Time Nick Message 01:03 stormchaser3000 hi 01:04 stormchaser3000 i seem to have an issue 01:04 stormchaser3000 after an update system update on arch 01:04 stormchaser3000 minetest segfaults when every i try to run it 01:29 est stormchaser3000, this is a known bug, see #3130 01:29 ShadowBot https://github.com/minetest/minetest/issues/3130 -- segfault on startup for irrlicht 1.8.2 01:48 est ah 01:49 est "Irrlicht log: Warning: The library version of the Irrlicht Engine (1.8.2) does not match the version the application was compiled with (1.8.1). This may cause problems." 01:49 est this seems to be an arch only bug 01:49 est stormchaser3000, can you try to remove minetest, and re-install it? 01:49 est it will trigger a rebuild of minetest 01:49 stormchaser3000 est: i am trying to downgrade irrlicht 01:50 stormchaser3000 and then i will rebuild minetest 01:50 est irrlicht 1.8.2 and 1.8.1 aren't binary compatible, as it seems 01:50 est and that has been the problem, I guess 01:51 est because I can't reproduce it 02:06 est Tesseract, are you around? 02:59 est Tesseract, nvm 03:16 stormchaser3000 etc: ok i downgraded irrlicht 03:16 stormchaser3000 and 03:16 stormchaser3000 still segfault 03:16 stormchaser3000 even after recompilinbg 03:16 stormchaser3000 and reinstalling 03:19 est this bug is really weird 03:19 est I've tried to build irrlicht 1.8.2 myself 03:19 est and there was no problem 04:02 est31 anybody around to approve this commit https://github.com/est31/minetest/commit/5c4769674e0887cae6476f1386b32b5a51ab0e59 04:03 est31 its fairly obvious, and tested by me 04:03 est31 so therefore I'll push it in 20 minutes if nobody minds 07:39 rubenwardy #3140 corrected to remove commits added in error 07:39 ShadowBot https://github.com/minetest/minetest/issues/3140 -- Cache some settings by rubenwardy 07:39 rubenwardy nrzkt, it was obviously a git mistake 07:43 nrzkt no it's your mistack because you add commits to the PR branch :) 07:44 rubenwardy ... 07:45 rubenwardy that sentence means the same as "it was obviously a mistake with git" 07:46 rubenwardy callgrind says that 60% of dedicated_server_loop is in ScopeProfiler::ScopeProfiler and ScopeProfiler::~ScopeProfiler 07:46 rubenwardy Is that accurate? 07:47 hmmmm i think you might be misunderstanding callgrind output. 07:47 hmmmm that does not sound right 07:48 rubenwardy https://cdn.pbrd.co/images/miqj8g0.png 07:48 rubenwardy yeah, I thought I would be 07:49 rubenwardy it's possible I'm also not sampling for long enough. That sample was for 30 ticks (the client packet graph pinged every few seconds, I assumed that that was sending of packets by server at end of tick) 07:51 rubenwardy Actually, it may have been 10, I forget 07:52 hmmmm yeah. 07:52 hmmmm i think that might be it. 07:52 hmmmm that screenshot is basically saying that 35% of execution time in the server is being used up by what is essentially a map lookup and an addition 07:53 hmmmm we know how slow std::map lookups are already - they're not that bad 07:53 rubenwardy if you're talking about ProfilerAVG, thats actually 35% of 56% as its relative to the parent 07:54 hmmmm erm 07:54 hmmmm are you sure that's relative to the parent? 07:54 hmmmm 55.88% breaks down into 14.02% and 34.48% 07:55 rubenwardy the parent function may take some? I enabled the "relative to parent" setting" 07:55 hmmmm that much? 07:55 rubenwardy yeah, it's doubtful 07:55 hmmmm the only work being done in ~ScopeProfiler is delete m_timer 07:56 hmmmm oddly that does not show up as a call to operator delete in the call graph.. 07:56 rubenwardy I also missed server::asyncRunStep or whatever, that takes about 20x what dedicated_server_Step does 07:57 nrzkt callgrind said std::map are the many calls, but not the time, don't misunderstand callgrind. call number != call time 07:58 rubenwardy is there a way to make a server run for X ticks then terminate? 07:59 hmmmm no, you'd have to add that yourself 07:59 VanessaE I dunno about X number of *ticks* but what about minetest.after() and call the shutdown command from there? 