Time Nick Message 00:00 red-001 I don't think meshs are need for trapdoors? 00:01 sofar nope, they work fine as is 00:01 sofar the door mesh is nothing more than a glorious nodebox anyway 00:01 sofar it needs to be a mesh only because nodeboxes can't texture larger than 1x1x1 00:02 sofar pushed a fix for the bad light inside the upper door node 00:02 sofar it's better now 00:03 red-001 why was check_player_priv renamed? 00:03 sofar because I intend to use it for other ways to access the node as well 00:03 sofar so the name can be more generic 00:04 sofar most of the code was actually entirely rewritten 00:04 sofar I started out from blank 00:04 red-001 so made can_access 00:04 sofar I wanted to make sure I wasn't copying old code that needed cleaning up 00:05 sofar yeah, it's just a name anyway 00:05 sofar the "access" priv needs to go into this as well 00:05 red-001 yes 00:05 sofar since that's about to get merged into minetest_game 00:05 sofar plus I need to somehow mod in my keys mod as well 00:05 sofar s/mod/patch 00:06 * red-001 needs to start up github for windows 00:06 sofar are steel doors not diggable by hand? I thought they were... 00:07 red-001 shouldn't can_dig be local to the file so the code is not duplicated? 00:07 sofar it is local 00:07 sofar ah you mean, defined only once? 00:07 red-001 yes 00:08 sofar then you'd have to lookup nodedef each caller 00:08 sofar not that that's so bad 00:08 sofar yes, I had it that way first 00:08 sofar there's not enough doors in the game to make a big difference 00:32 * red-001 needs sleep 00:45 paramat bunch of minor mapgen fixes i will merge soon https://github.com/minetest/minetest/pull/3556 00:47 paramat all trivial 00:48 sofar yeah, pretty much 01:29 paramat thanks for looking, now merging 01:39 paramat merged 01:42 paramat next i would like to improve (non-mgv6) caves by using a method used in c++ valleys mapgen: placing biome material on the cave floor in cave entrances, no longer bare and ugly and more a part of the world surface 01:43 sofar paramat: any time to look at the aspen/apple noise params? 01:45 paramat i need to consider those, should appletree and aspen be equal in distribution or should aspen be fewer in number? 01:45 sofar as I made it, aspen are a smaller percentage 01:45 sofar the median is slightly below the middle 01:46 sofar I found reasonable spreads whereever I looked, I'd say it's 25-75 or 33-66 or so 01:46 paramat i'm thinking some areas pure appletree, many areas mixed, should there be pure aspen areas though? 01:47 sofar at the far end of the probability, there will be almost 95% aspen 01:47 sofar so that's pretty much an aspen forest exclusively 01:47 sofar it'll be rare 01:48 sofar paramat: I've noticed that there's a huge grassland biome without any trees 01:48 sofar don't know which one it is 01:49 sofar but I was wondering if we could introduce random, small clusters of aspen in those 01:49 sofar very sparsely planted 01:49 sofar anyway that's a different patch 01:50 paramat ok now looking at noises 01:51 sofar toddler dinner time 02:14 paramat ok made the world all-deciduous forest. i feel appletrees should dominate and many areas should be appletree-only, then other areas a varying mix. 66% 33% by number seems good. will tune noise params 02:47 paramat ok added my suggestion to the thread. world is feeling richer due to the logs and mix of trees 02:52 paramat now it's my feeding time bbl 02:52 kaadmy imo apple trees should be very spare, and be rare. if hunger's ever added, you can just collect several stacks of apples 03:32 sofar kaadmy: "apple trees" are just the oak trees name 03:37 kaadmy ah 03:45 sofar uhhh 03:45 sofar how come seed doesn't change if I delete map* in a world folder? 