Time Nick Message 00:00 rubenwardy Wtf, jekyll suddenly works for me on minetest.net? 09:56 smalllargeone I've ran into a bit of an issue compiling 5.2.0 from source on raspbian with c++ 6.3.0. The build failed for me. I've tried with trunk/master but that failed for me at the same location. I have no idea how to proceed. 09:57 p_gimeno I don't think anyone can help without specific error messages 09:57 smalllargeone fwiw i'm a user attempting to compile from source, i'm by no means a developer 09:58 smalllargeone oh, np i make a paste 09:58 smalllargeone [ 31%] Building CXX object src/CMakeFiles/minetest.dir/client/clientlauncher.cpp.o 09:58 smalllargeone c++: internal compiler error: Killed (program cc1plus) 09:58 smalllargeone Please submit a full bug report, 09:58 smalllargeone with preprocessed source if appropriate. 09:58 smalllargeone See for instructions. 09:58 smalllargeone src/CMakeFiles/minetest.dir/build.make:110: recipe for target 'src/CMakeFiles/minetest.dir/client/client.cpp.o' failed 09:58 smalllargeone make[2]: *** [src/CMakeFiles/minetest.dir/client/client.cpp.o] Error 4 09:58 smalllargeone make[2]: *** Deleting file 'src/CMakeFiles/minetest.dir/client/client.cpp.o' 09:58 smalllargeone make[2]: *** Waiting for unfinished jobs.... 09:58 smalllargeone CMakeFiles/Makefile2:253: recipe for target 'src/CMakeFiles/minetest.dir/all' failed 09:58 smalllargeone make[1]: *** [src/CMakeFiles/minetest.dir/all] Error 2 09:58 smalllargeone Makefile:149: recipe for target 'all' failed 09:58 smalllargeone make: *** [all] Error 2 09:59 p_gimeno c++: internal compiler error: Killed (program cc1plus) 09:59 sfan5 doesn't "make a paste" usually mean linking to an external site with the text content? 09:59 p_gimeno ^ 10:00 p_gimeno anyway, that internal compiler error probably means that the compiler ran out of memory 10:00 smalllargeone sorry. yes if it is preferred then i'm able to paste on a external server. In my book pasting means just that, copy-paste :-) 10:00 p_gimeno you can try to exit X window to see if that helps 10:01 smalllargeone oh, out of memory ? even if i have enabled swap ? (i am aware it's a stretch on a pi) 10:02 smalllargeone i give that a swing then 10:03 sfan5 in case you are using "make -j4" or "make -j2", use "make" instead 10:08 smalllargeone ok thanks, i'll try that as well 10:31 p_gimeno if you're using the stock X window, you can exit by using Ctrl+Alt+F1 or Ctrl+Alt+F2 to go to a terminal, then logging in and entering: sudo service lightdm stop 10:31 p_gimeno the stock desktop* 10:39 smalllargeone thx, i'll keep a note of that and will apply when necessary. for the time being the compiler is progressing... it's a slow process though :-) 10:47 smalllargeone i was thrown off guard from the error statement an sich. there is no mentioning whatsoever from the compiler about running out of memory or something alike. I would have expected a more meaningful error-report but perhaps that has more tot do with regards to my inexperience with gcc 10:48 smalllargeone never too old to learn :-) 11:13 smalllargeone .. and done. thank you p_gimeno and sfan5 !! 11:34 p_gimeno sweet, enjoy! 11:36 p_gimeno OOM can cause all sorts of weird errors, it's very hard to tackle every instance of a memory allocation in order to report it, and sometimes you can't report it because you have no memory to report it 11:37 p_gimeno that, and the fact that it was a Raspberry Pi, made memory the first suspect 15:13 nerzhul sfan5 are you around ? 15:39 sfan5 yes 16:07 nerzhul #9641 is ready for a merge sfan5 :) 16:07 ShadowBot https://github.com/minetest/minetest/issues/9641 -- Move luacheck & clang-{tidy,format} to github actions by nerzhul 16:22 rubenwardy It was a trap! 16:23 nerzhul :o ? 16:43 sfan5 merging #9712, #9537 soon 16:43 ShadowBot https://github.com/minetest/minetest/issues/9712 -- Fix alias handling of get_content_id by sfan5 16:43 ShadowBot https://github.com/minetest/minetest/issues/9537 -- Reuse object_shader for "wielditem" and "item" entity drawtypes by dcbrwn 16:43 nerzhul rubenwardy, he didn't fall in my trap :D 16:44 nerzhul i adds comments on #9712 :D 16:44 ShadowBot https://github.com/minetest/minetest/issues/9712 -- Fix alias handling of get_content_id by sfan5 16:47 sfan5 will fix 17:03 sfan5 merging #9712 soon, again 17:03 ShadowBot https://github.com/minetest/minetest/issues/9712 -- Fix alias handling of get_content_id by sfan5 17:10 sfan5 it appears github actions does not support colored output 17:10 sfan5 how sad 17:14 sfan5 it also causes wget to output progress like this: 17:14 sfan5 7150K .......... .......... .......... .......... .......... 51% 120M 1s 17:15 p_gimeno DS-minetest: [off] is for off-the-record, not for off-topic 17:40 nerzhul which core metrics can be useful to expose ? (i want to add some for a first PR). I added player count, server step time, map save time, i will try to have loaded mapblocks number 17:41 rubenwardy see Buckaroo's dashboard 17:42 nerzhul link ? 17:43 rubenwardy it's 404-ing 17:43 nerzhul hu ? 17:43 rubenwardy https://monitoring.minetest.land/d/YUpouLmWk/overview?orgId=1&refresh=5s 17:44 rubenwardy make sure to check out other boards 17:48 nerzhul okay some are to be retrieved, after that i think anybody can contribute on any engine part :) 17:48 nerzhul * to add useful metrics needed 18:05 rubenwardy nerzhul: that should be somewhere in here: https://github.com/pandorabox-io/minetest_docker/blob/master/patches/old/pandorabox.patch (its a mess :laughing: ) 18:05 rubenwardy I asked for a list of stats 18:05 nerzhul np, after that if anybody wants to help after the PR is merged i'm fine 18:06 nerzhul i'm not sure adding all metrics on the PR is needed for a first shot , but it's improvable (and i have some work time on CI to do) 18:09 nerzhul the lua vm size is obviously the better start 18:09 nerzhul stat* 18:09 nerzhul i think there is some custom implementation inside it too 18:10 rubenwardy this too : https://github.com/pandorabox-io/minetest_docker/blob/19310e7c2b9cc166273bce149582e74ccdce0ce2/patches/old/pandorabox.patch#L2117-L2201 18:10 rubenwardy GitHub 18:10 nerzhul i will need to have the build PR merged before publishing this one as it will require ci changes 18:10 rubenwardy The PR should consider how stats gathering should be done 18:10 nerzhul you can per packet counter ? 18:11 nerzhul you want* 18:11 rubenwardy I'm showing that BuckarooBanzai has done, for inspiration 18:11 nerzhul yep it can be useful, i agree, but not sure what can we find by having per packet type counters 18:12 rubenwardy some useful bottle necks I guess 18:12 rubenwardy more useful are things like chunks sent, lost reliable packets/resends, etc 19:53 nerzhul yep and perf bottleneck but remember that the lua mods can be hard to track :D