08:00 hmmmm yeah ^ that's a decent solution for what you want to do 08:02 VanessaE use an on_globalstep to count the ticks 08:02 VanessaE (if you need to know a specific amount) 08:02 rubenwardy that's what I was thinking 08:03 VanessaE er register_globalstep, whatever 08:08 est31 is sth like #3142 a good idea? 08:08 ShadowBot https://github.com/minetest/minetest/issues/3142 -- Let travis check for git format mistakes by est31 08:08 est31 it'll print errors like https://travis-ci.org/minetest/minetest/jobs/78532719#L713 08:13 est31 perhaps nrzkt could add a jenkins job which only checks this, to better separate it from travis 08:13 est31 jenkins/format_check-nrz 08:14 nrzkt check which format ? 08:14 nrzkt ah i see 08:14 nrzkt the git connector doesn't check that, it's a little bit difficult to do 08:15 est31 that pr has a shell script 08:15 est31 it checks for whitespace errors in the diff 08:16 est31 and for properly formatted commit messages 08:19 nrzkt hmmm maybe i should download patch from github and analyze it, but i need some time to look and i don't guarantee this atm 09:49 Megaf is the irrlight related bug fixed already? 09:52 nrzkt no 09:55 Megaf hm, is it Irrlicht fault or MT fault? 09:55 Megaf I mean, can we fix it? 09:56 nrzkt irrlicht 09:56 nrzkt then no 09:57 nrzkt irrlicht packaging on arch has memory problems, if you look at gdb in debug mode your read memory outside program... i have a read on a std::map which should be empty but has more than 10 billions entries 10:01 Megaf So is that only on the Arch version? 10:02 Megaf I have been using the latest Irrlicht trunk from svn on my server for some time now and never had any problem with it 10:02 Megaf I'm going to make some local tests here anyway 10:08 Megaf But I did stop updating the server a couple of days ago when I read about these issues 10:21 Megaf Ok, I will test MT Server Stable-0.4 and Master against Irrlicht 1.8 and SVN http://i.imgur.com/PIbN6qV.png 10:24 nrzkt you can, but that doesn't change anything, i already test it 10:25 Megaf Couldn't this cause problems? 10:25 Megaf /dev/shm/MTtest/MinetestServer-0.4-IRR-1.8/src/lua/src/loadlib.c:76:21: warning: ISO C forbids conversion of object pointer to function pointer type [-Wpedantic] 10:25 Megaf lua_CFunction f = (lua_CFunction)dlsym(lib, sym); 10:41 est31 The according issue is #3130 10:41 ShadowBot https://github.com/minetest/minetest/issues/3130 -- segfault on startup for irrlicht 1.8.2 on arch linux 10:41 est31 nrzkt, you say you get the error even if you rebuild? 10:41 nrzkt yes 10:41 est31 e.g. not with the arch linux minetest package? 10:41 nrzkt archlinux minetest package crashed 10:41 nrzkt and a built package crash 10:42 nrzkt but not same stacktrace 10:42 nrzkt minetestserver binary also crashed, with another stacktrace 10:42 est31 and what happens if you download irrlicht 1.8.2 source manually, and build it, and then use that? 10:42 nrzkt i don't try it. But rollbacking to 1.8.1 fix the problem 10:42 est31 because that works for me 10:43 est31 I'm not on arch however, as ubuntu user i'll get 1.8.2 later 10:43 nrzkt as you said maybe it's related to irrlicht packaging on arch 10:43 Megaf est31: so, I will test either mt 0.4 git and master with irrlicht 1.8 svn and irrlicht dev svn 10:43 Megaf compiling all 4 MTs right now 10:44 Megaf server build 10:44 est31 what I'm wondering, why does the server fail 10:44 est31 I mean it doesnt link nowhere, does it? 10:44 est31 it just uses the #defines? 10:45 Megaf I think I will not have any issues though, Debian Jessie here. I suspect nrzkt is Arch related, perhaps kernel related? nrzkt can you try with another kernel? 10:45 nrzkt stop saying stupid things 10:45 nrzkt kernel hasn't been updated on my distron only the irrlicht package 10:46 nrzkt three days ago, irrlicht 1.8.1, works, two days ago, only irrlicht 1.8.2 upgrade crash 10:46 nrzkt if it was kernel memory allocation problem i think every arch user will have problems lol 10:46 nrzkt and all linux distros on this kernel version. 