03:45 sofar o_O 03:46 kaadmy there's a cfg file, world.txt iirc that has the seed 03:47 sofar map_meta.txt has it 03:47 sofar and I deleted it 03:47 sofar but every time I started that world again, it kept the same seed 04:34 sofar paramat: would you mind if I clean up 'saveschems' and fork it to minetest-mods? 04:34 sofar paramat: it's way too useful, we should really try and take care of it together 04:35 paramat ok go for it 04:37 sofar I will try and convert the existing schems to the format I used for aspen, since it's so darn easy to make trees that way 04:37 sofar hopefully it'll help others 04:38 paramat i have a local copy that gets changed a lot, the version at my github is probably a little behind 04:38 paramat ok 04:41 sofar ok, let me do some cleanups on it first. 04:41 sofar you know what, lowering the trunk x2 actually looks pretty decent 04:41 sofar so, I'll just do that 05:13 paramat ok 13:51 RealBadAngel http://i.imgur.com/fpSRahc.png 13:51 RealBadAngel live meshes in formspecs 13:52 RealBadAngel lotsa of inventory images code to be trashed soon 13:53 est31 I guess its connected with an fps drop 13:53 est31 but I'm too tired to discuss with you 13:53 RealBadAngel rather no drops 13:54 est31 (btw I havent tested whether paramat's bugreport is true whether there is a real fps drop from your commit) 13:54 est31 but drawing fixed images is far more easier than drawing meshes 13:54 est31 at least I assume 13:54 RealBadAngel main goal of this one is to get rid og preload item visuals 13:55 RealBadAngel so theyre not needed anymore, huge save on ram 13:55 est31 thats probably true 13:56 est31 but paramat will be against it I guess simply because it comes from you 13:56 RealBadAngel idc 13:56 est31 he'd want you to not do meshes but instead place single faces of them 13:56 RealBadAngel he may be conservative and keep to mt spirit 13:56 RealBadAngel i wont 13:56 est31 because its "minetest style" 13:57 RealBadAngel im coding modern game 13:57 RealBadAngel not the 80's 13:57 nrzkt RealBadAngel, i prefer having a formspec thing to have mesh viewer rotating the mesh :D 13:57 est31 even in the 80's it was totally ugly to place single face 13:57 RealBadAngel nrzkt, look closer 13:57 nrzkt and let users use this new formspec thing to do more things 13:57 RealBadAngel the meshes are rotating 13:57 est31 the meshes are rotating 13:58 est31 paramat will -1 it, and c55 will close it 13:58 RealBadAngel atm i let selected or hoovered ones to rotate 13:58 est31 just because they are rotating 13:58 RealBadAngel lol 13:58 RealBadAngel no problemo 13:58 RealBadAngel that will land in minetest HD then 14:00 RealBadAngel to make things clear, i dont completely care bout paramat's or c55's opionions 14:00 est31 probably its really the best idea if you forked it 14:00 RealBadAngel and i will code things only based on my vision and taste 14:00 RealBadAngel i dont wanna fight 14:01 RealBadAngel theyre having different one, so its ok 14:01 est31 minetest is open source, forking is part of open source life 14:01 est31 closed source games cant fork they have only one version 14:01 est31 minetest can have hundreds of versions and tastes 14:01 RealBadAngel ofc 14:02 RealBadAngel but in the long run mainstream will suffer 14:02 RealBadAngel but thats not my problem 14:02 est31 probably yeah 14:02 est31 minetest should be split up imo 14:02 est31 one lib for the network 14:03 est31 one for the graphicd 14:03 est31 graphics* 14:03 est31 etc 14:03 est31 then forks can keep common parts 14:03 est31 and improvements of e.g. network can be shared between all 14:07 est31 and the deserialisation and database code can be used by minetest mapper scripts 14:10 RealBadAngel est31, that may be propably good idea 14:10 RealBadAngel but im talking about replacing old, faulty and dead slow code 14:10 RealBadAngel and replacing it with native meshes 14:11 est31 you want to render them in real time 14:11 RealBadAngel i do 14:12 RealBadAngel 100 nodes rendered on one frame more? 14:12 RealBadAngel is that a big problem? 