Time Nick Message 02:14 RealBadAngel sofar, here? 02:46 sofar RealBadAngel: sort of, what's up? 02:46 RealBadAngel about bf culling again 02:46 RealBadAngel that thing is introduced in protocol version 17 02:47 RealBadAngel but somehow broken 02:47 RealBadAngel lemme show you something 02:48 RealBadAngel https://github.com/minetest/minetest/blob/master/src/nodedef.cpp#L941 02:48 RealBadAngel for regular tiles its set in code, for special tiles its being read from tiledef only 02:49 RealBadAngel that should be somehow fixed, like setting it once with code and alter it when nodedef is set to a different value 02:49 RealBadAngel for both sets of tiles 02:50 sofar we can change 02:50 sofar case NDT_MESH: 02:50 sofar to 02:50 sofar f->solidness = 0; 02:50 sofar and then NO break 02:50 sofar and NO f->backface_culling = false; 02:51 RealBadAngel but you are making changes to NDT_MESH only, while "backface_culling" in tiledef is general setting for all the drawtypes 02:51 sofar whooops didn't see the end of the switch 02:52 sofar grmbl I hate redoing PR's to fix other people's bugs :P 02:52 RealBadAngel hehe 02:53 RealBadAngel its quite old, but i suppose it was designed so 02:53 RealBadAngel that flag was propably meant only for water nodes and nothing more 02:54 RealBadAngel and it was way before we had such fancy drawtypes as meshes ;) 02:55 RealBadAngel on the other hand bugs are sometimes useful, thx to wield mesh hunt ive fixed rendering to texture for postprocessing :) 02:57 RealBadAngel that means in fact i could apply now water reflections :) 02:57 RealBadAngel how neat that could be? :) 03:09 sofar ok so this isn't out of the woods yet 03:09 sofar backface_culling inside tiles seems to be enabled by default? 03:09 sofar I can't seem to turn it OFF 03:09 sofar lol 03:10 RealBadAngel default is true i think 03:10 sofar yeah I'll figure it out 03:10 RealBadAngel for sure 03:10 sofar at least, I think I can 03:10 RealBadAngel tiledef.backface_culling = getboolfield_default( 03:10 RealBadAngel L, index, "backface_culling", true); 03:11 RealBadAngel so if you dont specify it at all, it will hold true 03:11 sofar oh we never read the tiledef culling 03:11 sofar gotcha 03:12 RealBadAngel so if if its set to "false" it means that modder altered it 03:12 sofar that doesn't work 03:12 sofar wait 03:12 * sofar tries st else 03:12 RealBadAngel you have to apply that around line #941 03:12 ShadowBot https://github.com/minetest/minetest/issues/941 -- Index the player list from minetest.get_connected_players() with player names by ShadowNinja 03:13 RealBadAngel su bot ;) 03:14 sofar still something not right, if I just rely on f->tiledef[j].backface_culling, it still is always enabled even if my mod disables it for a tile 03:15 sofar afk 1yr old girl needs attention 03:15 RealBadAngel hold on a sec 03:15 RealBadAngel something like that should work: http://pastie.org/10693785 03:16 sofar Website is offline No cached version of this page is available. 03:16 sofar Error 522 03:16 RealBadAngel if tiledef's is false it means that it is altered, so override nodedef's set one 03:16 RealBadAngel wrrr, wait 03:17 sofar im sure i have the same change 03:17 RealBadAngel http://pastebin.com/fvvcgHgz 03:18 sofar https://gist.github.com/sofar/b435a32d14cf70fbbfd5 03:18 sofar afk... cant code when hugs are needed... 03:19 RealBadAngel ok, take ur time 03:19 frustrated howdy 03:19 frustrated i'm having a strange problem 03:20 frustrated a mod i'm writing had a bug on line 185 of some certain file in it 03:20 frustrated i fixed the bug, and added logging above that line, yet the game keeps crashing with the same error on the "old" line 03:20 frustrated i even renamed the file and the bug keeps happening with the old file name and old line number 03:21 frustrated but if i delete the file altogether the mod won't load 03:21 frustrated if i restore the file, even with a new name, the bug persists 03:21 frustrated is there some sort of lua cache? 03:21 frustrated this thing is driving me mad 03:23 sofar paste the code/bug 03:25 frustrated http://pastebin.com/be5uCV5Z 03:26 frustrated behavior.lua doesn't even evist anymore, and all the paths are correct 03:26 sofar are you sure it's not running your mod from a different mod folder? 