Time Nick Message 00:37 paramat Sokomine hmmmm #3194 this needs testing though 00:37 ShadowBot https://github.com/minetest/minetest/issues/3194 -- Mapnode: Add rotateAlongYAxisFull supporting 24 facedirs by paramat 01:40 hmmmm scary 01:40 hmmmm if it works, it works then 01:43 hmmmm 00000100, 10000101, 10000010, 100000011 01:43 hmmmm hrmm 01:43 hmmmm I see no pattern at all 01:43 hmmmm i have no insight whatsoever into how 6d facedir works 01:43 hmmmm just saying that it looks downright scary 02:28 Sokomine why scary? 02:30 hmmmm it's completely opaque 02:33 Sokomine for normal facedir, it's still param2 % 4 for the information weather the node is facing north, south, east or west (rotation along the y-axis/vertically). the other values rotate along all the other available axis. people use a screwdriver to rotate the nodes 02:34 Sokomine wallmounted is diffrent and has only 6 values. havn't been able to detect any particular order there 02:35 Sokomine it helps to modify the screwdriver so that it prints out the current param2 value of a node after each rotation. easier to visualisize it 02:38 paramat my method is in the issue #3191 i stared at the table linked to and imagined rotations 02:38 ShadowBot https://github.com/minetest/minetest/issues/3191 -- Schematic rotation support for all 24 node orientations WIP 02:38 paramat if it works i'll be amazed 02:43 paramat also i still have to work on wallmounted 02:45 Sokomine yes. we all are whenever that has to be done and works. getting those tables right is always an issue when importing something 03:41 paramat http://i.imgur.com/9dUn1P9.png so far so good but more tests to do 04:56 celeron55 hmmmm: the 6d facedir values aren't completely unreasonable; they are set up in a consistent way but i can't say i would remember what the way is 06:46 hmmmm oooh 06:46 hmmmm i just triggered a falling object duplication bug 06:47 hmmmm something something to do with a water source surrounded by sand with no bottom 10:04 est31 hmmmm: congrats, SAOs are as we all know too unreliable 10:05 est31 @rotation code discussion: perhaps all this can be put into a single class, with comments explaining it, and methods to deal with it, so that its perhaps a bit easier to understand etc? 10:06 est31 but tbh, I don't even understand yet how screwdrivers work :D 10:07 est31 I'll have to give it all a closer look 10:28 nrzkt est31: wtf 11:02 T4im facedirs are btw. shortly described in: https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L537 11:03 T4im but some abstraction of dealing with them, might probably help everyone :D 11:45 est31 T4im, that explains it 11:45 est31 even though the author of the doc has mixed terms as it seems 11:45 est31 because facedir modulo 4 is a value between 0 and 3 11:45 est31 can never be 4 or 5 11:46 est31 and it precisely describes the two less significant bits 11:48 est31 one of those two things is surely wrong, most likely both are 11:52 T4im 00 = 0, 01 = 1, 10 = 2, 11 = 3 11:52 T4im looks right to me 11:53 est31 what is 4 modulo 4 ? 11:53 T4im what it however does fail to mention, is that they alawys describe the rotation around the same axis (I think it was the original y-axis) 11:53 est31 or 5 modulo 4? 11:54 est31 yea the next question how are the rotations applied 11:54 est31 in which order 11:56 est31 rotation group isnt abelian for n > 2 11:56 est31 for n=2 im not sure but I think it is 11:56 est31 https://en.wikipedia.org/wiki/Rotation_group_SO%283%29 12:00 T4im paramat: looks good (rotated by 180°), https://img.bi/#/cISQtt1!_OvStwhtOFkgtPh5lQFZA4CgWVGchQtIsjZwA4ZZ 12:02 T4im est31: I don't think there's a lot more logic to facedirs than the two lsb's 12:05 T4im if you wanted it to rotate cleanly, yea.. but... well.. it seems just to rotate around the same axis, after you told it, where that one is directed to 12:35 T4im if I'm not mistaken, it isn't really SO(3), and should be abelian: i.e."go into position fd=4 and leave your 2 lsb alone, then rotate by 90° around your original y" is the same as "rotate by 90° around your original y, then go into position fd=4 and leave your 2 lsb alone",.. 12:36 T4im not, that the screwdriver would actually do that (it resets the lsb's each time, I think) 12:39 kahrl no, it's not abelian 12:39 kahrl the group of orientation-preserving symmetries of the cube is isomorphic to the symmetric group S_4 12:40 kahrl http://groupprops.subwiki.org/wiki/S4 12:40 kahrl also https://en.wikipedia.org/wiki/Octahedral_symmetry#The_isometries_of_the_cube 12:41 kahrl so each possible rotation corresponds to a permutation of the four diagonals of the cube 12:41 T4im oh, interesting, thanks 12:41 VanessaE suddenly I'm reminded of that "timecube" guy :) 12:52 est31 yea its a subgroup of SO(3) 12:52 est31 http://groupprops.subwiki.org/wiki/Classification_of_finite_subgroups_of_SO%283,R%29 12:52 est31 4th 12:55 est31 S_4 good to know 12:56 est31 T4im, my main problem with the docs is that the two least significant bits are the same than the number modulo 4. 