14:13 RealBadAngel drawtime thingy doesnt move with that 14:31 RealBadAngel nrzkt, can you edit your comments on https://github.com/minetest/minetest/pull/3547 ? 14:32 RealBadAngel defining functions virutal for that class is not a subject of this PR 14:32 RealBadAngel you made look this PR buggy while its not 14:36 RealBadAngel btw, dead code to delete: https://github.com/minetest/minetest/blob/master/src/content_cao.cpp#L136 14:37 RealBadAngel and https://github.com/minetest/minetest/blob/master/src/content_cao.cpp#L280 14:37 RealBadAngel both those CAO's are not used 14:38 RealBadAngel (items are genericCAO) 14:45 nrzkt RealBadAngel, done 14:47 RealBadAngel k 18:13 paramat hmmmm celeron55 and all, any comments on the suggested fix for this important issue? https://github.com/minetest/minetest/issues/3530 '21-25% FPS drop with shaders disabled due to vertex tangents commit' ie split the renderer 18:16 * Fixer hears echo 18:17 RealBadAngel paramat, i said i can do that when needed, what for you need c55 at all? 18:17 RealBadAngel have you excavated another box without shaders support? 18:18 paramat since it's important issue i'm asking for other expert opinions 18:19 RealBadAngel no need to repeat that opinion each time im on the channel 18:24 paramat there's no opinion, just a question. the suggested fix needs discussion 18:32 paramat i decided to discuss this before i saw you were on-channel 18:32 RealBadAngel im reorginizing now that stuff, i can do that 18:33 RealBadAngel mainly, i can split things into textures (non shaders.old way) and meshes, native irrlicht way 18:35 RealBadAngel but frankly i am afraid old way will be slower 18:35 RealBadAngel it is already 18:36 RealBadAngel thx to that render-to-texture magic way 18:36 RealBadAngel if something is native in engine, why to use magic tricks to show it? 18:37 RealBadAngel if you cant, it means you just dont know how to use the engine :P 18:38 paramat i see. the main issue though is fps with shaders off, i guess the slowness of formspecs won't eliminate the 25% gain in fps 18:39 RealBadAngel main goal of the meshes for items change is to free ram 18:40 RealBadAngel now each item needs extra texture to be shown 18:40 RealBadAngel more items, more ram needed 18:41 paramat i see. do you need vertextangents for inventory meshes? 18:41 RealBadAngel also, to get the texture, costly process was called 18:41 RealBadAngel i do need that vectors for almost everything 18:41 RealBadAngel its like you have asked me do i need sun 18:41 RealBadAngel yes i do 18:42 RealBadAngel do you need mapgen noise gens? or other stuff? 18:44 RealBadAngel such things are basic for everything with textures 18:45 RealBadAngel nothing is possible without it 18:46 RealBadAngel i had to cheat to get absudly costly tanget space before 18:46 RealBadAngel i mean calculated out of 2 vertices only, in shaders 18:46 paramat yes 18:48 paramat before splitting the renderer lets get some other devs' input on the idea 18:53 hmmmm oh no please don't split that 18:54 hmmmm the better idea would be to find the root cause of the slowness instead of splitting things into slow and slow-but-not-quite-as-slow 18:54 hmmmm the best idea would be to use a different god damn graphics engine though 18:55 RealBadAngel hmmmm, i saw you mentioned that some time ago but i afraid its not that easy 18:55 RealBadAngel each and every triangle is using just more data 18:56 RealBadAngel 2 more vectors per each vertice 18:56 RealBadAngel thats why it is slower 18:57 