozuki> like hammer on and hammer of 17:04 < necropotame> Ok 17:04 < necropotame> I will try tomorrow 17:05 < necropotame> Btw, you should use a new graphic for the gun instead of reuse the vanilla Ines. They are already all taken for something else 17:05 < necropotame> Or course, vanilla need a fallback 17:06 < Ryozuki> it's not a single gun 17:06 < Ryozuki> it works with the gun 17:06 < Ryozuki> and the rocket 17:08 < necropotame> I see 17:08 < Ryozuki> if u want u can come and test it 17:08 < necropotame> I can't 17:10 < necropotame> Is it related to this? https://forum.ddnet.tw/viewtopic.php?f=48&t=5000 17:11 < Ryozuki> ye 17:11 < Ryozuki> but it doesn't do all what he says 17:21 < Learath2> heinrich5991: I need to base a branch off of your pr_ddnet_extendprotocol2, how to do that without making it horrible to merge later on? 17:21 <@heinrich5991> base it off that 17:21 <@heinrich5991> check my branch out, then do your changes 17:21 < Learath2> but its way too old and doesn't have cmake 17:21 <@heinrich5991> ah 17:21 <@heinrich5991> mh 17:22 <@heinrich5991> Learath2: I'll rebase it, that works, right? 17:22 < Learath2> sure but then you'll be force pushing a public branch :P 17:23 <@heinrich5991> to a feature branch, oh noes 17:24 < Ryozuki> necropotame, btw tiles used are 96 97 98 17:24 < Learath2> thus the ":P" 17:24 <@heinrich5991> Learath2: rebased 17:24 <@heinrich5991> Learath2: btw, do you use makefiles for compilation? 17:25 < Learath2> ? 17:25 <@heinrich5991> Learath2: how do you compile after doing cmake ..? 17:25 <@heinrich5991> make? 17:25 < Learath2> make 17:25 < Learath2> i should give ninja a go 17:26 <@heinrich5991> yes :) 17:26 <@heinrich5991> it seems to be faster, at least for builds where there is nothing to do 17:26 <@heinrich5991> or not much 17:28 < Learath2> whenever i get a new package i have the incredible desire to update everything while at it. not really good if you are using gentoo :P 17:28 <@heinrich5991> solution: don't use gentoo :P 17:29 < Learath2> no thank you i don't want systemd-* 17:30 <@heinrich5991> yea, then gentoo is appropiate maybe 17:31 <@heinrich5991> fstd_ would probably nudge you in the direction of *bsd now :) 17:32 < Learath2> unfortunately i still need my graphics card, so gentoo is right where i want to be 17:33 < Learath2> heinrich5991: the ninja build does seem to be faster but it doesnt look nearly as cool as the make build :P 17:33 <@heinrich5991> less terminal scrolling by -- you're less of a hacker now 17:34 < fstd_> tfw ninja build 17:34 < Learath2> the color cues were nice 17:34 < fstd_> isn't that reserved for rockstar brogrammers? 17:35 <@heinrich5991> ;) 17:38 < ddnet-commits> [ddnet] Learath2 opened pull request #668: Fix array subscript below array bounds (master...dd_pr_chatoops) https://git.io/vyKwX 18:34 < Learath2> heinrich5991 || anyone else: https://paste.pr0.tips/bI?c something feels off about this that i can't put my finger on 18:50 <@heinrich5991> Learath2: what kind of "off" do you have in mind? 18:50 < Learath2> in the OOP sense 18:50 < Learath2> i rewrote that file like 6 times 18:51 < Learath2> I ended up using a bool to signify if a CAccountManager instance belongs to the server 18:51 < Learath2> but a Kex operation between two CPeer's are agnostic of whether the instance belongs to the server or nor 18:52 < Learath2> s/nor/not/ 18:53 < ddnet-commits> [ddnet] heinrich5991 pushed 6 new commits to master: https://git.io/vyKi3 18:53 < ddnet-commits> ddnet/master 8b41829 Redix: Fix typo in CMakeLists.txt 18:53 < ddnet-commits> ddnet/master 3911504 Redix: Minor cmake fixes... 18:53 < ddnet-commits> ddnet/master dfdd08a Redix: Disable some warnings on win64 18:54 <@heinrich5991> redix: could you check whether Learath2's pull request compiles on windows? https://github.com/ddnet/ddnet/pull/652 18:55 <@heinrich5991> Learath2: so it's more of a CAuthConnection than a CAuthManager class? 18:56 <@heinrich5991> SeesionKey btw 18:56 < Learath2> well cant really call it a connection since it doesnt create any new connections between client and server 18:57 <@heinrich5991> dunno, I think on a higher level, it does 18:58 < Learath2> well sure could call it CAuthConnection 18:59 < Learath2> i used to not have any manager but instead had two CPeer objects for each CServer::CClient 18:59 <@heinrich5991> how does that class even interact with the environment? 