Time Nick Message 01:13 paramat "move the mapgen specific params to mapgen.h anyway, and declare them directly in the mapgenparams" no thanks, so i agree with you 01:16 paramat hmmmm see https://github.com/minetest/minetest/pull/4216#issuecomment-225747280 and following comment, PR seems ready now 01:28 hmmmm yeah but I mean vs. the old algorithm 01:30 hmmmm also just a tip, if you want more detail with a greater spread value, you can increase the lacunarity 01:30 paramat i've recreated the old behaviour through choice of noise parameters 01:31 hmmmm yeah but i want to actually see a comparison if that's okay 01:31 paramat ok 01:34 paramat i trust the params more than judging it visually, but okay 01:45 paramat eh ok so it was worth checking after all, need to tune the offset 01:46 paramat not enough moss 01:48 paramat ah that 01:49 paramat what is a good value for content to indicate 'nil' or 'null'? 01:49 paramat here i used '0' for when no alt_wall is wanted 01:50 paramat 'NULL' didn't work because content is an unsigned integer 01:56 paramat replied in the thread 02:05 paramat afk for food 03:06 hmmmm yeah paramat whenever you make an adjustment to noise params like that you need to actually see the difference you're making 03:07 hmmmm you can use CONTENT_IGNORE for "no content" 03:07 hmmmm of course content_t is not a pointer 03:08 paramat yeah i'll use content_ignore 03:08 hmmmm i wanna change all that stuff soon anyway 03:08 hmmmm cause look 03:09 hmmmm caves are very similar to dungeons and yet their interface isn't consistent at all 03:09 hmmmm with your PR you're now making multiple dungeons inside of DungeonGen::generate() instead of inside the mapgen 03:10 hmmmm whereas if you consider caves, makeCaves only generates one set of tunnels, and to make multiple cave systems in the chunk, the mapgen is the one responsible for doing it more than once 03:11 paramat makes sense 03:11 hmmmm i'm not telling you to go change it, i'm saying consider this 03:11 hmmmm i'm aiming to separate the concerns as best as possible and right now it's not exactly a mess but it could be better 03:11 paramat ok 03:12 hmmmm like for example caves and dungeons shouldn't have logic that determines their frequency or whether they occur above water_level or similar 03:12 hmmmm those are things that mapgens, the objects that actually use cavegens and dungeongens, worry about 03:12 hmmmm all a cavegen should worry about is making a cave at point XYZ 03:46 paramat hmmmm, PR updated 04:55 hmmmm btw paramat, https://github.com/kwolekr/minetest/commit/dacbc47994723a4615ba7060749997c451f0f352 08:54 NeueMann When following these instructions: http://dev.minetest.net/Compiling_Minetest#Building_without_Irrlicht_.2F_X_dependency what should one set for cmake? 08:57 Calinou NeueMann: -DBUILD_CLIENT=0 should be enough 08:57 Calinou you can install Irrlicht packages anyway, without requiring a desktop environment 09:54 NeueMann I'm having trouble when using biome_lib: http://pastebin.com/GNBXp43k 09:54 NeueMann This comes up when I try to join the game. 09:55 NeueMann I also get this when starting up: WARNING[Main]: system-wide share not found at "/usr/local/share/minetest" 09:57 asl97 NeueMann: did you make install it? 09:58 asl97 it seem like you might have build a system-wide built instead of a run_in_place 09:58 NeueMann No. I'm running from the folder. Do I need to? 10:01 * Fritigern uses this script and never has trouble... http://pastebin.com/D0d6hJU2 10:03 Fritigern Oh, you want to build without irrlicht? Never done that. Perhaps you can use the script as the basis for your own build 11:05 Calinou asl97: NeueMann: you can ignore that warning 16:40 paramat game#1137 boats were broken by an innocent looking commit, will fix today 16:40 ShadowBot https://github.com/minetest/minetest_game/issues/1137 -- Boats sink if they get under water. But they sink through stone too. 16:42 paramat nore i think game#1135 can be closed, any comments? 16:42 ShadowBot https://github.com/minetest/minetest_game/issues/1135 -- buckets won't run out of water/lava in creative mode by cdqwertz 16:43 paramat nice, no more mapgenfactory 16:43 hmmmm unfortunately it's only a 4 line reduction 16:44 asl97 #4091 is a small fix, hopefully it didn't gotten forgotten 16:44 ShadowBot https://github.com/minetest/minetest/issues/4091 -- mesh tiling bug and possible fix 16:44 hmmmm i never even knew about it 16:45 hmmmm huh i thought grouping up to 2 nodes only was the expected behavior 16:46 asl97 wat... why would that be the expected behavior? 