03:27 frustrated it's the only one one my whole system 03:27 frustrated the bug suddenly started happening in the middle of development, with no unusual activity on my part 03:29 sofar are you puttnig mod inside your compiled minetest tree? 03:29 frustrated i compiled with RUN_IN_PLACE, then put the mode in /mods of the minetest folder 03:29 frustrated *mod 03:30 sofar and it's not reading mods from ~/.minetest/mods/ or anything? 03:30 frustrated nope. i'm on gentoo. that folder doesn't even exist 03:31 sofar how about /usr/games/minetest or anything? 03:31 sofar have you tried `locate behavior.lua` ? 03:32 frustrated locate returns nothing 03:32 sofar find / -name behavior.lua 03:32 sofar crude, lol 03:32 frustrated here's a new development: 03:32 frustrated i made a new world and the bug is gone 03:32 frustrated but in the old world it's there 03:33 sofar worldmods/ ? 03:33 RealBadAngel have you checked world folder? 03:33 frustrated nothing in the world folder 03:34 frustrated the function is assigned to a variable in a table that for simple_mobs 03:34 frustrated is the function just getting pickled somehow? 03:36 RealBadAngel theres no cache for mods, you have this mod installed and enabled for the previous world 03:36 RealBadAngel when the error is gone in the new one it means that mod is not enabled propably 03:36 RealBadAngel so you have just find the folder with your mods 03:36 frustrated no, i enabled the mod in the new one and it behaves as expected 03:37 sofar load your old world 03:37 frustrated sofar: still crashes 03:37 sofar is there a worldmods folder in that world? 03:37 frustrated nope 03:38 sofar find / -name behavior.lua 03:38 sofar did that give you anything? 03:38 frustrated it's taking a while, i have a lot of files 03:39 RealBadAngel btw, it looks like a mod bug, the function is called on entity 03:39 RealBadAngel there shall be check for nil on params added 03:39 frustrated data.mob is a table not an entity in that code 03:40 frustrated but i should change it if the name was not clear 03:40 sofar dammit rba 03:41 frustrated when an entity is "saved" to the world file, what happens to anonymous functions on it? 03:41 frustrated are they serialized? 03:42 RealBadAngel function mkFindNodeNear(sel, _dist) 03:42 RealBadAngel if sel == nil or _dist == nil then return "failed" end 03:42 frustrated sofar: the find command turned up nothing (i renamed the original behavior.lua file) 03:42 RealBadAngel that should fix the mod's bug 03:42 frustrated i didn't add checks because they are called with constants in only one place 03:43 RealBadAngel its wise to check for nils in mods 03:43 RealBadAngel shitload of mod bugs are from that 03:43 frustrated sure, but the thing where it was erroring on a file that didn't exist anymore is weird 03:44 sofar FINALLY 03:46 frustrated thanks for the help, i think it was just me not understanding lua very well :) 03:46 RealBadAngel sofar, got it working? 04:41 sofar RealBadAngel: yeah, I think i got it 04:42 RealBadAngel can i see the code changes? 04:42 sofar I just got back from kiddo bedtime, sec 04:56 sofar https://github.com/minetest/minetest/pull/3578 04:56 sofar update PR 04:56 sofar I'm just relying on tiledef to provide the value, which works 04:56 sofar the trick was the damn tiles = {} syntax 04:56 sofar believe it or not 04:57 sofar that had me confused more than anything 04:58 sofar RealBadAngel: I don't see why we want to invert logic for some node drawtypes 04:58 RealBadAngel actually i pasted you example with syntax needed in my comment on the PR before ;) 04:59 RealBadAngel its general property, it doesnt mean some usages may be wise 04:59 RealBadAngel but better to have it that way than restricting it for some nodes imho 05:01 RealBadAngel hmmm, atm you have set it to tiledef only 05:01 RealBadAngel which means you have disabled all the nodedef.cpp set values 05:01 RealBadAngel point was to allow override those ones not to trash them 05:03 sofar right 05:03 sofar yeah I'm just quickly testing with my other meshes and I'll want to change this 05:04 RealBadAngel i see 05:05 sofar actually 05:06 sofar your change makese no difference 05:06 sofar and it can't I think 05:06 sofar follow me here 05:06 sofar formerly 05:06 RealBadAngel go on 05:06 sofar meshes, always disabled 05:06 sofar so now we want to enable it per specific mesh tiledef 05:06 sofar but it's enabled by default there 05:07 RealBadAngel we need 3rd state? ;) 05:07 RealBadAngel lemme think a moment 05:07 sofar no we need to tell lua to disable culling in