19:00 < Learath2> heinrich5991: which one? CAccountManager? 19:00 <@heinrich5991> I only see setters and a PerformKex method, no way to extract info 19:00 <@heinrich5991> I can only see CAccountManager 19:00 < Learath2> oh it doesnt interact much yet, i didnt want to bother designing the rest until i got a nice interface to perform the Kex 19:01 < Learath2> its going to have a void Encrypt(const unsigned char *pData, int Size); 19:01 <@heinrich5991> ah 19:02 < Learath2> and probably a int State() { return m_State; }; 19:02 <@heinrich5991> char *pOut, int OutSize 19:02 < Learath2> unsigned char *pDst, int DstSize would be more teeworlds esque 19:03 <@heinrich5991> k 19:03 < redix> heinrich5991: the pull request compiles :) 19:03 < Learath2> so you think the way i distinguish server from client is k? 19:03 < Learath2> redix: any warnings? 19:03 <@heinrich5991> any newly introduced ones 19:03 <@heinrich5991> it does have some warnings IIRC 19:04 < redix> only the ones that i also get with master branch 19:04 <@heinrich5991> redix: thanks for trying :) 19:05 < redix> but i didn't try with cmake (which has stricter rules for warnings on windows) 19:05 < ddnet-commits> [ddnet] heinrich5991 pushed 3 new commits to master: https://git.io/vyKih 19:05 < ddnet-commits> ddnet/master 9319a82 Learath2: Add /invite for locked teams. Fixes #647 19:05 < ddnet-commits> ddnet/master 0b0783f Learath2: Fix styling issues, use uint for bitmask 19:05 < ddnet-commits> ddnet/master ccb9db9 heinrich5991: Merge pull request #652 from Learath2/dd_pr_invite... 19:05 <@heinrich5991> it's ok 19:05 < Learath2> while we are mostly here can we talk about this timer thing... 19:06 < Learath2> just want to know if we are keeping it 19:06 <@heinrich5991> I'd like to keep it 19:06 < Ryozuki> btw, why the timer was shown in broadcast? i still don't know why 19:07 < Learath2> Ryozuki: you mean in the past or with necropotame's patch? 19:07 < Learath2> I'd like to put it behind a config option for mappers 19:07 < Ryozuki> in the past 19:07 < Ryozuki> oh 19:07 < Ryozuki> no, in the patch 19:08 <@heinrich5991> before allowing mappers to decide, I'd rather autodetect 19:08 < Learath2> we were (ab)using the game timer for race time which has the effect of breaking animation sync between clients 19:09 < Ryozuki> ah ok 19:09 < Learath2> then necropotame patched it to use the warmup timer, which allows us to have synchronized animations 19:09 < Learath2> heinrich5991: I'd rather do it by hand then use some autodetect 19:09 <@heinrich5991> that creates too much manual work 19:09 < Learath2> afaik animations relative to racetime work just fine 19:10 <@heinrich5991> we can't have a config option for everything, especially not for some niche stuff like vnailla client that not every mapper is aware of 19:10 <@heinrich5991> they do create a jump in animation when passing the start line 19:11 < Learath2> heinrich5991: If he is aware of moving freeze zones and synchronized animations you can bet they know about sv_force_sync or a tile in the map that forces sync 19:11 <@heinrich5991> oh you want animations synced to race time in ddnet client as well? 19:11 < Learath2> s/he/they/ 19:11 < Learath2> yeah ofc 19:11 <@heinrich5991> it creates that jump of animations, I don't like it 19:11 < Learath2> I want old maps to behave as they did for everyone 19:12 <@heinrich5991> it was ok for the hack 19:12 <@heinrich5991> but now we have a better solution 19:12 < Learath2> i mean this is 50x the hack that the old one was 19:13 <@heinrich5991> previously we had bad animations for everyone 19:13 < Learath2> and i don't really get how animations are more important then a nice timer in a race mod 19:13 <@heinrich5991> now we have bad timer for vanilla 19:13 < Ryozuki> why don't leave vanilla away c: 19:14 < Learath2> bad is really subjective, I'm yet to hear anyone complain about bad animations but i've heard plenty of people complain about the new timer 19:14 <@heinrich5991> yea, because the timer was a change 19:14 <@heinrich5991> people always dislike change 19:14 < necropotame> Auto detection of moving zones is easy, btw 19:14 <@heinrich5991> yes 