RealBadAngel its all about numbers 18:57 paramat since you (RBA) seem to be starting a HD branch or fork, it might be simpler and less work for you and everyone just to revert vertextangents in master 18:57 RealBadAngel more we "see" more data we got 18:57 RealBadAngel paramat, you may want to trash all my commits, youre free to do so 18:58 RealBadAngel just go on or shut up 18:58 RealBadAngel other way youre playin flat world madman just 18:59 RealBadAngel and frankly you shall start branch for museum workers all around globe :P 19:00 sofar you guys remind me of "the red and the blue" claymation 19:05 paramat heh. i'd prefer to keep vertextangents but also fix the non-shader slowness, if that's possible. reading the forum there are players who are, erm, very keen to see their fps return 19:06 RealBadAngel as i said, i can do that but only by separating 19:06 RealBadAngel fancy mind tricks using templates are out of discussion 19:07 RealBadAngel if something is good for all cases it means is worth shit 19:08 RealBadAngel simple, old, gold rule 19:14 paramat okay. well, hmmmmm suggests a split renderer may not be the thing to do, and i somewhat trust his judgement 19:14 RealBadAngel its not his code, he havent touched a sigle key coding it 19:15 RealBadAngel his knowledge is next to google one 19:15 RealBadAngel same as yours :P 19:15 RealBadAngel but you know the best 19:16 RealBadAngel what if i make a commit with separation and you will try it just? 19:19 RealBadAngel meanwhile, https://github.com/minetest/minetest/pull/3547 19:20 paramat well, i'm asking for consideration of the method before you start a lot of work. i do tend to trust c55 and hmmmmm more than you even on graphics 19:20 RealBadAngel paramat, c55 lately is using way more advanced methods in his farmesh branch 19:20 RealBadAngel say hello to VBO code 19:21 RealBadAngel and the same way say good bye to your low end boxes 19:23 RealBadAngel frankly, demand to be compatible with something more than 10 yrs old in gaming is like demand to be compatible with dinosaurs 19:23 RealBadAngel sorry, i dont like to fuck with them l;) 19:24 RealBadAngel OpenGL 1.2 19:24 RealBadAngel Release date: March 16, 1998 19:25 paramat my box is an intel core i5 with intel haswell integrated graphics :) 19:25 sofar what does android have nowadays? 19:25 RealBadAngel wonder if all the fridges and washing machines support 1.2 already 19:26 RealBadAngel some of the watches already do 19:26 RealBadAngel saw minetest running on the watch already 19:26 Fixer even my crappy laptop with Pentium-M 2GHz with Intel Integrated has OGL 2.2 iirc 19:26 Fixer and it already has red blocks issue 19:30 RealBadAngel fixer, im am developing now separate shaders for both GLSL and ES 19:30 RealBadAngel they have to be different 19:30 sapier great :-) 19:31 sapier so we might be able to use ogles2 soon? 19:31 RealBadAngel no way to have same for both 19:31 RealBadAngel yes 19:31 RealBadAngel i am able to use shaders for postprocessing for GLSL and ES already 19:31 sapier on the other hand as of speed there doesn't seem to be a difference, at least on my device 19:31 RealBadAngel @paramat, its called SEPARATION 19:42 paramat i'm just trying to discuss calmly and sort this out, it's tough work with your current attitude =P 19:44 Calinou what does android have nowadays? 19:44 Calinou OpenGL ES 3.0, 3.1 19:44 Calinou (well… Irrlicht is stuck on ES 2.0 since forever) 19:44 Calinou <+RealBadAngel> no way to have same for both 19:44 Calinou we could switch to ES 3.0 for everything, it's plenty 19:48 sapier Calinou: we use ogles1 and as I mentiones about 15 minutes ago I don't see any benefit on ogles2.0 atm do you know about any benefit 3.0/4.1 might have? 19:49 sapier just a higher number ain't exactly a valuable benefit ;-) 19:51 Calinou Irrlicht has support for OpenGL ES 2.0, it's even mentioned in the Minetest settings 19:51 Calinou sapier, having shaders? 