Time Nick Message 00:06 realbadangel fixed issue #1902 00:06 ShadowBot https://github.com/minetest/minetest/issues/1902 -- Meshnodes with facedir 0 have bad smooth lighting 00:16 zat hey devs remove items number limits on cooking recipes! 00:21 realbadangel zat. what for? 00:22 zat so recipes for alloys and such can finally be made 00:23 VanessaE he wants to be able to stick two iron in a single cooking slot I guess, like with technic's alloy furnace 00:23 VanessaE or similar. 00:24 zat iron and copper and make... whats the alloy that resuts from that? :P 00:24 zat or was it zinc and copper 00:24 VanessaE NOTHING reasonable is made from copper+iron 00:24 VanessaE (however two copper + one tin is bronze I think) 00:24 VanessaE (if one is using correct and proper recipes) 00:25 exio4 what kind of simple problem I could write that interacts with the user in a dumb way (reading from console) and does something fancy? 00:25 exio4 hmm 00:25 exio4 oh, this isn't main channel 00:25 exio4 sorry 00:26 zat thing is that I suggest that recipes registering should not enforce limits because 00:26 zat 1.- this would loose binding between recipes and nodes for crafting 00:26 zat 2.- whoever wants to make crafting tables or furnaces WON'T have to code their own framework for alternate recipes 00:31 paramat BTW zat is developing awesome stuff so as long as it's reasonable im backing these requests =) 00:31 zat (lol) 00:36 realbadangel zat, if something can be done easily with a mod it means that engine is flexible enough 00:36 realbadangel and doesnt have to be changed 00:37 realbadangel if that wouldnt be possible, that will be another story and a reason for request 00:37 zat it can be done with the engine, yes, easily? no... its duplicating code 00:37 zat it can be done with a MOD, yes, easily? no... its duplicating code 00:37 realbadangel technic does all that without modifications to the engine, so it is possible 00:39 zat that it is possible by adding a lot of duplicate code, modspace, and duplicating an already existing api in order to add a feature that could easily be added to the engine is possible, does it mean that it is the best? 00:40 realbadangel not really 00:40 realbadangel it means it is already working 00:40 realbadangel and we all have much more important things to do that reinventing the wheel ;) 00:40 realbadangel *than 00:41 realbadangel at some point we could allow that for engine, but now we shall focus on upcoming xmas release and fix more critical issues 00:42 zat what that mod does is working around a lack of flexibility of the engine. does it work? probably it does, still it makes using such features much more cumbersome, starting with the fact that I have to depend on a particular mod in order to gain said feature 00:42 zat realbadangel: I am not asking anyone to make it, I want to know how well accepted the idea would be 00:42 zat so I can implement it... 00:42 realbadangel if you code it i will merge it 00:43 realbadangel enough? 00:43 zat realbadangel: more than enough 00:43 realbadangel so deal 00:45 paramat kewl 00:49 realbadangel paramat, kewl what? 00:50 paramat the above 00:51 zat hint: read above 00:51 realbadangel ah 00:52 realbadangel zat, while you will be adding that feature, think of custom craft types 00:53 realbadangel not only hardcoded ones as we have by now 00:53 zat hows that? 00:53 zat you mean accept custom strings? 00:54 realbadangel no, i mean craft types registration by mods 00:54 realbadangel atm theyre hardcoded in the engine 00:54 zat no no yes I got that 00:55 zat so other than cooking or normal etc... 00:55 realbadangel they all should be registered by lua mod 00:55 realbadangel and engine should not carry ANY hardcoded type 00:55 zat sounds good 00:56 realbadangel this way we will get clean engine way to register any craft recipe 00:57 realbadangel and any subgame will be able to get own ones 00:58 zat has anyone worked in a mods manager that can resolve dependencies? 