12:57 est31 e.g. take the number 14, then 14 modulo 10, you have 4, and the "least significant digit" is 4 as well 12:57 T4im that's the idea, isn't it? 12:58 T4im oh, now I get what you mean.. the label is a bit misleading, indeed 12:59 T4im or rather.. yea.. ok, now it confuses me too x) 12:59 VanessaE bbl 13:02 est31 rej good catch 13:08 est31 #3192 is fine for me now 13:08 ShadowBot https://github.com/minetest/minetest/issues/3192 -- Clarify radii and distance types in documentation by t4im 13:10 Sokomine T4im: as far as screwdriver usage goes: left-click to rotate through the four normal facedir values. shift-rightclick to change the rotation axis 13:34 T4im Sokomine: what I meant is, that the "left click" rotation is reset if you would rightclick your way back through it again 13:54 Sokomine T4im: ah, yes. noticed that as well. the screwdriver used to behave diffrently in the past 14:15 est31 anybody can have a look at 3192? 14:15 est31 (second one) 14:17 nrzkt #3192 14:17 ShadowBot https://github.com/minetest/minetest/issues/3192 -- Clarify radii and distance types in documentation by t4im 16:43 est31 pushing in 5 mins is really triviak https://github.com/est31/minetest/commit/4f03f8d119283c16034836d6e8d617961a8150c9 16:43 est31 trivial* 16:43 est31 !title 16:43 ShadowBot lua_api.txt: fix typo · est31/minetest@4f03f8d · GitHub 17:23 kahrl est31: fixing typos in the docs? here's another one - I think ;) https://github.com/minetest/minetest/blob/master/doc/world_format.txt#L315 17:23 kahrl shouldn't name_len be key_len? 17:24 est31 ah yeah thx 17:24 kahrl (yeah, I know I said the commit that added that LGTM) 17:24 est31 and I was the author :/ 17:25 est31 well I can explain how it was created, name wasthe previous version of the "_len prefix", then I thought key fits better and renamed it 17:25 est31 and the rename was incomplete :) 17:25 est31 feel free to push a fix kahrl :) 17:25 kahrl been there done that, many times :) 17:26 kahrl doing an incomplete replace job, I mean 17:27 kahrl pushed 17:31 kahrl oh, and there's also the overzealous replace job: all the occurrences of "self program is free software" in builtin/ are proof of that 17:32 * est31 types grep -r "self program" 17:32 est31 ... and hopes the best 17:57 est31 fortunately the list of files that have this typo is limited to builtin/fstk 17:58 est31 but it occurs three times per file: 17:58 est31 ./builtin/fstk/tabview.lua:4:--self program is free software; you can redistribute it and/or modify 17:58 est31 ./builtin/fstk/tabview.lua:9:--self program is distributed in the hope that it will be useful, 17:58 est31 ./builtin/fstk/tabview.lua:15:--with self program; if not, write to the Free Software Foundation, Inc., 18:00 sfan5 ಠ_ಠ18:05 est31 https://github.com/est31/minetest/commit/1adc7bf5c60830cc8dcbab34c682f7fddde55558 18:06 est31 ok 2 push? 18:12 kahrl est31: lgtm 18:12 kahrl est31: I noticed the typo a long time ago, but never bothered to fix it because I thought it was somewhat funny 18:13 est31 lol 18:13 est31 sorry for ruining your joke xD 18:14 kahrl nah it's okay :) 19:12 est31 T4im, do we really use a manhattan metric for find_node_near? 19:13 est31 I mean we use the result of FacePositionCache::generateFacePosition 19:13 est31 but that one doesn't look like its using manhattan metric 19:20 est31 it more looks like a maximum metric 19:25 est31 added comments 19:25 est31 those with ok were to mark my reviewing progress 19:28 T4im hm, yea you're right 19:47 T4im hmm interesting, does [ci skip] not work for pr's? 19:48 T4im feels wrong bothering travis for textfile changes 19:48 T4im oh wait, that's a jenkins, that's why 19:48 T4im nvm 19:52 est31 thats no moon, thats a jenkins! 19:52 est31 be sure to get not into the range of its testbeam! 19:53 T4im :D 19:54 T4im travis-ci is used then for the actual builds though? 19:54 est31 they are all checkbots 19:54 est31 nightly builds are done by other bots 19:54 est31 all checkbots build from scratch 19:55 est31 the jenkins even builds irrlicht if i remember right 21:35 paramat hmmmm and all #3194 is updated, tested and ready for review 21:35 ShadowBot https://github.com/minetest/minetest/issues/3194 -- Mapnode: Add rotateAlongYAxisFull supporting 24 facedirs by paramat 21:43 T4im paramat: looked good here too, this afternoon: https://img.bi/#/cISQtt1!_OvStwhtOFkgtPh5lQFZA4CgWVGchQtIsjZwA4ZZ 21:44 paramat thanks for testing 21:44 T4im (that's 180° to each other) 21:44 T4im thanks for fixing :D 21:46 T4im or improving perhaps x)