19:51 Calinou OpenGL ES 1.0 is fixed-function, nobody should use it anymore. 19:51 paramat input on this is needed, but my feeling is it's best RBA starts a branch or fork, and we revert vertextangents here 19:51 Calinou OpenGL ES 2.0 is still very limited (source: reduz, the lead developer of Godot) 19:51 Calinou OpenGL ES 3.0 is liberating compared to 2.0 19:53 sapier Calinou: many ppl disable shaders on pc too ;-) 19:54 Calinou you should work on OpenArena, probably 19:54 paramat btw c++ valley's mapgen may be merged soon, last chance for reviews or objections https://github.com/minetest/minetest/pull/3508 < hmmmm 20:01 hmmmm I can't find any real objections 20:01 hmmmm if you guys like it and it works well, then so be it 20:02 hmmmm lots of copying from mapgen to mapgen though, I'll try to see if I can't generalize a couple of the most repetitive functions sometime 20:03 paramat ok thanks 20:03 hmmmm love how there's exactly 1,337 additions - wondering if duane intentionally did that 21:01 RealBadAngel hmm, showin native meshes instead of textures is circa 500 lines code less 21:02 RealBadAngel is that enough for you to like it? 21:03 RealBadAngel you have always pointed me to accomplish the task with minimal code needed 21:08 RealBadAngel anyway im pretty close to paramats demand 22:10 celeron55 what 22:31 RealBadAngel celeron55, im talkin about implementing advanced stuff, like VBO for example 22:31 RealBadAngel youre smuggling it in without any comments 22:32 RealBadAngel do you expect me to collect all the "non mt spirit" comments for ya? 22:33 celeron55 what the hell now again 22:33 RealBadAngel not any hell, please calm down 22:33 celeron55 i am not doing anything whatsoever at the moment 22:34 celeron55 and that branch will not be merged without further reviews and tests 22:34 RealBadAngel i am trying now to replace textures with meshes 22:34 celeron55 sounds dumb 22:34 RealBadAngel you know meshes are native 22:34 celeron55 textures aren't native? 22:35 RealBadAngel making scene to get a texture is wise? 22:35 RealBadAngel when you can just display it? 22:36 celeron55 well i don't really care about that 22:36 RealBadAngel ofc you dont 22:36 RealBadAngel thats why its fucked up 22:36 celeron55 make good code and don't create regressions for older hardware or something and i guess it's fine 22:37 RealBadAngel im polishing the code now, need a few days 22:37 sofar I think code speaks 22:37 RealBadAngel scales are weird for any case 22:37 RealBadAngel thats my problem 22:38 RealBadAngel i just raised the idea becuase i know it wil work 22:38 RealBadAngel and i love the selected item to spin :) 22:57 RealBadAngel sofar, howgh 22:57 sofar howgh to you too 22:57 RealBadAngel but ive seen that too many times 22:58 RealBadAngel mt is F16 flying bombmers codebase 22:59 RealBadAngel i mean "dont touch" 22:59 sofar fuck, the f16 is one of the best fighter jets in the air still 22:59 RealBadAngel ask hmmmm 23:00 sofar they've been trying how long to make a cheap, better replacement for it? and failed how many times? :) 23:00 RealBadAngel dont ruin my joke 23:00 sofar you may need to adjust metaphors ;) 23:00 RealBadAngel hmmm is involved here :P 23:00 sofar point taken, though 23:01 RealBadAngel anyway, we need to advance 23:02 RealBadAngel endless fiddling with bugs is a no waY 23:02 RealBadAngel i do really love your code on the custom drawtypes 23:03 sofar the fences? 