Time Nick Message 01:05 ar10ch Yes! 01:05 ar10ch fixed it 01:06 * cheapie offers to nibble on it until it's broken again 01:06 ar10ch haha it'll be broken by morning lol 01:14 ar10ch oh no, i didn't fix it. oh bugger 08:34 DivideByZero Does anyone use SkinsDB & downloaded all the skins? Have you noticed long server load times by doing this? 10:55 DivideByZero Is there instructions somewhere for how to interpret the /profiler data? I'm trying to figure out why my server is slow to start up or when clients connect, particularly Android clients which often timeout trying to connect. 10:58 DivideByZero Else is there someone knowledgable with it who is willing to take a look at my profile*.txt ? 12:18 sfan5 slow startup is solely due to what mods do at startup (duh), slow to connect should be mostly time spent in media transfer unless mods do something weird 12:19 rubenwardy on Android, it's due to network packet issues 12:20 DivideByZero In my log file the biggest delay in startup is occuring right after 'map' mod loaded (which I guess must be a build in part, because I didn't install any mods called 'map'. But I don't think it's the 'map' itself, but whatever occurs right after that, and it doesn't show anything happening for a while, it just seems to hesitate for a while seemingly doing nothing, although it may in fact be doing something and just not showing it in the verbose 12:20 DivideByZero log or STDOUT. 12:21 sfan5 rubenwardy: I thought those only affect aarch64 12:21 sfan5 hm well we're forced to ship aarch64 on the play store aren't we? 12:21 rubenwardy right, which is what Google Play will be giving to most Android devices 12:21 rubenwardy yeah 12:22 sfan5 DivideByZero: the verbose log prints times for every mod after it's done, are you looking at those? 12:22 DivideByZero yes, and they all seem acceptable times. 12:23 DivideByZero Whatever part is causing it, doesn't seem to tell me. 12:23 DivideByZero I was trying the /profiler sfan5 that you recommended to someone else to try in the forum. 12:24 DivideByZero But I also don't see anything listed in there that shows it taking a big time either. 12:24 DivideByZero I do see a couple longer times but they have a `-` in front of them, I don't know what that means. Looks like a negative number, but it likely means something else. 12:28 DivideByZero I put it up here if anyone could take a peak at it (the profiler data): https://pastebin.pl/view/0232b4ca 12:30 DivideByZero Thank you. 12:41 Andrew__ I don't get how "ardunio" can just hack in and DESTROY spawn 12:41 VanessaE more likely the spawn, whatever it is, isn't properly protected, or that user has elevated privileges 12:42 Andrew__ That user got the SERVER account 12:42 Andrew__ Ouch, the owner of the world gave a empty password to SERVER. What the hell 12:42 VanessaE I'd call that elevated privileges. 12:43 Andrew__ yup 12:43 MTDiscord /setpassword SERVER correct horse battery staple 12:43 MTDiscord /kick SERVER 12:43 VanessaE so no hacking happened. you just have an idiot admin :) 12:43 Andrew__ :\ 12:43 Andrew__ Yup :\ 12:43 VanessaE :) 12:43 * VanessaE changes appguru's password to Tr0ub4dor 12:44 MTDiscord is hacked 12:44 VanessaE (too lazy to look up the xkcd to check the spelling ;) ) 12:47 Andrew__ well turns out that they deleted the whole map 12:48 Andrew__ Compiling minetest 5.4 on a raspberry pi, irrlicht does not get found 12:54 Andrew__ http://ix.io/2VuN 12:55 MTDiscord 1. Clone Minetest's Irrlicht fork 12:55 MTDiscord 2. Compile it 12:56 VanessaE 5.4.0 doesn't use the MT fork 12:56 MTDiscord oh 12:56 sfan5 yeah just apt-get install libirrlicht-dev and try again 12:56 VanessaE or doesn't *need* it anyway 12:56 Andrew__ oh, thx 12:57 Andrew__ Wait, it's a raspberry pi 12:57 sfan5 you want a client to run on there right? 12:57 Andrew__ No, a server, so i dont get why irrlicht is necessary 12:57 MTDiscord headers 12:57 sfan5 it's not, well not entirely 12:58 sfan5 in that case you can in fact clone https://github.com/minetest/irrlicht and then point -DIRRLICHT_INCLUDE_DIR= to there plus /include 12:58 sfan5 (and make sure to use -DBUILD_CLIENT=FALSE) 12:58 Andrew__ ah 12:58 Andrew__ DBUILD_CLIENT=FALSE is what i need 12:59 Andrew__ but seems that the old version in the raspbian repos start working 12:59 * Andrew__ compiles directly on a Pi rather than cross-compiling 13:01 Andrew__ Is there a way to version-control the map 13:01 Andrew__ git doesn't diff binaries 13:02 VanessaE use level db for the map database and take backups periodically 13:02 sfan5 particularily with leveldb incremental backups work 13:02 VanessaE mmhmm 13:02 MTDiscord Postgres should support history too 13:03 * Andrew__ doesn't use postgres though 13:03 Andrew__ Compiling stuff on a RPi takes a long time 13:12 Andrew__ i 13:12 Andrew__ done! y 13:12 Andrew__ thanks 13:37 Andrew__ I wonder how people could delete the whole map? 13:38 MTDiscord With server privs it's not that hard 13:39 MTDiscord Using clearobjects you can get rid of all items, mobs etc quickly 13:40 MTDiscord And using deleteblocks, you can get rid of everything else 14:21 Wuzzy Any JT2 players here? 14:48 MTDiscord Does minetest on raspberry pi have goto capabilities? 