10:46 nrzkt and not only minetest 10:47 nrzkt i thought it was glibc related, but not, i tried the 3 last versions of glibc and only minetest causes problem 10:47 est31 yea 10:47 est31 kernel is highly unlikely 10:47 nrzkt i don't try another irrlicth dependant problem 10:47 nrzkt i think we should find another arch package with irrlicht lib to be sure :) 10:47 est31 nrzkt, is the stacktrace the same every time you start? 10:48 est31 also does it occur when you start the mainmenu, or when you join a game? 10:48 est31 when does it happen for the server? 10:48 est31 when you start it? 10:50 nrzkt yes, exactly same backtrace at each launch on a specific MT version 10:50 nrzkt and i have EXACTLY same backtrace as the issue on github, on the minetest mentionned version 10:50 est31 okay 10:51 est31 can you give the line number inside ClientActiveObject::registerType 10:51 est31 also I wonder what happens if run with valgrind memcheck 10:53 * est31 tries to run minetest with irrlicht 1.8.2 with valgrind 10:53 nrzkt i do it est31 10:53 nrzkt and valgrind said: WTF this shit, i cannot do anything, stop 10:53 nrzkt xD 10:54 est31 have you disabled luajit? 10:54 nrzkt it's disabled on my build 10:54 nrzkt but i tried both 10:56 nrzkt i go to lunch, see you 10:56 est31 good idea 10:57 Megaf interesting, MT compiles faster with Irrlicht 1.8 than Irrlicht Dev 11:00 Megaf est31: I can confirm nrzkt crash on Debian 11:00 Megaf kind of, MT Master with Irrlicht 1.8 SVN did not crash, all other versions crashed 11:01 est31 ? 11:01 est31 which versions again? 11:04 Megaf est31: ok, I will give the paste of all four tests 11:05 Megaf First test. Minetest Git Master with Irrlicht 1.8 SVN Dev. https://paste.debian.net/plain/310200 11:06 Megaf Second test. Minetest Git Master with Irrlicht SVN Dev https://paste.debian.net/plain/310203 11:07 est31 hah 11:07 waressearcher2 is anyone here ? 11:07 est31 thats not the crash Megaf 11:07 est31 its another crash 11:07 est31 which happens if you try to run two servers on the same port 11:07 waressearcher2 can I ask a question ? 11:07 est31 waressearcher2, yes 11:07 Megaf ah, sorry about that 11:08 Megaf let me fix that 11:08 est31 waressearcher2, yes 11:08 est31 dont ask to ask just ask 11:09 waressearcher2 can I ask about minecraft ? 11:09 waressearcher2 do you guys think its a good thing that Mircrosoft bought Minecraft ? 11:10 Megaf est31: I can not confirm nrzkt crash, tried different combinations of MT and Irrlight and none of them cause crash 11:10 Megaf waressearcher2: most of us just don't care. 11:10 Megaf waressearcher2: and #Minetest is a better place to talk about that 11:11 waressearcher2 I never player minecraft on PC just a few hours on smartphone so I don't care much either 11:12 waressearcher2 so the main goal of minetest is to clone minecraft or impletemt stuff that minecraft doesn't have or its just a hobby project for you developers ? 11:14 Megaf They are separate and diferent games. Just the same style. 11:14 Megaf again, #Minetest is a better place to talk about that 11:14 waressearcher2 I once saw video of "shaders" mod for minecraft, it looks really gorgeous does minetest support something like those shaders ? 11:14 waressearcher2 Megaf: I think I ask "developing" related questions, didn't I ? 11:14 Megaf Now you did 11:15 Megaf We have shaders yes. And they look pretty good and getting better 11:15 est31 yea, the question about the goal is indeed on-topic 11:16 est31 minetest is a voxel game like minecraft yes 11:16 est31 and both games share other parts too 11:16 Megaf waressearcher2: so, there are several game genre, like FPS, Strategy, Voxel, racing... 