16:49 hmmmm i approve of the change though 16:50 hmmmm paramat: http://i.imgur.com/NtHJVtc.jpg this is how big a river valley actually is lol 16:50 paramat nice 16:51 paramat i'm not keen on mgvalleys, lumpy and too small-scale 16:51 hmmmm mgvalleys is like a bunch of creeks 16:51 paramat but it's very popular 16:52 hmmmm well, at least it's adjustable 16:52 hmmmm i wish you could modify the river depth too 16:52 hmmmm noise based rivers kinda suck though 16:53 hmmmm i think millersman has the right idea; global map generation should be the way forward and there's really no way to do this without persistence 16:53 hmmmm i was going to do a similar thing except with a low-res bitmap image stored in the map directory, not a database 16:56 hmmmm in any case, if you want noise-based rivers similar to valleys except without all the extra math, log(x^2) seems to work well 17:18 diemartin celeron55, ShadowNinja, anyone else with permissions, would it be possible to add a mobile-friendly theme to the forums? 17:18 diemartin excuse the offtopic 17:18 Krock is it really the theme or would it need a forum software upate too? 17:19 celeron55 do you have a theme suggestion or what action do you think needs to be taken? 17:21 celeron55 i can add a secondary non-default theme if someone knows a good one; it would be nice if it would be customized a bit first though; generic themes give a cheap feel 17:22 diemartin I'm not entirely sure on what needs to be done, but there's this: https://www.phpbb.com/community/viewtopic.php?f=73&t=2138277 17:23 diemartin I don't really care about the colors but looking at the forums on mobile is a pain 17:24 paramat confirmed the fix for game#1136 will fix today 17:24 ShadowBot https://github.com/minetest/minetest_game/issues/1136 -- Searching blocks/items glitch, in new creative interface. In version 0.4.14 17:27 paramat not really a theme then, more an alternative simplified layout 19:00 hmmmm you know what 19:00 hmmmm i think i wanna get rid of the flags 19:00 hmmmm flags are an implementation detail of an array of booleans, and users shouldn't be directly exposed to that 19:00 hmmmm does anybody agree? 19:02 celeron55 of course they are 19:03 hmmmm i mean agree to get rid of them in favor of a bunch of booleans 19:03 celeron55 however, it doesn't directly imply flags should never be used in an interface 19:03 celeron55 but it could 19:03 celeron55 you might have noticed that in no part of the Lua API that i designed there are any C-style flag fields 19:05 celeron55 (altough... there might be none in the implementations either) 19:05 celeron55 s/either/also/ (how does this even work) 19:05 DI3HARD139 Question. How do I change the release setting to debug when compiling from github? 19:06 celeron55 you pass -DCMAKE_BUILD_TYPE=Debug to cmake 19:06 celeron55 (it's a standard cmake thing) 19:07 DI3HARD139 So just add it to the end of cmake . -DRUN_IN_PLACE=1 -DENABLE_GETTEXT=1 -DENABLE_FREETYPE=1 -DENABLE_LEVELDB=1;? 19:07 celeron55 assuming of course that you are not using visual studio, which cmake handles specially 19:07 celeron55 yes 19:08 DI3HARD139 I'm compiling using the 1-line script from Calinou 19:11 DI3HARD139 ty. I just reimaged my server using the base build of ubuntu-server and added the xfce 19:25 Hijiri If I set the metatable of a LuaEntitySAO, will it have the same metatable when the object is fetched again e.g. from a punch callback? 19:47 paramat hmmmm, about flags: the 'trees' and 'flat' flags were moved from mg_flags to mgv6_spflags, but to maintain backwards compatibility i had to leave them duplicated in mg_flags, just not documented. this can be confusing because they still show up in map_meta.txt in mg_flags. so it might be good to sort this mess out by breaking compatibility flagswise 19:48 paramat obviously we need to support old mgv6 flat worlds, but the mgv6 'trees' flag could be replaced by the new global 'decorations' flag 19:50 paramat as for flags vs bools i have no opinion so far 19:53 paramat personally i would rather users are forced to change a flag or two than continue with this flag mess i created 19:56 Ambistic I'd like to help translate the wiki to french, is there a core dev that may help me ? 