14:48 MTDiscord Like goto continue 14:48 MTDiscord I mean, insofar as I could tell you where you could goto with a question like that, I guess...? 14:49 sfan5 you mean Lua? that depends on whether you build with luajit 14:49 MTDiscord ah okay thank you 14:49 MTDiscord Will "goto" ever be supported in non-luajit versions? 14:49 rubenwardy no, because we depend on Lua 5.1 14:50 MTDiscord You essentially can't safely use it in a mod even if JIT allows it anyway 14:50 sfan5 IIRC the issue wasn't that we couldn't move to Lua 5.2 but that luajit doesn't provide exactly the 5.2 feature level 14:51 MTDiscord Oh crap 14:51 MTDiscord Just use a local function 14:51 MTDiscord How am I supposed to program in "goto" as a local function? 14:51 rubenwardy I've never needed to use a goto 14:51 rubenwardy try rephrasing your code 14:51 sfan5 you never need to use anything 14:52 rubenwardy splitting out code into another function can help 14:52 rubenwardy Lua is missing continue much more than goto 14:52 MTDiscord I'm using it to write cleaner code, but this means that some things are available in only half the versions of the builds available 14:52 MTDiscord The cost of an extra function call should be tolerable under JIT, especially if you're encapsulating your nested loops all in that function, which you'd be doing if you felt the need to escape them all. 14:53 MTDiscord So testing on the same system with the same build with different compiler settings in the same code can crash oh shit 14:53 MTDiscord If you ever figure out a way to use goto to write cleaner code, you should publish. 14:54 MTDiscord This actually isn't the point anymore warr 14:54 MTDiscord I almost ran into a performance inversion between JIT and PUC that was significant enough that it would have compelled me to detect which one and swap out code based on that ... but it turned out not to be reliable under different system loading. 14:55 MTDiscord There is not feature parity, which I have to test using non-jit builds from now on then test them into the jit build 14:55 MTDiscord sure ... or just remember the handful of features that don't have parity and don't use them. 14:55 sfan5 apart from goto it's a some obscure stuff, doubt you're using them accidentally 14:56 MTDiscord I was apparently local-testing using PUC and running my servers on JIT for a couple years and didn't even notice until I tried to test for JIT... 14:56 MTDiscord other than the "apart from goto" part, yeah 15:10 sfan5 I looked it up and the most relevant change 5.1 to 5.2 probably is the removal of getfenv and setfenv 15:17 MTDiscord Don't take those away from me 15:35 MTDiscord yes take those away from you 15:35 MTDiscord iirc you can set the _ENV variable to change the environment of a block of code 15:38 MTDiscord what are ColorStrings? 15:39 MTDiscord setfenv can do very dangerous things and you shouldn't use it normally 16:10 MTDiscord well it accepts a stack level, which is kinda dirty 16:12 MTDiscord I could live with _ENV though 17:31 VanessaE BuckarooBanzai: Please test biome_lib commit d06ab90e . You shouldn't notice much if any change in-game but in the background, you'll find better old-block timeout control, and from the terminal/cmd prompt, better block log reporting, especially at shutdown. 17:32 VanessaE others: it's in the Content Store and Dreambuilder as of ~5 mins ago. 17:39 VanessaE or at least it *should* behave better, if I did my job right :P 19:31 MinetestBot 02[git] 04sfan5 -> 03minetest/minetest: Update Android build config 1357218aa https://git.io/JOvQO (152021-04-09T19:16:45Z) 20:34 MinetestBot 02[git] 04Wuzzy2 -> 03minetest/minetest: Don't reseed stars when changing star count 13e89e6c8 https://git.io/JOvAn (152021-04-09T20:33:45Z) 20:36 MinetestBot 02[git] 04sfan5 -> 03minetest/minetest: Fix server favorites not saving when client/serverlist/ doesn't exist… 130abc1e9 https://git.io/JOvAa (152021-04-09T20:36:10Z) 20:36 MinetestBot 02[git] 04Wuzzy2 -> 03minetest/minetest: Update German builtin translation 13a0e7a4a https://git.io/JOvAw (152021-04-09T20:34:18Z) 20:36 MinetestBot 02[git] 04Wuzzy2 -> 03minetest/minetest: Update builtin locale files 138c7e214 https://git.io/JOvAr (152021-04-09T20:34:18Z) 23:14 TechDude Hi 23:14 TechDude i have a question 23:14 TechDude is it possible to create a new world from the command line? 23:15 TechDude If I could give the world a seed, that would be even better 23:20 Glorfindel I'm pretty sure, yes. windows? 23:34 Calinou --worldname some_world_that_doesnt_exist 23:35 Calinou that creates a new world if it doesn't exist 23:35 Calinou as for specifying the seed, not sure if you need to do it in minetest.conf, or even by modifying some configuration file in the world folder 23:35 Calinou (after the world is created, so you have to manually clean up all the already-generated mapblocks somehow) 23:40 cheapie Calinou: --worldname just gives an error if the world doesn't exist - seems that you have to use --world and specify the full path to a nonexistent world if you want to actually create one. 23:41 cheapie As far as deleting all generated mapblocks go, you can just delete the map DB file and it'll create a new one when started next.