Time Nick Message 00:02 sofar It's strange to see git describe --tags say 0.4.11-git-bla and not 0.4.12-git-bla 00:43 paramat yes supporting 6d rotation would be a challenge, 24 orientations x 3 rotations to consider =P i'll try rotating some horizontal branches again to see if rotateAlongYAxis actually works for 6d, i don't think it does but should check again 01:13 paramat nope horizontal branches are not rotated, for now then i'll disable that function for param2 > 3, and check wallmounted 01:14 paramat i just added a sandstone grassland biome to my biome system, the sandstone cliffs and mountains are beautiful, and make this material a resource 01:25 paramat param2 seems to be used differently for wallmounted, so i'll only alter the facedir code 01:42 paramat 'rotateAlongAxisOfEvil' 01:55 paramat hmmmm here it is for review #2666 01:55 ShadowBot https://github.com/minetest/minetest/issues/2666 -- RotateAlongYAxis: For facedir case, return if param2 >= 4 by paramat 02:30 hmmmm sure, looks good. 02:31 paramat cool 02:41 paramat now pushing 2666 02:46 paramat push complete 02:46 hmmmm remember to flush the toilet 02:52 paramat hehe 16:47 VanessaE https://github.com/minetest/minetest/issues/2667 16:54 rubenwardy Do any devs actually work on minetest_game? 16:54 VanessaE a few do 16:54 Krock Currently, it doesn't look very active 16:54 VanessaE paramat is most active 16:54 rubenwardy There are a lot of good PRs (bug fixing) that haven't been merged 16:54 rubenwardy game#200 16:54 ShadowBot https://github.com/minetest/minetest_game/issues/200 -- Make furnaces burn items even when they are unloaded by Novatux 16:54 rubenwardy no, game#395 16:54 ShadowBot https://github.com/minetest/minetest_game/issues/395 -- Make furnace work even when unloaded by Kalabasa 16:55 rubenwardy game#499 game#489 16:55 ShadowBot https://github.com/minetest/minetest_game/issues/499 -- Optimize all sounds by kilbith 16:55 ShadowBot https://github.com/minetest/minetest_game/issues/489 -- Rename to Minetest Game (fixes #480) by BlockMen 16:55 Krock same topic - other pull :3 16:55 kilbith making VanessaE or Calinou _game maintainers with half-voice on PRs would revive that repo 16:56 kilbith they have experience on that 16:57 VanessaE eh. 16:57 rubenwardy They say the aim of Minetest is "customisation" but even that isn't done properly. It's only fully customisable for subgame developers, mods can't override subgames properly. Even then, most players can't make mods. I'd love to see more modularity, like being about to download decorative items individually rather than getting loads of redundant ones in a mod. 17:05 VanessaE I'm content to maintain my own game and pull stuff from minetest_game as needed 17:06 VanessaE trying the furnace patch in dreambuilder though 17:20 VanessaE https://github.com/minetest/minetest_game/issues/395 seems to work well on dreambuilder ;) 17:32 paramat hmmmm this is interesting and concerning #2665 checking watershed the 2D perlinmaps have z = 1 as is correct 17:32 ShadowBot https://github.com/minetest/minetest/issues/2665 -- get_perlin_map consumes all memory and brings down servers 17:43 hmmmm ouch 17:43 hmmmm sounds crazy 17:43 hmmmm FWIW, you do use a lot of 3d noise maps :) 17:44 hmmmm "If I understand correctly, a single perlin map should only contain 343KB of data. 777*16 * 64bit data = 343KB. " 17:44 hmmmm where did this guy these numbers from... 17:51 hmmmm I could (theoretically) make memory usage of perlin noise more efficient 17:52 hmmmm it would become much more awkward to use though :) 17:54 hmmmm also that guy is full of shit.. the theoretical maximum of the memory consumption from those noise params would be 36 MB 17:54 paramat yeah i do, riverdev is worse =) 17:54 hmmmm sounds to me like he edited them to have insane values 17:55 hmmmm like maybe "oh more detail is good, I'm going to increase number of octaves to 20" 17:58 paramat "it would become much more awkward to use though" then of course, please don't 17:58 paramat yeah so i'll label it unconfirmed bug 18:00 hmmmm i feel that if something is technically superior but more difficult to use, it should still be there so the willing may use it 18:00 hmmmm we can have a line-by-line buffer and have the caller interpose perlin noise calls with they functions they use the noise in 18:00 hmmmm so it'd be like 18:01 paramat sure 18:01 hmmmm for (s16 y = min; y <= max; y++) { float buffer[80]; noise_foobar->getPerlin3D_line(buffer, x, y, z); for (s16 x = min.X; ...) 