00:59 realbadangel not my area 01:00 zat that would be good 01:00 zat anyone who is interested: make it like nodejs's npm :P 01:04 realbadangel such things are problematic always. and best resolver is always human's mind 01:05 realbadangel but some are just too lazy ;) 01:05 zat naaaaaa 01:10 realbadangel anyway i can just comment such thing, havent even touched that code before 02:22 Zeno` realbadangel, are you here? 02:47 Zeno` hmmmm, #1904 02:47 ShadowBot https://github.com/minetest/minetest/issues/1904 -- Fix MSVC build by SmallJoker 03:16 electrodude512 gitk 03:16 electrodude512 derp, nvm, wrong terminal 04:20 paramat in lua, within 'minetest.register_on_mapgen_init(function(mg_params)' how would i check the presence of a certain mapgen flag?: 'if mg_params.flags "biomes" then' 04:27 paramat while on the subject, here's my addition of a 'biomes' mapgen flag, i guessed and used 0x12 in '#define MG_BIOMES 0x12' in mapgen.h 04:27 paramat https://github.com/paramat/minetest/commit/6268e42296ef95c9ddcf72831ded4575688ef0c5 04:28 Zeno` it should be 0x20 04:29 Zeno` 0x12 is 0x08 | 0x04 (i.e. DUNGEONS + LIGHT) 04:29 Zeno` err not that, but still 04:30 Zeno` it's a combination of two bits, and these look like they're bitwise flags 04:30 n4x if you are using bits as flags, the next value should be 0x0F 04:30 Zeno` no 04:30 VanessaE uh, no. 04:30 n4x er 04:30 Zeno` F is 1111 04:30 VanessaE unless 0x10 is used, then 0x20 is correct. 04:31 VanessaE (and 10 is indeed used) 04:32 kahrl I guess you guys aren't into the decimal system but have you celebrated the milestone of 1000 closed PRs yet? 04:33 Zeno` kahrl, I haven't I will celebrate tonight! 04:33 Zeno` paramat, that 0x20 was meant for you, by the way ;) 04:33 kahrl we should have a dev party 04:33 Zeno` count me in! 04:33 Zeno` kahrl, you're bringing the beer 04:33 kahrl nah I'm poor 04:34 Zeno` I'll see if a sponsor will give us a few kegs then 04:34 paramat thanks 04:41 paramat i need to learn about bitwise stuff 05:09 hmmmm well sheezy 05:09 hmmmm paramat: I have a proposal 05:10 hmmmm if you were serious about multiple BiomeManagers, why not have a "null" biome manager that generates without biomes? 05:10 paramat hi 05:10 hmmmm adding stuff like new mapgen flag parameters is okay but you have to realize they're permanent once a release is made 05:12 paramat 'multiple biomemanagers' meaning multiple biome systems? 05:12 hmmmm yeah, that's how it would be implemented 05:13 hmmmm class BiomeManagerParamat : public BiomeManager { u8 getBiome( .... ); // this is my override of getBiome! ... 05:14 paramat so the one null biomemanger would access each mapgen and create base terran only? 05:14 hmmmm hmm 05:14 paramat sorry that my programming skills and knowledge of terminology is not great 05:14 hmmmm I think not having any biomes is the wrong way to think about it 05:15 hmmmm "no biomes" would just be having a single biome registered, grass or whatever 05:15 paramat (i used to program BASIC in the 80s heh) 05:15 hmmmm 80s? 05:15 hmmmm you're older then lol 05:15 paramat lol 05:15 hmmmm i don't know why i always thought you were in your early 20s 05:16 paramat started programming at age 11 with a 48kb sinclair ZX spectrum 05:16 hmmmm that's awesome 05:16 paramat oh im very youthful =] look half my age too 05:16 hmmmm did you just never bother making a career out of coding? 05:16 hmmmm it's very lucrative 05:16 paramat i should never have stopped 05:18 paramat when i left home age 22 i stopped programming 05:18 paramat yeah good idea i'll look into coding 05:18 paramat as i prefer to work from home 05:20 paramat hm anyway i guess a flag is too permanent yes 05:21 paramat i can't release default biomes for mgv5/v7 without allowing players to disable them 05:48 hmmmm hmmm 05:49 hmmmm well 05:49 hmmmm when you say default biomes you actually are referring to biomes included with the minetest default game 05:49 hmmmm when i say default biome i mean the biome that's hardcoded into the biomemanager so it doesn't explode if nothing's registered 05:51 hmmmm i think it's best to include a 'clear biomes' api and make a note that one or more mods will compete with each other if they both register biomes 05:51 hmmmm mods shouldn't even be using biomes right now because it's technically undocumented and may be changed at any given time 05:51 hmmmm because biomes suck because i never got around to fixing them 05:52 hmmmm there was always something more important that