11:16 est31 but minetest is no strict 1:1 clone of it 11:16 est31 tetris 11:16 Megaf MT and MC are Voxel games, that's what they have in common 11:17 Megaf they are written in different languages, use different libraries and one is closed source and other is open 11:17 est31 minetest is far more generic than minecraft 11:17 est31 you can have much more blocks here 11:17 Megaf Minetest is a blank canvas that allows the user to do whatever one ones 11:17 est31 and if you want, you can have quite different blocks than the default 11:21 Megaf est31: so, back to the crash, I can't reproduce it either, so I presume is libc related? Or kernel? It's not caused by Irrlicht nor LuaJIT 11:21 Megaf And it is not MT fault either 11:21 est31 Megaf, seems to be the case 11:22 est31 it would be great if an arch user could try with their self compiled irrlicht 11:22 Megaf libc 2.19 here and gcc 4.9.2 11:22 Megaf est31: wait, no need to compile Irrlich to build server 11:22 est31 first an irrlicht from the PKGBUILD, self built, then one manually self built 11:23 est31 hrmmm 11:23 est31 then its a header bug? 11:23 est31 very interesting 11:23 Megaf nrzkt: can you try a server build with cmake flags like cmake ../ -DRUN_IN_PLACE=1 -DBUILD_CLIENT=0 -DBUILD_SERVER=1 -DIRRLICHT_LIBRARY=/run/shm/MTtest/Deps/irrlicht-svn/ ? 11:24 Megaf don't compile Irrlicht, just download it and point cmake to it 11:24 waressearcher2 it uses irrlich engine right ? is it possible to change "irrlich" to "darkplaces" or to change engine entire game have to be remastered ? 11:25 Megaf waressearcher2: everything is possible 11:25 Megaf nrzkt: svn checkout svn://svn.code.sf.net/p/irrlicht/code/branches/releases/1.8 irrlicht-1.8 11:25 Megaf waressearcher2: but it would require a lot of work 11:27 Megaf almost a complete rewrite :P 11:27 waressearcher2 does minetest uses some physics engine ? 11:28 Megaf hm, its complicated 11:28 waressearcher2 I assume its very simple internal physics engine 11:28 waressearcher2 just to calculate collision with blocks 11:29 est31 yea 11:29 Megaf we have colisions and player movement physics I think 11:29 est31 honestly we only have one single method 11:29 Megaf nothing fancy 11:29 est31 that does physics in the sense of "mechanics" 11:29 est31 ofc we simulate other physical effects 11:29 est31 like sunlight 11:30 est31 or gravity for water 11:30 est31 these are handled outside 11:30 est31 but the real mechanical stuff is inside a single huge method 11:31 est31 the only place where we have physics is for player and mob movement 11:31 est31 well, sand and gravel can fall down 11:31 Megaf but it's just colision physics 11:31 T4im no relativistic physics :( 11:31 waressearcher2 I mean newtonian mechanics 11:31 Megaf est31: we have physics for dropped items too 11:32 * T4im wasn't serious about that 11:32 Megaf waressearcher2: that's fancy, I said we have nothing fancy 11:32 waressearcher2 what about when arrow hits the player ? 11:32 T4im handled by the mod 11:32 Megaf We dont have arrows :P 11:33 rubenwardy at a guess, the arrow looks around itself whilst moving to look for a player. If it finds a playing in that radius, it assumes it hits. 11:35 T4im isn't that something you should rather solve by a particle than an entity, rubenwardy? 11:36 T4im :) 11:36 T4im technic used that once for the mining laser -> ended in stuck entities that weren't removable, not even by /clearobjects iirc 11:41 * Megaf waits for nrzkt 11:45 Megaf est31: nrzkt bug has been reported on Arch https://bugs.archlinux.org/task/46176?string=minetest&project=0&type%5B0%5D=&sev%5B0%5D=&pri%5B0%5D=&due%5B0%5D=&reported%5B0%5D=&cat%5B0%5D=&status%5B0%5D=open&percent%5B0%5D=&opened=&dev=&closed=&duedatefrom=&duedateto=&changedfrom=&changedto=&openedfrom=&openedto=&closedfrom=&closedto= 11:46 rubenwardy doing grep -nr "TODO" src/* is interesting 11:47 Megaf rubenwardy: Im afraid of the output of that 11:48 Megaf whats -n by the way? 11:48 rubenwardy line number 11:52 rubenwardy also grep -nr "#if 0" src/* 11:52 rubenwardy and grep -nr "FIXME" src/* 11:54 T4im those greps people do over sources against swearing can be interesting too 11:54 est31 Megaf, I know 11:55 rubenwardy / Version 19 is fucked up 11:55 rubenwardy thank you, IRC 11:55 rubenwardy // Version 19 is fucked up 11:55 T4im only 3 hits on that one, though the danish translation one might have to be checked... not that that actually appears on danish clients 11:56 est31 https://github.com/minetest/minetest/issues/3130#issuecomment-137411743 11:56 est31 interesting 11:56 T4im ... and isn't actual acceptable danish 11:56 est31 this is an irrlicht bug 11:57 Megaf est31: iti s just weird that it works on Debian 11:57 est31 Megaf, yeah 11:58 Megaf Ok, I will download the 1.8.2 zip form Irrlichts website 11:58 est31 ermm, with what have you done the checks? 