Time Nick Message 00:34 Megaf lol, my in my laptop Minetest can be as heavy as Unigine in Ultra settings 00:34 Megaf this shows how well developed Minetest is... pfff 00:35 * Megaf runs 00:35 Megaf Zeno`: ^ 00:35 Megaf that was a fact by the way 00:37 hmmmm ... what is your point in saying that 00:38 Megaf Unigine in in light mode, http://forum.casemodbr.org/Unigine_Heaven_Benchmark_4.0_20160926_2208.html 00:39 Megaf and in ultra http://forum.casemodbr.org/Unigine_Heaven_Benchmark_4.0_20160926_2235.html 00:39 hmmmm and? 00:39 Megaf hmmmm: I just mean that Minetest is extremely slow and inefficient 00:40 hmmmm do you know anything at all about the inner workings of the engine? 00:40 Megaf Very little in fact. I know that little are being done. 00:40 hmmmm are you going to do something to fix it? 00:41 Megaf sure, I forked Minetest and my fork is much faster already :) 00:41 hmmmm :) 00:41 Megaf so did proller 00:41 hmmmm do you wanna get banned like proller? 00:41 Megaf hmmmm: unfortunetally, my changes will never be accepted back into the project 00:41 hmmmm no? I wonder why not 00:41 Megaf because the things I believe, maybe 98% of the core devs disagree 00:42 Megaf I believe in simplicity and efficiency, I believe in hunting down bugs and optimizing existing code before adding a lot of new features 00:42 Megaf I believe in removing dead code 00:42 hmmmm look, i don't know what your fork is or what modifications you've done, but coming to #minetest-dev, a channel for development chat only, and verbally shitting over minetest to promote your own fork, does not look good for you 00:43 Megaf and sure removing 20 lines of inactive code from a cpp file won't make any difference, but all things together do 00:43 hmmmm GTFO. 00:43 thePalindrome :facepalm: 00:43 Megaf hmmmm: my fork is for my own enternaining. And the way I phrased the things I just said is just to cause some movement. 00:44 Megaf I'm not competing with Minetest in anyway, never will 00:44 hmmmm i'll help you with your movement 00:44 Megaf hmmmm: All I want is the game to be better developed 00:44 thePalindrome There are better ways to do that though 00:44 thePalindrome Like pull requests 00:44 Megaf less new features, less new databases support 00:44 Megaf here we go 00:45 hmmmm I helped him move out of the channel 00:47 hmmmm what's this going on with gregorycu now... 00:48 thePalindrome ? 01:19 Megaf I deserved the kick. I do apolagize. As a redemption gift, I ofter the rebirth of a bug. https://github.com/minetest/minetest/issues/1934 01:23 Megaf Zeno`: about #4306 I read somewhere that this is easy to fix, I was presented with the solution but unfortunetaly I can't remember right now. It's about how cmake is searching for the library, not a bug on Debian actually. 01:23 ShadowBot https://github.com/minetest/minetest/issues/4306 -- Minetest fails to link JSONCPP on Debian 9.x if -DENABLE_SYSTEM_JSONCPP=1 10:07 Fixer PR 4530 works fine for me and bugs I reported earlier seem to be fixed, feel free to merge! 10:10 Fixer #3770 10:11 ShadowBot https://github.com/minetest/minetest/issues/3770 -- Fix superflous shader setting updates by ShadowNinja 10:11 Fixer this old had bugs in it that gregorycu fixed in new pr #4530 that works fine for me 10:11 ShadowBot https://github.com/minetest/minetest/issues/4530 -- Fix superflous shader setting updates by gregorycu 10:11 est31 Fixer, there might still be a memory leak 10:14 Fixer maybe, i've checked memory use just few times, numbers were reasonable at that time 10:27 Zeno` someone else today suggested there might be a memory leak in that PR 10:27 Zeno` but I haven't tested 10:37 Zeno` #4544 looks a lot better now 10:37 ShadowBot https://github.com/minetest/minetest/issues/4544 -- mapblock_mesh_generate_special improvements by gregorycu 10:37 Zeno` apart from the snarky comment :/ 11:04 paramat good grief yet more negativity from Megaf 11:18 Fixer_ just for fun I've launched minetest (just test map with few mods, nonclean), put its window in background with same scene, noted mem usage, will see how it changes in, say, 1 hour or more 11:19 paramat hmmmm Megaf wants to remove all mapgens except v6, despite extra mapgens having no effect on game performance and their files only being 20kB, 8 times smaller than a single font 11:19 MoNTE48 Fixer_ 11:19 MoNTE48 Fixer_ 11:19 MoNTE48 @Fixer_ minetest begins to work very slowly over 24 hours 11:19 Fixer_ client or server? 