needs to be coded *right now* 05:56 Zeno` hmmmm, can I merge #1904 please? I'm asking you specifically because of the typo (see comments) 05:56 ShadowBot https://github.com/minetest/minetest/issues/1904 -- Fix MSVC build by SmallJoker 05:57 hmmmm oh shoot 05:57 hmmmm how the hell did that pass my own unit tests 05:57 Zeno` false cast to 0.0 05:57 hmmmm hmm 05:58 Zeno` np.persist == 0.7 -> false 05:58 * paramat reads 05:58 hmmmm why 05:58 hmmmm OHH 05:58 hmmmm because of floating point round off 05:58 Zeno` yep 05:58 hmmmm and it just so happens that under MSVC, their atof() is more accurate 05:58 Zeno` so it would then get cast to 0.0 and the unit test passes ;) 05:59 hmmmm this is such an awesome bug 05:59 Zeno` I'm guessing that on MSVC an implicit cast from bool to double is an error 05:59 Zeno` rather than a warning that probably only shows up with -Wextra 05:59 hmmmm as for src/hud.cpp, that's sapier's doing, not my own. i did make a stdint.h mistake once though. 06:00 Zeno` I have no idea about the hud.cpp changes but they look obviously correct 06:00 hmmmm thank you smalljoker 06:00 hmmmm i'm glad there are people out there who actually try building with MSVC 06:02 Zeno` yeah :) 06:04 Zeno` ok, will merge now 06:12 Zeno` I'd also like to merge #1853 (which is my last outstanding PR) so I can clean up my @*$(@ working repo 06:12 ShadowBot https://github.com/minetest/minetest/issues/1853 -- Add name of node pointed at to debug by Zeno- 06:13 hmmmm looks good 06:13 Zeno` ok will do. Thank you 06:27 paramat time to nap before a possible orion launch O/ 11:25 Zeno` Anyone here familiar with particle spawners? 11:29 jin_xi kind of 11:30 Zeno` jin_xi, do you know where/how they are supposed to be deleted? 11:30 Zeno` because they're leaking 11:30 Zeno` a lot 11:31 jin_xi i made a pull (very wip) to replace them with irrlichts ones 11:32 Zeno` hmm 11:32 Zeno` http://pastebin.com/CDBDPvDw 11:32 Zeno` those lines are the source of the leaks 11:34 jin_xi there are many 'new's 11:35 Zeno` yeah, and I cannot find where they're being deleted (they're obviously not); I guess the better question might be where are they supposed to be deleted? 11:36 Zeno` and when? 11:36 jin_xi theres delete_particlespawner but i dont think it deletes anything else but the id of the spawner... 11:37 jin_xi whole particles code is a mess, i specially like how parameters are read from lua. 11:38 jin_xi https://github.com/minetest/minetest/blob/master/src/script/lua_api/l_particles.cpp#L62 11:40 Zeno` hmm yeah 11:46 jin_xi historically particles were first done entirely via lua and entities, then c++ + GL_QUADS was added... later it was messed up even more when forced horizontal particles were added 11:46 Zeno` they can't even BE deleted 11:47 Zeno` https://github.com/minetest/minetest/blob/master/src/game.cpp#L3007 11:48 jin_xi entire system should be replaced and irrlicht particles used instead 11:48 Zeno` I can't see how that memory allocation can be deleted because the reference to the memory is not even stored (as far as I can tell) 11:50 Zeno` same with line 2992 11:52 Zeno` I can't even understand what that's supposed to be doing lol 11:59 jin_xi so particles were originally done this way because it was not clear if/how irrlicht can do colliding particles 12:01 jin_xi https://github.com/minetest/minetest/pull/1700/ this pull adds irrlicht particles and collision detection but does not much else 12:03 Zeno` hmm 12:03 Zeno` that PR doesn't address the issue in game.cpp though 12:03 Zeno` the code that seems to do nothing at all 12:04 Zeno` Do you know who wrote the current code? Maybe they know 12:04 jin_xi jeija i think 12:10 jin_xi well, im trying to get someone capable looking at particles for a long time now, as my knowlege of c++ is as lacking and ancient. 12:20 kilbith realbadangel ^ 12:23 Zeno` jin_xi, good memory 12:23 Zeno` Jan 2013 12:29 Zeno` well it's broken, that's all I can tell at the moment 12:33 Zeno` jin_xi, if you can fix all these issues I'll support your PR :P 12:33 jin_xi yeah right 12:33 Zeno` ? 12:34 Zeno` I will... the current system is broken. It's even more confusing than code I write 12:34 jin_xi as i said, i can not do it... 12:35 Zeno` you want me to chat to RBA if you're not around? 