11:59 est31 I have done all checks with that zip 11:59 Megaf est31: they said the server build crashes, so I got Irrlicht using SVN 12:00 Megaf got the latest 1.8 branch and the latest dev branch 12:00 est31 ok 12:00 Megaf none of them crashed 12:00 est31 where is it said? 12:03 Megaf What do you mean? 12:03 est31 est31: they said the server build crashes 12:03 est31 can you give an url 12:03 Megaf oh 12:04 est31 I wonder because if I recall right, they only provide windows binaries, no? 12:04 Megaf https://github.com/minetest/minetest/issues/3130#issuecomment-136857003 12:04 Megaf "Minetestserver has a different backtrace but always same across versions" 12:04 Megaf I presume that means the server crashes as well? 12:05 est31 yes 12:05 Megaf anyway, I will compile the server and then the client and test that too 12:06 est31 Megaf, all irrlicht examples fail for suxxnet42 as well 12:06 est31 Megaf, this means this is a clear irrlicht bug 12:06 Megaf well, I cant compile all examples here, a lot of dependencies to be solved 12:06 Megaf est31: did you try any example? 12:07 est31 Megaf, there is no point if I cant repro the crash 12:07 est31 but no 12:07 Megaf I want to troll Arch bug I know I should not 12:08 Megaf s/bug/but 12:08 est31 arch linux users do an important job: they betatest software 12:09 est31 until it came to ubuntu, all serious bugs are fixed 12:09 Megaf I have years of experience with Arch Linux, I had been using it back on version 0.7.2, before it was even rolling release 12:09 est31 and all non serious bugs are fixed as well, but dont get backkported to ubuntu 12:09 Megaf yep, Arch is not for stability, is for beta testing :P 12:11 Megaf well, when I finish the testing on this machine I will go to another Debian machine that is running Debian testing, it might be closer to Arch and perhaps I can reproduce the bug there 12:13 Megaf est31: ok, tested MT Master against Irrlicht 1.8.2 zip and it works just fine 12:14 Megaf a server build 12:14 Megaf I'm compiling Irrlicht now and will test a client build 12:24 Megaf Irrlich 1.8.2 demo works too 12:24 Megaf not Irrlicht bug est31, Arch bug 12:31 est31 Megaf, agreed 12:31 est31 arch irrlicht bug 12:31 est31 not arch minetest bug 12:34 waressearcher2 so where does irrlicht ends and minetest begins ? 12:35 est31 waressearcher2, you know what a library is? 12:35 est31 as a programmming term 12:36 waressearcher2 I know how to link a library yes 12:36 est31 well, irrlicht is a library 12:37 est31 it takes care of rendering the stuff minetest tells it to render 12:39 T4im you mentioned an idtech2 engine earlier... darkplaces was it? that's not just a library, that's a complete engine (the equivalent would "minetest" here, including irrlicht) 12:39 waressearcher2 isn't minetest developed enought as platform for modding ? 12:40 Megaf yep 12:40 Megaf minetest itself is a game engine that uses Irrlicht to render stuff 12:40 Megaf you can make a subgame for minetest 12:40 rubenwardy "game engine" is pretty meaningless, really 12:41 waressearcher2 I mean if you take "emacs" its development in "C" language stopped long time ago now it developed only in "lisp" so shouldn't minetest development stop where it is now and except bugfixes and all other stuff should be written in lua ? or there are lot features to add and they can't be added using "lua" ? 