23:03 RealBadAngel faster and saves shitload of ram 23:03 RealBadAngel connected nodeboxes 23:03 sofar I'm going to try and finish that one up this week 23:04 sofar I had it on a backburner for a bit thinking about other stuff 23:04 RealBadAngel but one con 23:04 sofar but I think I can finish it 23:04 RealBadAngel youre good to go with pipes, tubes etc 23:05 RealBadAngel you will fail with mesecons climbing walls 23:05 sofar yeah, mesecons is semi-special since it attempts to mimic minetest redstone 23:05 RealBadAngel thats why i told u to contact nore 23:05 RealBadAngel he made that math 23:05 sofar and redstone can connect to 9 other nodes, not 6 or 5 23:05 sofar frankly, I don't think I want to fix it 23:06 sofar I'd be happy if it connects 6-ways 23:06 RealBadAngel if you wont i will do that 23:06 RealBadAngel thats crucial imho 23:07 sofar there's no 3d model you can make that can cross an infinitely small line with a density larger than 0 23:07 RealBadAngel since a few weeks im takin all the steps ive promised earlier 23:07 RealBadAngel one by one 23:07 sofar anyway, premature 23:07 sofar maybe we will figure out a way 23:08 RealBadAngel just ask nore 23:08 sofar but first i have to implement the matching API 23:08 sofar way more important 23:08 RealBadAngel its a real fucked up math 23:08 RealBadAngel see open source redstone 23:09 RealBadAngel creator of it, female, said she went drunk one day 23:09 RealBadAngel next day she woke up she have workin ones 23:11 sofar wait, who? 23:13 RealBadAngel eloraam 23:14 sofar sorry, I don't know the old coders of lore ;) 23:15 RealBadAngel eloraam is creator of RedPower 23:15 RealBadAngel she left the scene in sake of creating own game 23:16 RealBadAngel some folks recreated the mod called RedPower 23:17 RealBadAngel and it is open source 23:18 RealBadAngel i am coder, 40+, i miss ability to create complex logic machines in mt 23:18 RealBadAngel in mt ive build a tree farm, it took me a few months to build 23:19 RealBadAngel shitload of wires and gates 23:19 RealBadAngel mt is not capable of running such shit at all 23:20 RealBadAngel in mt it is possible to build emulations of 8 bit CPUs even 23:20 RealBadAngel thats what makes MT epic 23:21 RealBadAngel EPIC even 23:21 sofar I had not suspected you to be an old fart like me... ;) 23:21 RealBadAngel i am 23:22 RealBadAngel i was coding candies before most of contributors to mt was even born 23:22 RealBadAngel so i have right to shout 23:23 sofar I prefer to use less words and confuse the less wise with well-placed cryptic messages 23:24 Megaf funny, you all still adding more and more stuff to Minestest, I forked Minetest and all my work so far was removing all that stuff 23:24 RealBadAngel youre pretty new to this place 23:25 Megaf my fork + my subgame = 9.5M 23:25 sofar I wasted a year or two on Bukkit 23:25 Megaf compiled 23:25 RealBadAngel you will sooner or later get to the point i am 23:25 RealBadAngel total frustration 23:25 RealBadAngel i am devoted to mt 23:26 Megaf What I have removed so far, stupid useless bitmap fonts, mapgen flat, v5, v7 and soem other 23:26 RealBadAngel 4 yrs already 23:26 RealBadAngel youre new 23:26 RealBadAngel so pick 23:26 Megaf I have removed support for Redis, LevelBD and Dummy 23:26 Megaf there is really lots of redundant stuff in Minetest 23:27 Megaf and, RealBadAngel, shaders is disabled on my fork, and so is minimap 23:27 sofar I love the minimap 23:27 Megaf and I'm having a hard time to find all related files/line to that stuff to properly remove it 23:27 sofar I could love bumpmapping but it's so darn hard to make textures work well, and I don't like the default textures much 23:27 Megaf All that I mean is. stop adding features to minetest 23:28 Megaf and begin removing old/redundant code 23:28 sofar nothing wrong with a stripped down fork imho 23:28 Megaf fixing the BLOODY MEMORY LEAKS 23:28 sofar well those fixes should go upstream 23:28 