11:19 MoNTE48 server 11:20 Fixer_ i'm running client with PR4530 11:21 thePalindrome The reason as far as I could tell Megaf wanted them gone was because something about adding new features when old features needed fixing was bad 11:21 paramat "I think I finally got a dev in bad mood" you managed that before several times with me ;] 11:22 paramat if i had not worked on new mapgens i would not have been bugfixing instead, they were created through obsession 11:22 Fixer_ i don't see much bloat in minetest 11:22 Fixer_ it is underbloated actually 11:22 Fixer_ it is just slow though 11:23 Fixer_ MoNTE48: some people reset server every 24hr 11:23 paramat hmmmmm actually strongly supports focussing more on fixing stuff instead of new features 11:23 Fixer_ MoNTE48: to keep it relatively fast 11:24 paramat Megaf hmmmmm was quite gentle with you heh 11:25 thePalindrome I noticed 11:25 MoNTE48 I started server with commandline. 4days and i get minetest.exe crash 11:25 thePalindrome oof, a windows memleak 11:25 * thePalindrome doesn't know the windows equivalent of valgrind 11:46 Fixer_ just by sitting idle rendering one scene, mem usage went quickly from 330 to 340 and stayed there for most of the time 11:48 sfan5 MoNTE48: i suggest not testing performance or memory usage on windows 11:48 sfan5 it will probably be better on linux :) 11:51 paramat looks like #4530 might need to be continued by someone else 11:51 ShadowBot https://github.com/minetest/minetest/issues/4530 -- Fix superflous shader setting updates by gregorycu 11:54 Megaf Fixer_: I'd like more testing on #4530 More code inspecting 11:54 ShadowBot https://github.com/minetest/minetest/issues/4530 -- Fix superflous shader setting updates by gregorycu 11:54 Fixer_ go ahead 11:54 Megaf est31 so far had done a great job at it 11:55 Fixer_ i'm already playing with it 11:56 Megaf paramat: I know, me and devs have a strong hate/love relation. I love them, they hate me. 11:57 Megaf Btw, why do we have two "- Fix superflous shader setting updates"? 11:57 Megaf MoNTE48: When you noticed that? My server keeps the speed after 24 hours, or becomes even faster 11:58 Megaf And the memory leak is really small/slow now a days. I can keep the server running for a whole week if I want 11:58 Megaf Zeno`: ^ 11:59 MoNTE48 My server has no memory leaks. Just a game starts very slowly when playing about 30 people during the 12-24 hour. 12:00 Megaf hm 12:00 Megaf it's been a while since the last time there was 30 people in my server... 12:01 MoNTE48 (Only if I run the GUI server. Not terminal. Tested on Ubuntu 16.04, 15.10, Windows 10 x64) 12:01 Megaf MoNTE48: anyway. It think is just a configuration issue maybe. Or regression, because I had a server with 30 people running on a Raspberry Pi, the first generation, speed was always constant 12:02 Megaf MoNTE48: GUI? Do we have a GUI now? 12:02 MoNTE48 30 players on RPI? 12:02 Megaf MoNTE48: if you launch the server from the minetest client, then the client part of it can make it slower over time yes 12:02 MoNTE48 I run server use Server tab 12:02 Megaf from the command line you will not have this problem 12:03 Fixer_ Megaf: 3770 is outdated and closed, 4530 is never with important fixes 12:03 MoNTE48 But why? 12:03 MoNTE48 This is client bug, but no memory leaks, cpu overload or more IO 12:03 Megaf MoNTE48: I am not sure. Just more things to be processed I guess. 12:04 Megaf MoNTE48: but why do you say it gets slower? Lag increases, IO is high or what? 12:05 MoNTE48 Map not loading , I or any player does not receive a block if it breaks down. Map is not updated at all 12:05 Megaf ok, what database are you using? 12:06 MoNTE48 i test on sqlite and leveldb 12:07 Megaf emergequeue_limit_total = 512 12:07 Megaf emergequeue_limit_diskonly = 256 12:07 Megaf emergequeue_limit_generate = 128 12:07 Megaf set those lines like that 12:07 Megaf maybe increase emergequeue_limit_total to up to 10000 12:07 MoNTE48 i use server_map_save_interval = 900, emergequeue_limit_total = 2048, emergequeue_limit_diskonly = 512 emergequeue_limit_generate = 512 12:08 MoNTE48 no result 12:08 Megaf max_simultaneous_block_sends_per_client set to 4 12:08 Megaf server_map_save_interval, I have this set to 3 seconds on my server, the shorter the period, the less thing will have to be writen and the faster it will be 12:08 MoNTE48 But as soon as I began to run minetest via terminal, everything works fine on the default settings! 