12:42 rubenwardy No, more c++ work is needed 12:42 rubenwardy Lua is really only good for adding new blocks 12:42 rubenwardy and maybe a chat command 12:44 waressearcher2 I believe its pain to add more and more stuff using incremental development, usually when games are created there first goes creation of design document with planning ahead all stuff in advance and only as a last step goes actuall programming, I mean if you want to add feature it can take you to review lots of code and sometimes remaster tons of 12:44 waressearcher2 lines of code 12:45 waressearcher2 so its not the case with minetest ? 12:46 Megaf minetest is a rollercoaster or whatever you spell that 12:46 rubenwardy Minetest wasn't planned very much, it was a personal project 12:46 Megaf minetest is just like the Universe, it just happened 12:46 rubenwardy You don't need to rewrite tons of lines of code, it depends on the feature 12:46 Megaf with a powerful force behind it 12:54 T4im btw, was output coloring now canceled or deferred after the last PR was closed? 12:54 est31 deferred 12:54 rubenwardy chat color? 12:55 rubenwardy I can't believe that we still don't have a simple feature like that, come on guys 12:55 T4im general output, rubenwardy; I'm also very interested in terminal output for coloring test results 12:55 rubenwardy meh about that 12:55 T4im I'm losing overview about what fails and what passes 12:55 T4im but chat color too, yes 12:55 T4im (was afaik the same pr, wasn't it?) 12:56 est31 rubenwardy, gcc had no color output until their 5.0 version 12:56 est31 but now they have 12:56 rubenwardy #1587 12:56 ShadowBot https://github.com/minetest/minetest/issues/1587 -- Colour codes in chat and console 12:56 rubenwardy I don't care about color in the console 12:58 T4im you are mod developer, aren't you? shakedown is for you: https://github.com/t4im/shakedown you might want to start caring about color in console now, too :P 12:59 rubenwardy I literally just say that 12:59 rubenwardy * saw 12:59 rubenwardy very cool 12:59 rubenwardy I'm also a Minetest contributor 12:59 T4im then: https://github.com/t4im/coretest 13:07 T4im apropos, est31: you guys talked about wanting to take over coretest; is that still wanted? I mean it would probably make sense, once the whole thing is sufficently well runnable for you (that's why I splitted it out, in the first place) 13:07 T4im I mean hmmm is right, that it makes sense 13:08 est31 T4im, I guess so 13:08 est31 its good to have a lua testing framework for minetest 14:35 waressearcher2 what is "lua testing framework" ? 14:35 T4im lua is the scripting language minetest uses for modding 14:38 T4im the testing framework provides a way to define tests to spot faults/bugs in these mods; and the included smoketest already checks for common known pitfalls with these, so you don't have to implement these tests yourself 14:39 T4im think starting the computer and smoke comes out.. no need to test any further ;) 14:46 Megaf_ T4im: life should come with one of those 14:49 T4im It often comes with two of those. 14:50 Megaf_ I had only one, its several thousands kilometers away now =/ 14:52 T4im sorry to hear that :/ 15:50 rubenwardy paramat: https://www.youtube.com/watch?v=t_eHBqVYa8A 15:50 rubenwardy it was you that tamed MT fire, right? 16:38 Robert_Zenz I'd like to change to change the signature of check_player_privs, from check_player_privs(name, privs) to check_player_privs(player_or_name, ...), with the three dots being vararg for a list of strings. 16:39 Robert_Zenz The reason for this would because it would become a little easier to use and I can do that in a backwards compatible manner...is there interest in this? 16:39 T4im the question would then be: is it OR or AND chained? 16:40 Robert_Zenz T4im, I would not change the behavior, only how the parameters can be handed to it. 16:42 T4im ah, it's a table atm, didn't remember that x) 16:42 Robert_Zenz T4im, yeah, with varargs you could do minetest.check_player_privs(player, "shout", "fly", "something"). 18:31 est31 I'd welcome a player_or_name change, but I'm not sure if giving away flexibility (e.g. to have an or ordered list later on) is worth sparing the two chars { } 18:32 rubenwardy you could always use the {} methos 18:32 rubenwardy to do or 18:33 T4im or table.pack(...) ;) 18:38 Robert_Zenz est31, depends what you want to do later on with the list. I would have of course kept the compatibility. 