19:56 paramat since you're making big changes to mapgen this would be a good time to do it and break flags compat at 0.4.15 20:03 hmmmm celeron55: i wish you would've stopped me when i tried to add the first flags field 20:03 hmmmm the way i envisioned it, it was like the flags on a command line 20:03 hmmmm --fno-foobar --something-else --etc 20:04 hmmmm but that's not very intuitive i guess 20:04 celeron55 well, you can always add a compatibility converter from flags to something that will be used for the foreseeable future 20:04 hmmmm paramat: i am not implying anything about what you did, just a question about design in general 20:05 hmmmm yeah in fact I had an idea for making a SettingsMigrator class 20:05 Ambistic I'd like to help translate the wiki to french, is there a core dev that may help me ? 20:05 hmmmm you hard code a list of old names, a list of new names, then when Settings loads a config file, it does the conversion silently 20:05 hmmmm and then it removes any instances of the old names on save 20:06 hmmmm Ambistic: either nrzkt or Calinou could probably help here 20:08 paramat also we have mg_flags = dungeons in defaultsettings.cpp for some complex compat reason 20:08 hmmmm ANYWAY THE POINT IS we can easily do settings name and type changes in a reverse-compatible manner and I can't figure out any reason why we shouldn't 20:09 paramat if you set mg_flags = nocaves for a world that will bizarrely disable dungeons 20:09 paramat you have to end up doing mg_flags = nocaves, dungeons 20:09 hmmmm yeah that's an unintended consequence 20:09 hmmmm this is pretty messy. 20:10 hmmmm but that's not what i'm talking about 20:10 hmmmm i'm just talking about the whole concept of exposing flags to the end user 20:10 hmmmm it's actually not that great of an idea 20:10 hmmmm it should be an internal thing only 20:10 hmmmm what do you think, do you like flags? 20:22 paramat i don't mind them, they have been a little confusing, maybe docs are unclear 20:23 paramat i might prefer a set of seperate bools, having the one string of many flags is perhaps not ideal 20:24 paramat a bool is clearer than using a 'no' prefix 20:27 paramat ok so i'll put it this way, i think the system could be better and clearer 20:30 paramat but more important to me is to, at least, sort out the 'trees' and 'flat' mess 20:40 sapier hello 20:42 Fixer hi sapier 20:43 est31 hello sapier 20:43 est31 how are you doing 20:43 est31 long time no see 20:44 sapier true I've been extremely busy in my paid work ;-/ ... well not much better now but I'm gonna try to answer questions ... maybe I remember whats required to know 20:53 nore hi sapier :) 20:54 paramat our android dev! we have been looking for a new one 20:55 hmmmm i'm of the opinion that we should pluck android support for mt if we don't have any android developers 20:55 hmmmm it's in a very sorry state right now and shows no signs of getting fixed 20:56 paramat i've suggested that too. android seems to cause lots of grief 20:56 est31 hmmmm, cant that be said for whole minetest 20:56 Fixer hmmmm: critical bugs? it kinda works at least 20:57 sapier guess it did never evolve beyond the state I left it? 20:57 est31 not much 20:57 est31 almost not at all 20:57 est31 kept on life support basically 20:57 est31 I've done some fixes to the build system 20:57 est31 and monte has done some smaller improvements 20:57 est31 thats it 20:57 sapier sad ... I was hoping this would turn into some sort of self running thing after the initial work is done 20:59 paramat we're really understaffed recently 20:59 est31 We are better staffed than irrlicht 21:00 est31 basically one person project atm 21:00 est31 with second person sending patches now and then 21:00 paramat nore sfan5 i'll merge this later, trivial fixes for 2 bugs game#1139 21:00 ShadowBot https://github.com/minetest/minetest_game/issues/1139 -- Fix boat falling through world and grey screen in creative search by paramat 21:01 sapier one person at irrlicht? ... explains why it's android support is that buggy 21:01 est31 at least right now. 21:03 sapier I'd like to say I'm gonna have a look at it but I'm quite sure I'm not gonna have enough spare time till september to do any really productive work 21:13 hmmmm yeah like sapier said it "works" but it's not "evolving 21:13 hmmmm it barely works 21:13 hmmmm let's face it, people who develop for mobile platforms are usually have a profit motive 21:13 hmmmm they are not interested in developing open source 21:14 sapier1 sad to admit it but you could be right hmmmm 21:15 est31 hmmmm, agreed, for >= 95% this is the primary and only motive 21:15 est31 monetisation 21:15 est31 but still there is the 5%, maybe less 21:15 hmmmm fwiw you can make a shitton of money on mobile 21:15 hmmmm ads are like 0.003 cents per view or whatever 21:15 hmmmm but you scale that by millions and it's still a lot 21:15 est31 there are many many open source apps 21:16 Fixer adaway installed :} 21:16 sapier1 well if nothing changed our main issue is we don't have anyone really capable of designing a really good mobile ui 21:16 hmmmm that one guy who ripped minetest and changed some default settings and w/e made like $100k off of it 21:16 hmmmm lol :( 21:16 est31 hmmmm, https://f-droid.org/ 21:16 Fixer hmmmm: no way, proof? 