19:14 <@heinrich5991> just check whether the map uses teeuniverse features 19:15 < Learath2> I didn't want the new behaviour for old maps, in that sense auto detection is meh at best 19:15 < Ryozuki> people always dislike changes, but if they don't see a client release in some time, they start crying :P 19:15 < necropotame> But keep in mind that creating animated zones will be easy using TeeUniverse, so you may expect to see them more often once I've implemented it 19:15 < Learath2> maybe i should have made the autoupdate more visable 19:15 < necropotame> lol 19:16 < Learath2> and we just destroy the f1 console of vanilla clients 19:16 <@heinrich5991> ? 19:16 <@heinrich5991> no 19:16 <@heinrich5991> that's a ddnet "feature" 19:16 < Ryozuki> xD 19:16 < Learath2> s/vanilla/unupdated/ 19:16 <@heinrich5991> which should maybe have rate-limiting 19:16 <@heinrich5991> because we're not the only server using broadcast for continuous information 19:17 < necropotame> I've a solution: put some sort of version tag in the map (automatically added by the editor). Old maps will be detected by that way 19:17 < Learath2> necropotame: i'm fine with that but that introduces bad animations back 19:17 < Learath2> for the old maps i mean 19:17 < necropotame> Infclass player hates you <3 19:18 <@heinrich5991> you can also turn off the broadcast thing IIRC 19:18 < necropotame> Learath2, you said that you want the same behavior for them 19:19 < Learath2> necropotame: I do heinrich5991 doesn't im guessing you don't either 19:19 < necropotame> Btw, all of this happens because ddnet don't use branches :/ I'm still for an animation branch. Or should I host it? 19:19 < ddnet-commits> [ddnet] heinrich5991 closed pull request #601: Implementation of Portable TeeUniverse Maps (master...master) https://git.io/vyKXv 19:20 < necropotame> Well, I can understand that old maps should stay as they was 19:20 <@heinrich5991> the problem is independent from that 19:20 <@heinrich5991> we'd have this problem sooner or later 19:20 < redix> It works well for the latest ddnet version... so why even bother with people who use old clients? And for vanilla the broadcast should be fine 19:20 < Ryozuki> ^ 19:20 < necropotame> Because some peoples must use sdl1 19:21 < Ryozuki> why? 19:21 <@heinrich5991> because we haven't fixed the problems yet 19:21 <@heinrich5991> they can however merge the code if they want 19:21 < Ryozuki> go for it then (?) 19:21 < redix> then better fix the problems with sdl instead of changing the timer again... 19:21 < necropotame> Now I remember why I said SDL2 is not in Ubuntu 14.04. It's in the repository, but it didn't work well withb DDNet (input errors) 19:22 <@heinrich5991> ah, good starting point 19:22 <@heinrich5991> what kind of issues appear there? 19:22 < Learath2> w/e we can keep the new timer I guess, I'll try patching up some of the sdl2 problems 19:23 < necropotame> heinrich5991, I'm lost with the animation implementation. Just tell me what I've to do 19:23 <@heinrich5991> necropotame: ? 19:23 < necropotame> I don't remember exactly. I think Henningstone experienced the same 19:24 <@heinrich5991> Learath2: cool! they've been lying around for some time 19:24 < necropotame> It was something like you can't use properly the mouse wheel, but I'm not sure 19:24 <@heinrich5991> necropotame: I don't get this message: [19:22:59] heinrich5991, I'm lost with the animation implementation. Just tell me what I've to do 19:26 < necropotame> @heinrich5991, oh you created a branch :P thanks 19:27 < necropotame> I've just seen the IRC message, so I thought you just close it 19:32 < Ryozuki> btw 19:32 < Ryozuki> https://cdn.discordapp.com/attachments/255333887679397888/290552501831467008/Screenshot.png 19:32 < Ryozuki> the heart is weired 19:32 < Ryozuki> (latest git version) 19:33 <@heinrich5991> lol, the screenshot is buggy :) 19:33 < necropotame> heinrich5991, there is several point to discuss concerning TeeUniverse (like what zones can be animated, how to adapt the editor, ...). Maybe a dedicated label can be useful 19:33 <@heinrich5991> from a lot of different frames 19:34 < necropotame> Or a github project, but from what i've seen from it, it's kinda useless 19:34 < Ryozuki> well its timakro who send me it xd 19:34 < Ryozuki> he says the heart was well positioned before 19:34 < Ryozuki> hi 19:36 < timakro> https://misc.timakro.de/heartbug.png 19:36 < timakro> note the two floating hearts 19:36 < Ryozuki> oh, in middle of scren 19:36 <@heinrich5991> Learath2: ^ 19:36 <@heinrich5991> that's your fault :P 19:37 <@heinrich5991> and mine, because I accepted the PR :) 19:37 < timakro> i liked it better how it looked before, just color should change 19:37 < timakro> little bit hacky code wont hurt this mess ^^ 19:37 <@heinrich5991> that attitude is how it became a mess :) 19:55 < ddnet-commits> [ddnet] heinrich5991 opened pull request #671: Completed the refactor (master...pr_ddnet_cytral_refactor) https://git.io/vyKMs 19:55 < ddnet-commits> [ddnet] heinrich5991 closed pull request #514: Refactor source (master...Refactor) https://git.io/vKshj 19:58 < ddnet-commits> [ddnet] heinrich5991 closed pull request #581: cl_overlay_entities_show_unused (master...hideUnused) https://git.io/vX4N4 21:39 < Learath2> heinrich5991: is it the extra space im putting there? 21:39 < Learath2> and why are they floating up there 21:39 <@heinrich5991> I guess that's the real question ;) 21:40 < Learath2> i mean they are rendered exactly as the lines themselves are rendered 21:40 < Learath2> timakro: hao2 reproduce? 21:45 < Learath2> i can't seem to get them to weirdly offset nor make them float 21:45 < Learath2> rather you seem to have an extra space 21:47 < timakro> i use thisbind t "+show_chat; chat all" 21:47 < timakro> maybe thats the reason 21:48 < Learath2> timakro: isnt that the default t bind? 21:48 < Learath2> also how do you prepend the client id there? 21:49 < timakro> oh maybe they made it the default 21:49 < necropotame> Btw, why DDNet needs a lot of binds? There should be at least more entries in settings/input, or some new buttons in the ESC menu 21:49 < Ryozuki> "also how do you prepend the client id there?" it's a client option 21:49 < timakro> cl_show_ids 1 21:49 < Learath2> oh need to initialize m_Friend so it doesnt happen to system messages 22:18 < Learath2> chat.cpp:L716 is basically str_format("%2d: ") but even tho %2d and that construct generate exactly the same string they are rendered differently 22:19 < Learath2> %2d ends up in the middle 22:32 < ddnet-commits> [ddnet] Learath2 opened pull request #672: Forgot m_Friend isn't initialized, also align the messages (master...dd_pr_hearts) https://git.io/vyK5s 22:32 < Learath2> heinrich5991: not sure about this one ^^ 22:39 < Learath2> also circle ci seems to be working fine just not for PRs anymore? 22:40 < Learath2> Which is because we now use the primitive webhook instead of the service integration 22:44 <@heinrich5991> Learath2: currently looking into fixing, unless you're also doing that right now 22:45 < Learath2> nah i wanted to ask here if there was a particular reason we were using the primitive webhook before touching anything 22:45 <@heinrich5991> let's find out 22:50 < Learath2> timakro: i couldn't really reproduce the floating heart 22:55 < ddnet-commits> [ddnet] heinrich5991 opened pull request #673: Test CircleCI (master...pr_ddnet_test_circleci) https://git.io/vyKdZ 22:56 < Learath2> \o/ 22:57 < ddnet-commits> [ddnet] heinrich5991 closed pull request #673: Test CircleCI (master...pr_ddnet_test_circleci) https://git.io/vyKdZ 22:57 <@heinrich5991> works :) 22:57 <@heinrich5991> needed to tick a box on CircleCI's website 22:57 <@heinrich5991> "test pull requests" 22:58 < Learath2> oh :P 22:58 <@heinrich5991> Learath2: btw, do you know what would be great -- actual tests :) 22:58 <@heinrich5991> but I have no experience writing tests tbh 22:59 < Learath2> every time i look at circleci i think that, then i notice idk wat to test except for system.c which is stable anyways 22:59 <@heinrich5991> very-high level tests: client code can connect to server code, e.g. 22:59 < Learath2> anyways onto finishing my 6th iteration of the account system :P 23:00 <@heinrich5991> :) 23:00 <@heinrich5991> not sure if tw code as is allows for such testing though 23:00 < Learath2> well cant even have a headless client 23:00 < Learath2> so cant even abuse that 23:28 <@deen> start an X server, test that correct pixels show up at expected spots 23:41 <@heinrich5991> deen: serious? :)