18:02 hmmmm all you REALLY need are two points from the noise lattice to avoid recalculation 18:02 hmmmm the current interface is there for convenience 18:02 hmmmm i never saw memory consumption be a problem though 18:03 paramat ah i see, that's reasonable 18:04 hmmmm oh jeez 18:04 hmmmm i know, it's the lua table that's consuming the majority of the memory 18:04 hmmmm but it shouldn't be taking up gigabytes... no way 18:05 hmmmm how on earth do you measure consumption of lua? 18:06 paramat no one else has had this issue with my monster mapgens 18:07 hmmmm it's funny, i profiled this a while ago. writing the noise values to a lua table takes up about 10x as much CPU as generating them does 18:13 paramat nore, sfan5 see https://github.com/minetest/minetest_game/issues/459#issuecomment-98395281 could i have permission to add the mgv5/v7 biome system in a few days? then soon after i'll sort out the sapling growing 18:14 hmmmm why is it 0.5.0 18:14 hmmmm did we not already agree the next major version would be 1.0.0?? 18:17 paramat yes, whatever, my mistake 18:18 hmmmm it's a more general question. we first said 1.0.0 then somebody said 0.5.0 and now there's an 0.5.0-dev branch in upstream 18:18 hmmmm and now everybody refers to it as 0.5.0 18:18 paramat some call it 0.5.0 so i used that 18:22 paramat i guess like me, many are just repeating what they see written 18:22 hmmmm so what do you like better? 1.0.0 or 0.5.0? 18:22 hmmmm the numbering system is FUBAR and it'll never reach 1.0.0 if we continue along this path 18:24 paramat personally i like 0.5.0 because 1.0.0 sound too 'finished' hehe 18:24 paramat (sounds) 18:24 nore paramat: I see no problem with adding it, if it works :) 18:24 paramat that or 5.0.0 18:25 paramat thanks 18:25 hmmmm we thought about 5.0.0 but people were against it because the sudden leap would confuse everybody 18:25 nore did we not already agree the next major version would be 1.0.0?? <-- I haven't seen that... but that's good! 18:26 hmmmm a couple weeks back we had a talk about it 18:26 hmmmm erm, months rather 18:26 hmmmm I don't think the next version is going to be this major version though 18:26 hmmmm a defining feature of the next major version is the introduction of client-side scripting 18:27 paramat yes 0.4.13 next, and soonish since there's been so much mapgen progress 18:27 hmmmm huh 18:28 hmmmm i wonder if there's enough material for another release 18:28 hmmmm if we had a changelog from 0.4.12 to right now, what would it look like? 18:28 paramat erm midsummer .. june 21st perhaps 18:28 hmmmm first of all, the bugs that currently exist need cleaning 18:29 hmmmm i think it's plenty clear at this point that nerzhul will not fix the player damage flash so we'll fix that 18:29 paramat shadow glitches fixed and snow biomes in mgv6, mgv5 and mgv7 stable and with biomes 18:29 hmmmm the git-describe tag issue will need fixing too 18:30 hmmmm there are so many bugs 18:30 hmmmm just go to the issue tracker and scoop up a handful 18:30 hmmmm i need to figure out the mac osx thing too 18:31 paramat nerz is apparently busy with RL stuff so the HP bug may need fixing by someone else 18:34 hmmmm i was busy with real life a lot and i still found time to fix large bugs 18:34 hmmmm that one time there was a crash on 0 length packets in connection, i was 30 minutes late for work but i fixed it 18:35 hmmmm though that was a pretty bad vulnerability 18:35 hmmmm anyway how about the flat player bug? 18:44 paramat this sloped landscape looks good https://forum.minetest.net/viewtopic.php?p=177683#p177683 18:46 hmmmm hmm 18:46 hmmmm it does look cool but it's ultimately unusable until the nodebox lighting bug gets fixed 18:47 hmmmm also i wonder how that guy added the slopes... the simplest method i can think of is to generate map like normal, and then do another pass to replace gaps with angled nodes 18:47 hmmmm but then you have the cutoff-at-y=47 problem 19:01 Sokomine hmmm: yes, transfering huge amounts of data from the engine to lua seems to be expensive. getting the mapgen data and writing it back is a considerable time factor regarding lua mapgens 19:04 Sokomine fixing the nodebox highlighting bug might cause problems. we sometimes use it to draw with light on purpose 19:06 Sokomine hmmmm: i've already considered reading and writing only part of the mapchunk in order to increase speed. it's not too slow in general - even my phone can run the lua mapgen. but still - saving time would be nice 19:47 hmmmm i have no idea if it would increase performance breaking things up into smaller chunks, but i'll try 19:47 hmmmm in theory the larger chunk would be faster because it's the same amount of work with the lowest amount of overhead and optimal cache coherency