21:17 hmmmm i don't have any proof but after looking at the math i actually do believe it's possible 21:17 Fixer people play with ads? mazohists 21:17 est31 Fixer, indeed they do 21:17 est31 many players join servers 21:17 est31 and say 21:17 hmmmm most people are normies and they think ads are just part of life 21:17 est31 "your game is shit, it is full of ads" 21:17 Fixer ads should die 21:17 hmmmm ad injection 21:18 est31 they think the game is called "mine builder" or something 21:18 hmmmm because we use udp for file transfer we're vulnerable to get MITMed 21:18 hmmmm I bet the next step in advertisement evolution is to silently replace the regular textures with advertisement images when it passes through their router 21:18 hmmmm rofl 21:18 est31 lol 21:19 est31 wait for google fiber to start with it 21:19 est31 free internet, but your minetest blocks will have ads on them 21:19 Fixer i don't mind, people want join isis or watch ads? fuck em, thats their choice 21:19 est31 nodes* 21:20 Fixer i will not go like Stallman trying to stop people eat shit %) 21:20 est31 its their choice to join isis and its the choice of the usa to send drones to kill them 21:21 hmmmm ISIS recruitment mod 21:21 Fixer i wonder are there any ISIS ads? 21:21 est31 lol 21:21 sofar there's a solution to all this 21:21 sofar just block forks from connecting to servers 21:21 hmmmm play as a jihadi militant and slaughter as many innocent villagers (mobs) in a randomly procedurally-generated sand village 21:22 hmmmm ace of spades meets terrorism 21:22 est31 sofar, thats the totally wrong approach 21:22 sapier1 @zeno thanks for pushing https://github.com/minetest/minetest/commit/7ea4a03c835d68a6fb58aa55aa6a6315ec80b79f I don't know why I missed to push this obvious error fix to minetest myself 21:22 sofar est31: sure, it's not a nice approach :) 21:22 est31 sofar, the problem is equivalent to "prevent people from saving the online video to their disk" 21:22 est31 requires DRM essentially 21:23 sofar if we want to "protect" the minetest community (and game), we should consider all methods available seriously 21:23 hmmmm sorta reminds me of warcraft 3 21:23 est31 another good way would be to go GPL 21:23 sapier1 come one wc3 is way to new ... :-) 21:24 sapier1 wc1 rocks 21:24 hmmmm the client will refuse to connect to a server if the signature doens't match the IP address padded with a bunch of NULs 21:24 est31 and create a special licensing organisation or something 21:24 est31 which gives play store exceptions to selected forks 21:24 sofar or trademark enforcement.... 21:24 Fixer sapier1: i'm not totally sure, but i've noticed smth interesting on mt with this patch (can't say before patch), from time to time (rarely) received packets go in big spikes of 50-100-200-400 21:24 Fixer sapier1: at least on Extreme Survival Minetest server 21:24 Fixer not sure if mod problem or patch problem or fubar in general 21:25 sapier1 fixer could you try to find out if this is a counting issue? e.g. maybe droped packets are counted as "received" 21:25 nrzkt i agree with sofar 21:26 est31 nothing against registering a minetest trademark 21:26 Fixer sapier1: not a dev, but i think i didn't noticed much dropped packets on f5 graph 21:26 est31 but what would it help against an app that calls itself "mine builder pocket edition"? 21:26 Fixer sapier1: occasionally there are 1 or 2 dropped packets 21:26 est31 and is essentially re-branded minetest 21:26 sapier1 maybe the ones dropped due to client timeout are counted as received 21:27 Fixer sapier1: maybe i need to observe this more, people often go in and out 21:27 est31 but shrug, I think re-branded versions of minetest aren't something really bad 21:28 est31 minetest is open source this is what happens 21:28 est31 and it brings in new users 21:28 est31 there were many players who started with one of these adware infested apps 21:28 est31 and then switched to the official app 21:28 est31 or got minetest for the desktop 21:29 sapier1 fixer at least by looking at the code I see no reason how this patch could cause it ... but that doesn't rule out it somehow triggers another flaw in code which was hidden 21:29 Fixer est31: maybe servers should say that "If you see ads on your game, you are scummed, download official client" 21:30 est31 sapier1, you know why the client first sends a packet with two 0 bytes? 