Megaf you can't even fix old bugs and are adding 3D support for fecks sake 23:29 Megaf maybe make a good branch for Minetest Stable and a branch called Minetest Experimental where all that stuff can be tested 23:29 sofar that's already been proposed and accepted 23:29 sofar so in reality you can start pushing your cleanups and leak fixes to master 23:30 sofar I suppose RealBadAngel actually should create the branch for his work already 23:30 Megaf sofar, I will not 23:30 Megaf zeno has done some clean up, and so sapier did, but they seem like lone players, with not much support from other devs 23:31 sofar I suppose they need your help? 23:31 Megaf $ du -hs minetest/fonts/ 23:31 Megaf 15M minetest/fonts/ 23:31 Megaf 15 MB in bloody ttf fonts 23:31 sofar why stand at the sidelines screaming foul if you are wearing the team outfit? 23:31 Megaf I'm just asking that people do basic and easy stuff 23:31 Megaf step by step 23:31 sofar it's OSS, don't ask. You can do. 23:31 Megaf and stop trying to bloat up Minetest with these features 23:32 Megaf sofar, I can not help, I can't even code, yet I managed to do this cleanup 23:32 Megaf s/can/can't 23:32 Megaf ops 23:32 Megaf can not is correct :) 23:32 sofar you obviously can 23:32 sofar it really, doesn't, take, much 23:33 Megaf sofar, I already tried to proposed changes, did submit pull requests 23:33 Megaf but all the time minetest devs seems to know more, know whats better 23:33 Megaf I've seen two devs that wanter to reduce code, optimize and hunt bugs, sapier and zeno 23:33 Megaf and that's about it 23:34 Megaf zeno more than sapier perhaps, sapier came up with the 3D support 23:34 sofar I don't think anyone is against optimizing and reducing bugs 23:34 Megaf it works great indeed, yes it does and I used it a lot. But for now, is just bloatware 23:35 Megaf sofar, how many people you think will be in favour of removing mapgen flat, v5, v7 and the other that I don't recall the name? 23:35 Megaf singlenode 23:36 sofar ah, that's a very good point 23:36 Megaf all mapgens can make flat terrain, that makes mapgen flat redundant, v6 was created only because v5 had a bug, but it had nice features so they decided to keep it 23:36 sofar I can see v5 getting removed 23:36 Megaf v7 was created because v6 had a bug, it couldn't make mountains 23:36 Megaf sofar, is bloat on top of bloat 23:36 sofar totally agreed 23:37 Megaf now, think about of hours of work put into making each magen 23:37 rubenwardy singlenode has a use. v5/v6/v7 has purposes, however it should be easer to remove them for a build 23:37 rubenwardy the dummy backend is used by CTF and is very fast 23:37 sofar more #ifdefery ? 23:38 Megaf now, about databases, sqlite is fantastic, and it needs very few adjustments to make it faster and use less space for the db, like enabling vacuum 23:38 rubenwardy and useful for testing mods 23:38 rubenwardy database backends are more of a server thing, and shouldn't be included in the client by diffault (other than a good all round one) 23:38 Megaf rubenwardy, on my tests dummy wasnt any faster than sqlite properly set up 23:39 sofar I thought only sqlite is enabled for the client by default? 23:39 Megaf in real life tests I eman 23:39 Megaf mean*, actually playing on a server using dummy 23:39 rubenwardy ~web title https://github.com/rubenwardy/capturetheflag/pull/36 23:39 ShadowBot rubenwardy: Recommend the Dummy backend by asl97 · Pull Request #36 · rubenwardy/capturetheflag · GitHub 23:39 Megaf and that's other thing I question, do devs actually play Minetest? 