Time  Nick         Message
02:38 Fritigern    Megaf: If you want to know if it's safe to use the dev version or Irrlicht, perhaps you should just try it, and rollback if it makes things go kablooey
06:13 RealBadAngel https://github.com/minetest/minetest/pull/1871
06:13 RealBadAngel i just made a pull with extruded plants and torches for testing
08:14 kilbith      RealBadAngel: extruded meshes deserves to be in the settings GUI
09:35 Zeno`        Does mapgen5 need to be reverted? Apparently it's causing memory leaks?
09:36 Zeno`        s/reverted/fixed
13:57 Megaf        Hi all
13:57 Megaf        so, I have this issue. http://paste.debian.net/plain/133255
13:59 Pennth       looks like you need to allocate more ram to the container where you run the server.
13:59 Megaf        I see Mineteset server using lots of ram without need
14:00 Megaf        I see a memory leak, since it was using 512 MB of ram + 512 MB of swap
14:05 Pennth       https://wiki.gnome.org/Apps/MemProf
14:07 Pennth       valgrind will also work from the command line and appears to be the go-to tool on debian
14:07 Pennth       https://packages.debian.org/sid/valgrind
14:58 Megaf        Pennth: Problem is, that stuff can slow the server down a lot and bring a bad experience to players
20:25 Pennth       megaf: crashing due to memory leaks also annoys the players. Do you have enough proc/ram on your host to run it under valgrind in a separate container and just invite some beta-test users there?
20:27 sapier       valgrind can only detect leaks not "stall" memory
20:28 Megaf        Pennth: VanessaE has reported a memory leak too
20:29 Megaf        and the worse thing is that is not always that happens, is in fact rare
20:29 VanessaE     I did?
20:29 VanessaE     which leak did I report?
20:29 VanessaE     I don't even remember :P
20:30 sapier       the only "leak" I remember reported by vanessae is the excessive memory usage by textures ... which is exactly what I meant with stall memory, vallgrind doesn't bother about memory which is still accessible
20:31 sapier       I suspect for a long time that on megafs system memory is somehow stalled in some queues due to low cpu power those queues never run empty but build up ... but ... thats just a guess I can't really proove
20:35 Megaf        sapier: well, Now I run it on an Intel Xeon...
20:35 kaeza        soo, any comments on builtin cleanup? or will it be left to rot yet again? or should it be closed because WONTFIX?
20:35 Megaf        And I still have exately the same issues
20:35 Megaf        so, not my ARMs fault as you used to blame
20:35 sapier       megaf and there you have still ram usage building up?
20:36 Megaf        sapier: well, in this configuration it happened once
20:36 sapier       how much time did it take?
20:37 sapier       kaeza can you be more precise about what pull you're talking about?
20:37 kaeza        sapier, #1341
20:38 ShadowBot    https://github.com/minetest/minetest/issues/1341 -- Builtin cleanup, part 1 (game). by kaeza
20:40 Megaf        VanessaE: https://github.com/minetest/minetest/issues/1846
20:40 VanessaE     OH  yes
20:40 VanessaE     I forgot about that one
20:40 kaeza        also fixes #1874
20:40 ShadowBot    https://github.com/minetest/minetest/issues/1874 -- Fix some undeclared global variables by CraigyDavi
20:41 sapier       kaeza to much to check if this works merge it BUT if you don't fix the bugs you caused by it, meaning even things you changed intentionally but didn't mention the api change I'll never ever tell you to merge anything ;-P
20:43 sapier       AND fix the commit messages!
20:43 sapier       if not I'm gonna revert it immediatly ;-P
20:43 kaeza        everybody here ignores my commits and proceeds with NIH anyway, so it won't be a change anyway :P
20:44 kaeza        the commit messages don't matter; it will be squashed
20:44 sapier       I expected you to squash it I just wanted to be sure you can't tell later I allowed you to merge something with that crappy messages ;-)