tiledefs for meshes 05:07 sofar by default 05:07 sofar and then just use my patch 05:07 sofar that maintains the default settings 05:08 RealBadAngel so just inverse my logic 05:08 sofar you can't fix this in nodedef.cpp alone 05:08 RealBadAngel if false is default, true will mean modified and have to be used 05:09 RealBadAngel tiledef.backface_culling = getboolfield_default( 05:09 RealBadAngel L, index, "backface_culling", true); 05:09 RealBadAngel you need to change that in c_content.cpp 05:09 sofar ah, yes, that's a way to do it 05:09 sofar very good indeed 05:09 RealBadAngel but hold on 05:10 RealBadAngel if we have true in nodedef and wanna disable it? 05:10 RealBadAngel same shit :) 05:10 sofar same problem 05:10 sofar yes, so we're back to always using the tiledef setting, but setting it to false for nodes where it's currently false 05:11 sofar bool culling_default = (NDT_MESH) ? false : true; 05:11 sofar something like that 05:11 RealBadAngel lemme have a cofee, need that for better thinkin 05:11 RealBadAngel atm i do have my brain full of texture tear problem and slabs ;) 05:15 RealBadAngel i think that "modified" flag is needed 05:16 sofar no, it's not that hard 05:16 sofar hold on 05:26 sofar someone explain me the difference between "visual = "mesh"" and "drawtype = "mesh"" 05:34 sofar you know, this whole bloody mess, is pissing me off 05:34 sofar it's all about how the hell lua gets parsed 05:34 sofar because effin shit 05:34 sofar tiles = { "foo.png" } 05:35 sofar is not the same as 05:35 sofar tiles = {{ name = "foo.png" }} 05:35 sofar that's the whole damn problem 05:36 sofar I'll see if I can fix it up, tho 05:38 RealBadAngel visual i think is for entities 05:38 RealBadAngel drawtype ofc for nodes 05:40 RealBadAngel it shall be the same, i mean all the 6 tiles shall be filled with this one 05:40 sofar I've got a solution now 05:47 RealBadAngel need to reboot and check something brb 07:29 sofar LOL 07:29 sofar signlike... is used by? ladders 07:29 sofar what are signs? nodeboxes 07:29 sofar seeds on the ground? 07:29 sofar signlike! 07:40 sofar https://github.com/minetest/minetest_game/pull/793 07:40 sofar enough PR's for today 07:40 sofar I'm done, have a good one 08:07 red-001 how do I get the default value for a setting in C++ 08:10 celeron55 red-001: in no way without modifying Settings to support it 08:11 red-001 so is it ok to hardcode defaults or should I change settings 08:15 kahrl red-001: what are you trying to do? 08:16 red-001 and a reset keys to the change keys dialog 08:16 kahrl I see 08:17 kahrl I'd probably add a method to Setting that removes a given setting (and reverts it to its default value, if any) 08:17 kahrl Settings* 08:17 red-001 yes I might do that 08:21 kahrl actually Settings::remove(name) exists 08:22 red-001 doesn't that just remove the setting? 08:22 kahrl yes, isn't that what you need? 08:22 kahrl if a setting isn't defined, the default is used 08:22 red-001 ok then 08:24 kahrl (don't expect its return value to mean anything, I'm pretty sure it always returns true) 08:29 kahrl hmmmm: that ^ was a bug introduced in f2c18511 08:29 kahrl the "delete m_settings[name].group" creates an entry if it doesn't exist, so the erase below it always returns 1 08:31 kahrl it didn't matter until now since nothing in minetest calls Settings::remove, if I checked right 08:48 hmmmm i do not even remember writing that commit 08:52 hmmmm well that's a simple enough fix (the Settings::remove thing - just use std::map::find, instead of using the convenience operator notation) 09:08 kahrl untested: https://gist.github.com/kahrl/3e37088d40f9346196db 09:13 hmmmm lgtm 09:13 kahrl thanks, merging 09:13 hmmmm aah I got a +1 on this https://github.com/kwolekr/minetest/commit/f8b3395227b2f15ea67844415c96e076f6c097ef 09:14 hmmmm you have any thoughts before I merge? 