12:35 jin_xi look, all i did was figureing out how to do collision with irrlicht particles, as it was given as the reason mt has rolled its own 12:36 Zeno` maybe your way is better 12:36 jin_xi so, i hope it helps someone to replace current particles, but i can not do it 12:36 jin_xi anyway, bbl 12:36 Zeno` ok, I'll talk with RBA and see what he's doing with it 13:45 Zeno` kahrl, I managed to get a sponsor to buy some beer for the party 17:09 * Krock thinks, Zeno` should merge the pull requests slowlier but more exact (reference flood at #1887 ) 17:09 ShadowBot https://github.com/minetest/minetest/issues/1887 -- Lighting artifacts 17:10 GhostDoge Krock: *more slowly 17:13 Krock whatever. 17:34 Zeno` Krock, what? 17:36 Zeno` 1887 is not a merge 17:36 Zeno` it's an issue 17:39 * Zeno` thinks Krock should concentrate on more important issues 17:42 Krock Zeno`, I meant the messages of "Zeno- referenced this issue from a commit in Zeno-/minetest 4 days ago" 17:42 Zeno` Yeah I have no idea why that happened. The merges are the same 17:43 Zeno` Perhaps some person(s) asked me to change the commit 17:43 Zeno` I can't remember 17:43 Zeno` it's not an issue 17:44 Krock Zeno`, it happens because someone changed the commit's contents, I would call that 'amend' 17:44 Zeno` of course 17:45 Zeno` Stop concentrating in minutia and look at the bigger issues. Please. 17:45 Zeno` there was only ONE commit/merge 17:45 VanessaE FINALLY someone said it 17:45 realbadangel hi all 17:47 VanessaE hi RBA 17:48 Krock VanessaE, said what? 17:49 VanessaE Krock: [12-05 12:48] Stop concentrating in minutia and look at the bigger issues. Please. 17:49 VanessaE that happens a lot around here. 17:50 Krock Really? I haven't seen it that frequent yet. Nvm. It's just a 'view' issue 17:50 realbadangel i guess we will have more issues like that until we change lighting system 17:50 Krock I've the feeling, a new lighting system will require more CPU power 17:51 VanessaE unlikely 17:51 VanessaE and it would use the GPU probably. 17:52 Krock Yeah. GPU is a much better soulation for those things 17:57 twoelk Eh? why is it only a much better solution? I thought it was invented to be THE solution, it is the reason the GPU exists :-D 18:08 Krock twoelk, just minetest doesn't use all GPU power yet :/ 18:09 twoelk all? does it use any? 18:09 twoelk not counting what the OS assigns anyways 18:10 Krock I think OpenGL is hardware-based, so it's more effective than the software render 18:11 Krock -think +know 18:11 GhostDoge twoelk: even the software renderer uses gpu power 18:12 Krock GhostDoge, yeah. the screen must be drawn somehow <.< 18:12 twoelk ah 18:12 * twoelk doesn't like his own tone and goes for a mining spree enjoying the deep down wonders of lavalakes 20:33 paramat MTgame team, lua pinetree generation and pine sapling growth is ready for merge https://github.com/minetest/minetest_game/pull/366#issuecomment-65749940 20:34 VanessaE no moretrees nodes' compatibility :'( 20:35 paramat explain what needs to be done, i fix 20:35 VanessaE nothing, it just means I'll have to add some if's and aliases in moretrees :) :) 20:35 * VanessaE erases the extra smiley. 20:35 paramat phew okay 20:36 VanessaE I have to wonder why you use a vmanip here instead of L-systems though 20:37 VanessaE (I'm sure it works perfectly fine, just seems like a ton of code by comparison) 20:38 paramat pines are near-identical to mg pines, it's sfan5's design 20:48 GhostDoge VanessaE: The l-system implementation is a ton of code too ;) 20:49 hmmmm so fun that nobody is going to implement L-system decorations aside from me :/ 20:56 Krock L-system is too difficult to understand for me somehow 21:06 VanessaE in such a situation, I'd just spawn saplings at mapgen time, and let an ABM with chance 1, interval 1 grow them. 21:09 realbadangel what is so difficult in l-system trees? 