12:09 Megaf also full_block_send_enable_min_time_from_building can be very high, like 20, I have it set to 200 on my server 12:09 Megaf MoNTE48: I know, the settings I'm saying are settings that will affect your performance over time 12:10 Megaf running a server on a raspberry pi forces you to learn one thing or two about optimization 12:10 MoNTE48 server_map_save_interval - i have 16gb 2400mhz ram. 900 value use 4-5gb ram only 12:11 Megaf MoNTE48: so, it doesnt make sense your server slow down with that configuration. The other thing you want to check in your config is item_entity_ttl. If the value is too long, items will be acumulating around 12:11 Megaf I have it to 30 seconds 12:12 paramat running any mods like moretrees or plantlife? 12:13 MoNTE48 I will try what you say ... 12:13 paramat try disabling all mods 12:14 MoNTE48 i use this mods https://github.com/MultiCraftProject/MultiCraft/tree/master/games/default/files 12:14 Fixer_ MoNTE48: 0.4.14 stable or dev? 12:15 MoNTE48 0.4.14 stable 12:16 paramat sorry that was directed to Megaf 12:16 Megaf paramat: I have moretrees, plantlife and a lot more mods. No impact on performance. Not on x86, not sure if you are asking about my current server or the RPi 12:17 paramat ok. just a general question 12:17 Fixer_ MoNTE48: there was a bug that was fixed after release, that may be your problem 12:17 paramat basic nodes shader now slows stuff down btw 12:18 MoNTE48 @Fixer_ pls url on commit 12:18 Fixer_ MoNTE48: symptoms like I seen on ESM server, were fixed after updating to -dev 12:18 paramat hm but that's client 12:19 Fixer_ MoNTE48: https://github.com/minetest/minetest/commit/7ea4a03c835d68a6fb58aa55aa6a6315ec80b79f 12:19 Megaf I got, cya 12:21 MoNTE48 @Fixer_ ok i try 12:24 paramat Zeno` please could you consider taking over on #4530 ? :] 12:24 ShadowBot https://github.com/minetest/minetest/issues/4530 -- Fix superflous shader setting updates by gregorycu 12:26 paramat funny spammer https://forum.minetest.net/viewtopic.php?p=234588#p234588 12:31 Zeno` why me? lol 12:32 Zeno` I'll see if the suspected memory leak is real or not though 12:33 paramat heh sorry 12:33 Zeno` why has gcu abandoned it? 12:37 Fixer_ Zeno`: i'm running it in background, mem usage: 330-340 - first hour, 340-346 - second hour 12:37 est31 Fixer_, its likely not a big leak 12:37 est31 also leaking behaviour only happens when you repeatedly leave and join a game 12:38 Zeno` well, it should be easy to fix anyway (just looked at the source) 12:38 Zeno` I really need to get off this chair. I'll bbiab (for real this time) 12:39 Fixer_ oh... should not be much then, how many times you restart it per day? 10-20 at most... 12:40 Zeno` the leaked memory (on most if not all modern operating systems) is reclaimed after shutting down anyway 12:40 est31 yes 12:40 Zeno` it's more of a concern if you don't restart it often 12:40 est31 still saying the leak is small and therefore doesnt need a fix is like when saying you build a ship and it has a hole and water is coming in and you are not closing the hole because you say it doesnt make the ship sink 12:41 Fixer_ i was not saying that 12:41 Zeno` Oh, I am not saying it shouldn't be fixed at all :) In fact I think it's an easy fix just by glancing at things 12:44 Fixer_ good you spotted it, since if you merge it like that, small leak will be hard to spot later I guess 12:50 Zeno` I think hmmmm spotted it 12:51 Fixer_ right 16:04 rubenwardy will merge game#1295 as "Farming: Remove unnecessary overriding of groups" in 5 mins 16:04 ShadowBot https://github.com/minetest/minetest_game/issues/1295 -- Remove un-needed group overrides by tenplus1 16:06 rubenwardy also, I've been busy lately with different real life issues (not just uni) but should be more available now 16:17 paramat ok 16:22 sofar oh wow github added `rebase and merge` 16:24 Calinou sofar: is that different from "squash and merge"? 16:24 sofar yes 16:24 sofar it doesn't squash :) 16:24 sofar so if there's 3 commits in the PR, they stay 3 commits 16:27 rubenwardy that's cool 20:27 Megaf [13:40:53] the leaked memory (on most if not all modern operating systems) is reclaimed after shutting down anyway 20:27 Megaf Was that a joke? 20:28 Megaf That's what I expect after shutting down a "program" or the whole OS