21:30 est31 https://github.com/minetest/minetest/blob/423d8c1b0d3841f6ce9756ab0d6b2e10408fc89b/src/network/connection.cpp#L1742 21:31 est31 I've wondered about it when I did https://github.com/minetest/minetest/commit/423d8c1b0d3841f6ce9756ab0d6b2e10408fc89b 21:31 est31 not really important, if you dont know its no problem 21:31 sapier1 that code was modified past my changes but if I remember correct this is regular join 21:32 sapier1 client doesn't know it's own peer_id by that time so there's not much it can tell to server 21:32 est31 thats not the part I am wondering about 21:33 est31 i mean the client sends its peer id in order to speed up packet routing 21:33 est31 on the server side 21:33 est31 thats the only reason, otherwise one could just use the ip address and port 21:33 sapier1 sorry I don't really know the new style login mechanism, the old style one was suboptimal 21:34 est31 its before any login 21:34 est31 btw, I implemented the new style login mechanism, so I know it well 21:34 est31 if you have any questions, ask 21:34 est31 :) 21:34 sapier1 there hasn't been information about ip adress and port in original networking code 21:35 est31 it may have been added later on 21:35 sapier1 maybe it's just "legacy" 21:35 est31 as a security fix 21:35 est31 sapier1, because if there is no ip address based check, a malicious client can fake the peer id e.g. of an admin 21:35 est31 and fake a chat command packet with /grant all 21:36 sapier1 most likely I completely rewrote that low level code but as my goal was to write a compatible solution till a decision is made for a better one I couldn't fix all issues 21:36 est31 sapier1, nobody is blaming you. 21:37 Fixer sapier1: hmm, joined server, for 2 minutes observed it, no people left in this period, i've observed rare 150 packets spikes 21:37 sapier1 est31 wouldn't bother anyway it's more then a year old, I've learned quite a lot since then ;-) 21:37 Fixer what are those guys? 21:37 sapier1 fixer did anyone connect? 21:38 Fixer sapier1: yes 21:38 Fixer sapier1: randomly, no correlation for me 21:38 sapier1 spikes in sending or receiving? 21:38 est31 Fixer, they have a different charset than the server 21:38 Fixer sapier1: received 21:38 est31 Fixer, and they send special chars 21:38 est31 Fixer, the server converts it to utf-8, and then back 21:38 Fixer est31: ah 21:38 sapier1 that's really strange 21:39 est31 but if the charset doesnt match it will make the utf-8 string broken 21:39 sapier1 fixer maybe it's a reliable packet beeing lost 21:39 est31 wchar is platform dependent 21:39 sapier1 e.g. one packet is lost but following packets are received 21:40 sapier1 once the lost packet is retransmitted all follow up packets can be processed immediatly 21:40 Fixer sapier1: it is just i never noticed this before patch, maybe it was, but I have not noticed :( 21:41 sapier1 if this explanation is true it should've happened before too 21:42 sapier1 if you find a way to reproduce that error in an environment where you can do packet capturing that'd help 21:42 sapier1 but you can't do packet capturing on a production server 21:43 paramat hmmmm #4216 is updated 21:43 ShadowBot https://github.com/minetest/minetest/issues/4216 -- Dungeons: Generalise. Add new parameters by paramat 21:43 sapier1 sorry have to get some sleep now, hope to see you soon 21:43 Fixer sapier1: you want server side? doubt it, while he hosts it on his PC, he is not completely crazy to debug it 21:44 sapier1 yes a server side dump could help, but of course not if there's any real user data in it 21:44 sapier1 minetest port only of course 21:44 sapier1 another option would be a description how to reproduce the issue 21:45 sapier1 sorry have to go now good luck 21:47 Fixer i really miss that celeron server days 21:47 Fixer with actual dev in game 21:47 Fixer and his c55.me server 22:22 paramat merging game#1139 22:22 ShadowBot https://github.com/minetest/minetest_game/issues/1139 -- Fix boat falling through world and grey screen in creative search by paramat 22:23 paramat and game#1120 , almost forgot 22:23 ShadowBot https://github.com/minetest/minetest_game/issues/1120 -- Mossy cobble slab and stairs. by sofar 22:36 paramat merged 22:38 Glorfindel I noticed that in #minetest lol 23:54 Fixer Glorfindel: thats expected, no worry