18:39 Robert_Zenz est31, which would mean that I'd check the first value of ... if it is a table, and if yes use that (same behavior as currently). 18:44 Megaf_ nrzkt: [03.09.15-19:44:15] Megaf_: svenstaro pushed a new $pkgrel, have you tried it already? 18:44 Megaf_ nrzkt: the bug should be fixed I think 18:44 Megaf_ ^ est31 18:45 nrzkt Megaf_: not synced in my current repo: community/irrlicht 1.8.1-2 -> 1.8.2-1 18:45 nrzkt always the 1.8.2-1 18:45 Megaf_ nrzkt: are you on #ArchLinux ? 18:45 nrzkt no i'm busy :p 18:46 nrzkt maybe my repo will be synced tomorrow 18:48 est31 Robert_Zenz, well, if you check all args for strings, its fine 18:49 CraigyDavi Any feedback on #3143 ? 18:49 ShadowBot https://github.com/minetest/minetest/issues/3143 -- Add CONTRIBUTING.md by CraigyDavi 18:49 Robert_Zenz est31, not exactly all args, just the first if it is a table, to keep compatibility with the old behavior. 18:50 rubenwardy CraigyDavi, no need to explain how to use Git 18:50 est31 Robert_Zenz, well, I agree to the general idea, now its just about the implementation :) 18:51 est31 rubenwardy is right, a link to a github help page is enough 18:51 Robert_Zenz est31, then I'll whip up a pull request so you see what exactly I mean. :) 18:51 Robert_Zenz Different question, was it ever considered to document the API with something like LDoc? 18:51 rubenwardy just say that you should make pr on their own branch on your personal minetest repo 18:51 rubenwardy but phrased better 18:52 est31 Robert_Zenz, what's so bad about the current format? 18:52 CraigyDavi rubenwardy, I see what you mean, I'll change that 18:53 est31 CraigyDavi, also no need to explain all this subsystem maintainer thingy 18:54 est31 its not lived 18:54 CraigyDavi So subsystems aren't in use anymore? 18:54 est31 yes, they only exist as issue labels right now 18:54 est31 and if Realbadangel has a fight with blockmen 18:55 Robert_Zenz est31, I was just curious if it ever was considered. 18:55 est31 CraigyDavi, and this list of criteria a pr has to fulfill is seen very critical by me as well. 18:55 est31 CraigyDavi, in order to confirm 5. you usually have to read most of the code 18:56 est31 CraigyDavi, I dont know we can do this if some dev sais this eases PR reviewing for them 18:57 est31 well, sorry, you havent added the "who can confirm this and that" 18:57 est31 so nvm that list can stay 18:58 est31 but generally I think that contributing.md should only be targeted at people new to the project 18:58 est31 not at core devs 18:59 est31 but yeah it can help for people new to know what PRs have to fulfill in order to be merged 18:59 est31 or that "two have to +1" rule 18:59 CraigyDavi Ok, I'll make some changes 19:05 Megaf_ nrzkt: I'm talkign to the guy who makes the package of irrlicht on Arch. sveenstaro, the bug has to do with gcc 5.2 19:05 Megaf_ est31: it works for us because even Debian Stretch is on gcc 4.9 19:06 est31 4.9.2 19:07 svenstaro why hello there 19:08 est31 hi! 19:09 svenstaro anyway I need an arch tester 19:10 svenstaro just test new irrlicht -2 with minetest for me 19:10 est31 nrzkt, ^ 19:10 est31 no i'm busy :p 19:10 est31 maybe my repo will be synced tomorrow 19:10 est31 :( 19:11 svenstaro eh ok, in that case I'm gonna nap 19:11 svenstaro post results into that bug 19:11 svenstaro pretty sure irrlicht doesn't test on gcc 5.2 19:13 Megaf_ Thanks svenstaro, nrzkt is our Arch user and Minetest contributor 19:13 T4im est31: needs_fallback_font? (https://github.com/minetest/minetest/blob/master/src/fontengine.cpp#L226) 19:13 T4im "yes" for unicode necessary? 19:14 T4im or.. something more complicated? (because german is set to "no" despite umlauts) 19:15 est31 dunno 19:32 CraigyDavi Updated #3143 \o/ 19:32 ShadowBot https://github.com/minetest/minetest/issues/3143 -- Add CONTRIBUTING.md by CraigyDavi