09:16 kahrl I don't think I know the emerge queue well enough to check that one 09:16 hmmmm it seems inconspicuous enough and it seems to work well... heh... heh... 09:17 hmmmm I'm going to write a unit test for the emerge manager at some point, I promise 09:18 hmmmm later 09:19 kahrl cya 12:33 lezzy Irrlicht: Could not open file of texture: <= all textures show correct but this fucking error is spamming it to death what is going on, i never had this issue before 15:18 paramat !tell hmmmm a freeminer schematic commit in case it's useful https://github.com/freeminer/freeminer/commit/551eb8515450be4167504f12653d111e09b0ee32 15:18 ShadowBot paramat: O.K. 15:44 paramat nice looking mars mod https://forum.minetest.net/viewtopic.php?f=9&t=13913 < celeron55 15:45 sofar I don't think Jordach noticed I wasn't using entities for the chests 15:46 sofar interesting mod 15:49 RealBadAngel he meant animations 15:50 RealBadAngel since the very start 15:50 RealBadAngel you have made just on/off 15:51 RealBadAngel we cant use animatons for regular nodes 15:51 RealBadAngel and propably never will be able to 15:55 kaadmy sofar: for signs and ladders i use nodeboxes now 15:55 kaadmy looks better and cleaner 15:56 RealBadAngel but anyway, sofar, youre asking the very same questions 15:56 RealBadAngel wonder where that will lead you 16:08 paramat luiz we know the cause, RBA will be hopefully be working on a fix. it's only 21-25% drop in FPS, not 70% 16:11 celeron55 there's a massive system load on my VPS and i'm now temporarily disabling a bunch of minetest stuff in order to diagnose it 16:15 RealBadAngel paramat, i will after texture tear problem, inventory meshes 16:15 RealBadAngel atm im workin on texture tear problem 16:17 paramat okay that's cool 16:18 RealBadAngel those problems stand in my way 16:19 RealBadAngel if i fix them i will be able to make reflections 16:19 RealBadAngel so you can bet i will fix them 16:36 luizrpgluiz sometimes I wonder if it would be interesting to put an option to choose to play the minetest more processor cores as a non-binding option 16:39 luizrpgluiz I know many developers like more OpenGL, but the game has a bug in the graphical engine that when you open the game in OpenGL 1.2, players can not see the icons of the items in survival mode and creative mode 16:41 sofar RealBadAngel: I'm prefectly happy with mesh nodes - I do not want to do animations as entities right now 16:41 sofar RealBadAngel: if one day mesh nodes can animate, then great 16:43 sofar Having chests behave visually like doors seems entirely coherent and natural to me 16:43 sofar RealBadAngel: plus I managed to get the chest-close sounds perfectly timed with people closing chests too 16:43 sofar even with multiple people looking inside the same chest 16:45 sofar the only sad thing... the formspec mostly obscures the chest as you open it :D 16:58 RealBadAngel the solution you proposed is imho right in the middle 16:58 RealBadAngel im ok with it 16:59 sofar I think it's fancy enough to make a difference, but simple enough so that it doesn't stand out with the rest of minetest_game 16:59 sofar and on purpose so 16:59 sofar I thought about node animations, and they're cool (I certainly want to learn blender better so I can make those) 17:00 sofar right now I can redraw the chest mesh in maybe 30 minutes 17:00 sofar so I'm getting more proficient in that 17:00 sofar which is fun 17:00 RealBadAngel you can do animations is separate scene nodes 17:00 sofar I know 17:00 RealBadAngel easily 17:00 sofar I made an animation for a different mesh already 17:01 sofar but, that's besides the point 17:02 sofar the things I'm proposing for minetest are things that I think should be in minetest_game 17:02 sofar having a few more mesh nodes in minetest_game will help mod makers show what's possible 17:02 RealBadAngel youre following the pretty same way to mine 17:02 sofar similarly I think we should have 2-3 monsters in minetest_game ultimately 17:02 sofar not 13 17:02 sofar just 2-3 really well-done 17:03 sofar I should perhaps also git commit my .blend files 17:03 RealBadAngel get client side lua first 17:03 sofar right 17:03 sofar perhaps that's what is needed first 17:03 RealBadAngel you have saw proof of concept already? 17:03 sofar no 17:03 RealBadAngel by Jeija 17:04 sofar oh, I know what needs an upgrade first in minetest_game? 17:04 sofar torches :) 17:04 RealBadAngel im talkin client side lua 17:04 RealBadAngel it possible since ages 17:04 sofar I'm sure it is 17:04 luizrpgluiz put monsters back in minetest would be a good idea, but do not need to put in the source code of the game, can be in the same language script 17:04 RealBadAngel back? we never had any 17:04 sofar luizrpgluiz: right, not in C++ 17:05 luizrpgluiz yes 17:05 sofar he's just saying "merge" 17:05 RealBadAngel floating sprites do not count 17:06 RealBadAngel sofar, https://github.com/Jeija/minetest/tree/clientside_mods 17:07 sofar RealBadAngel: I'm sure it's