21:09 VanessaE nothing at all, but I think the issue here is not having the necessary hooks in the mapgen "decorations" system to spawn them natively 21:13 realbadangel https://github.com/minetest/minetest/blob/master/doc/lua_api.txt#L2294 21:13 realbadangel all the code needed to spawn a tree 21:15 realbadangel same structure is aviable on engine side, so its same dead simple 21:18 realbadangel https://github.com/minetest/minetest/blob/master/src/treegen.h#L38 21:18 realbadangel and then just call this: https://github.com/minetest/minetest/blob/master/src/treegen.h#L68 21:23 hmmmm oh, that seems pretty simple 21:23 hmmmm so lsystem decorations would be just a decoration wrapper around TreeDef and some "spawn near node" type of things 21:24 realbadangel https://github.com/VanessaE/moretrees/blob/master/tree_models.lua#L1 21:24 hmmmm and generate() could just call make_ltree() 21:24 realbadangel here you go with all the models you would need 21:24 hmmmm the only thing i fear is the cutoff 21:24 VanessaE hmmmm: yes, more or less like that. take a look how I use the code in moretrees/plants_liv 21:24 VanessaE lib* 21:24 hmmmm those lsystem trees get really tall 21:25 realbadangel not all of them 21:25 hmmmm alright 21:25 VanessaE that cutoff issue was fixed a long time ago, I thought? 21:25 hmmmm nope =/ 21:25 hmmmm i wrote code for it but long story short it sucks 21:25 realbadangel beech is about the size of old trees 21:25 VanessaE strange, I never see even the biggest ones get cut off in my test worlds (delete map, re-gen, repeat) but, *shrug* 21:25 hmmmm only issue I see is if I add this, i'm going to get moaning and groaning from people filing bug reports about how their lsystem trees got cut off 21:26 VanessaE (biggest being the giant sequoia) 21:26 VanessaE they USED to get cut up pretty bad, but I could swear someone fixed that a while back. 21:27 realbadangel hmmmm, i havent saw any cut off trees for quite a while 21:27 hmmmm that's because you write them to the map with spawn_ltree.... 21:28 VanessaE gotta run. bbl 21:43 kahrl is this a known issue? mgv6 sometimes generates air "holes" between land and ocean 21:43 kahrl for example, next to the spawn jungle on seed 0: http://i.imgur.com/HYQa3Wx.png 21:44 kahrl before I could make the screenshot, the nodes below the junglegrass and where the flowing water is were air 21:44 kahrl of course then liquid update kicked in 21:45 kahrl in a previously generated map with the same kind of issue, liquid update never happened and there were just holes 21:48 Krock kahrl, that's nothing new 21:48 kahrl ah, I thought so 21:48 kahrl never noticed it though for some reason 21:48 Krock I only noticed it with extreme mountain and ocean worlds 21:49 Krock with many height diffrernces and caves 21:49 kahrl you mean with modified noise params? 21:49 Krock exactly 21:49 kahrl well this is with the default params so perhaps it could be more seen as a bug 21:50 hmmmm did it exist in 0.4.3? 21:50 kahrl let me check 21:50 hmmmm (i.e. did I mess that up somehow??) 21:51 hmmmm if anything i bet that's caused by caves 21:52 Krock Question: Are there any 'vector.*' functions I should add in my next pull request (except of vector.floor())? 21:53 GhostDoge Krock: vector.ceil maybe? 21:54 Krock added 21:54 kahrl hmmmm: looks like it existed 21:55 kahrl liquid update happens quite fast in 0.4.3 though so it's hard to see 21:56 kahrl yeah, it's definitely there in 0.4.3 21:57 Krock kahrl, does the function name 'math.get_sign(x)' look okay for you? 21:58 kahrl Krock: where is it? 21:58 Krock kahrl, on my screen. it should return -1 when the number is negative and 1 when it's positive 21:58 Krock useful for directions 21:59 kahrl sounds good, isn't it usually called just sgn though? 21:59 kahrl so math.sgn 21:59 Krock *math.sign? 22:01 kahrl http://en.wikipedia.org/wiki/Sign_function calls it sgn 22:01 kahrl my math professors usually called it that as well 22:02 Krock hmm google returned me Math.Sign() from C# 22:03 kahrl you could add both ;) 22:03 kahrl math.sign = math.sgn 22:05 Krock :/ 22:18 Krock #1907 22:18 ShadowBot https://github.com/minetest/minetest/issues/1907 -- Add Lua helper functions by SmallJoker 22:21 Krock Slightly extended the math.sign() function with a limiter - it prevents from rounding if required 22:26 kahrl s/If the number is smaller/If the absolute value is smaller/ 22:27 kahrl I'd call the variable limit in the implementation too for clarity 22:29 Krock Thanks for the improvement. Updated the commit 22:29 Krock hmm