23:40 rubenwardy sqlite3 (syncronous & ramdisk) 1.105558 dummy 0.012 23:40 rubenwardy actually, that's only for deleted 23:40 rubenwardy o 23:40 rubenwardy acta 23:40 sofar I'm no dev, but yes, I actually play 23:40 rubenwardy actually, it's only faster at deleting, not at editing 23:41 rubenwardy Most of this could be solved by adding C++ mod support - core mapgens could be compiled modules 23:42 Megaf I'm developing my own fork now, I have been running MT servers for about 4 years and playing as long 23:42 rubenwardy less bloat is good, I guess. 23:42 Megaf other thing, Lua is great indeed, but Python *might* be better 23:43 rubenwardy heh 23:43 Megaf and now, contradicting myself, why not add support for modding done in Python too? Keep the lua thing, and python, Irrlicht already supports it AFAIK 23:44 rubenwardy "Irrlicht already supports it" 23:44 rubenwardy what? 23:44 rubenwardy Yes, there may be wrappers allowing you to add scene nodes etc 23:44 rubenwardy But that would only be applicable to client side modding, not to Minetest's server mods 23:45 Megaf ok, unofficial biding for Python https://en.wikipedia.org/wiki/Irrlicht_Engine 23:45 rubenwardy Where is your fork, Megaf? megaf/minetest? 23:45 Megaf rubenwardy, nope, is not even on Megaf account 23:45 Megaf I don't want to make it public just yet 23:46 rubenwardy I think that the minimap is useful for gameplay if they made it better. Ie, the ability to require the crafting of an item to use. 23:46 Megaf and at the moment it is very dirty, the dbs/magen removal was very dirty (since I can't code) 23:46 Megaf rubenwardy, I like the minimap, a lot actually, I just think is not the time to have it 23:46 Megaf it could have been done via a mod perhaps 23:46 rubenwardy slow 23:46 Megaf not adding compiling time and memory use 23:47 Megaf so add features, yes, but make room for it first 23:47 Megaf or something line that 23:47 sofar client-side modding has been desired for a long time 23:48 sofar it could certainly render a minimap if done correctly 23:50 Megaf sofar, bottomline is, I'm kinda frustrated with Minetest and I don't see my suggestions getting accepted by Minetest dev team anytime soon, so I'm forking it, selling it, making money from it, and donating part of the money back to Minetest 23:50 Megaf so far* 23:51 sofar so assuming you're properly adhering to the licenses, I'd say you've found a perfectly valid way of contributing to the project in your own way 23:52 Megaf I'm not sure about the licenses, but I am giving all the credit to Minetest and c55 23:53 Megaf I didn't create the thing, just made it smaller and faster 23:53 rubenwardy You also have to share the full source code of the engine to anyone that buys/has a executable version of your fork 23:53 sofar let me buy a copy from you :) 23:54 Megaf rubenwardy, the source code is already available and will always be 23:54 Megaf as long as Minetest is Open Source 23:55 rubenwardy your version needs to be too 23:55 Megaf I'm charging only for my work hours and the compilation 23:55 Megaf rubenwardy, that's what I mean, my version will be Open Source and has it's code available as long as Minetest is too 23:55 Megaf and Irrlicht, and OpenGL 23:56 rubenwardy ? 23:56 sofar well no 23:56 rubenwardy I don't understand what you mean 23:56 sofar you are not allowed to refer to minetest.org for source code 23:56 rubenwardy You have to publish modifications, even if they're removals 23:56 sofar right 23:57 rubenwardy (also, minetest.net) 23:57 sofar irrlicht is not GPL, you don't need to give people the source code for that 23:57 sofar but minetest is 23:57 Megaf ok, rephrasing everything. The source code of my fork is available. Credits do go to Minetest 23:57 sofar do you have it on github? 23:57 Megaf so I am giving the source for my fork and telling where to get the source for Minetest 23:57 Megaf sofar, yep, 23:58 Megaf not on Megaf 23:58 Megaf I have to go now, good night everyone :) 23:58 rubenwardy o.