not difficult 17:07 RealBadAngel ofc not 17:07 luizrpgluiz I know I do monsters in 3D takes work, but I have a friend who can program well in LUA and he made a monster using the 4 horsemen of the apocalypse where it only comes when you have bone that has not yet been removed from the map, and the mod checks that no longer has the same monster like him 17:08 RealBadAngel but propably against the spirit of the mt ;) 17:08 RealBadAngel and its simplicity 17:08 sofar I dunno 17:08 sofar what are you going to have the client-side api do? 17:08 RealBadAngel move things 17:08 RealBadAngel without client asking server 17:09 RealBadAngel mobs, items, carts etc 17:09 sofar how does the server keep track of your position then? 17:09 RealBadAngel dont ask me 17:09 RealBadAngel idk 17:10 RealBadAngel but m$ paid billions for that 17:10 RealBadAngel and indie guys were able to make it 17:11 luizrpgluiz for now I use the windows to play, but Linux for me is the best system to develop something 17:11 Fixer win is degenerating anyway 17:12 RealBadAngel win? i had something like that a few yrs agp 17:13 luizrpgluiz but minetest runs better on Linux than in Windows 17:13 RealBadAngel depends on the drivers sadly 17:13 luizrpgluiz yes 17:13 RealBadAngel nvidia is beter supported 17:14 luizrpgluiz in my notebook I realized that the game had a very good improvement, but of course my laptop is 10 years Ago 17:15 RealBadAngel im said that i do develop high end candies 17:15 RealBadAngel but i do that on most cheap i3, 4 gb ram, 6 yrs old GTX550ti 17:16 RealBadAngel so i usually want to say "stfu" 17:17 luizrpgluiz My notebook is simple, the configuration is a Pentium dual core, 3GB of ram and Intel HD graphics card 17:17 RealBadAngel 5-6 yrs old gaming platform for coder is like having sex with dinosaurs 17:17 sofar LOL 17:17 sofar painful? 17:17 RealBadAngel they do have STANDARDS to take care of 17:18 RealBadAngel and yet you know those standards are all gone 17:18 RealBadAngel and dinosaurs are exctint already 17:18 luizrpgluiz hehehe 17:19 RealBadAngel but mt HAVE to be compatible with all those all dead poor dinos 17:19 Fixer I usually update every 10 years or so 17:19 Fixer and now you don't need to update at all 17:19 luizrpgluiz the minetest be done in C#,many developers would not be here 17:19 Fixer it is good enough 17:19 RealBadAngel i had an ati card, 2 yrs ago or something 17:20 RealBadAngel AMD dropped drivers for that card then 17:20 RealBadAngel so i wasnt even able to get drivers for it 17:21 RealBadAngel it was a few times faster than minetest reference i945m 17:22 RealBadAngel ive "sold" this card to work witch text editors and data sheets at my uncle office 17:23 RealBadAngel that topic about conspiracy has more sence when you stop laughing and think about things goes here 17:34 sofar hmm, I have 13 pending PRs 17:34 sofar this is getting out of control 18:07 sofar RealBadAngel: client-side modding would be nice for sounds, effects and the like 18:08 sofar RealBadAngel: I've been thinking of adding weather, the effects for that should all be client-side tho 18:08 sofar RealBadAngel: similarly, explosions, fire, etc. could be client side 18:08 sofar RealBadAngel: sound effects 18:14 twoelk actually individuall fixed position sounds should be all client side as they are different for each client - but somehow I thought sounds where allready processed clientside only 18:15 sofar not the scheduling 18:15 twoelk mh - now where is that toilet flush in homedecor 18:16 sofar servers tells client where to play sounds 18:22 twoelk as it should as the position is part of the map but the calculating what is actually heard is calculated by the client, I though (or maybe just hoped) 18:23 rubenwardy Ambient sounds should be client side 18:23 rubenwardy Lua or not 18:24 twoelk can I actually decide in a mod who proccesses something? 18:24 rubenwardy ? 18:25 rubenwardy What do you mean by who, twoelk 18:25 twoelk client or server 18:25 rubenwardy No. Mods are entirely client side, there is no lua environment on a client 18:25 rubenwardy (There's the main menu, but that is destroyed by the time the client is running) 18:26 twoelk whatabout all the stuff in builtin? 18:26 rubenwardy that's server side 18:27 rubenwardy It's just a hidden mod that always runs first 18:28 twoelk ah, so the client runs lua on sdtartup but blocks lua when connecting tio a server 18:28 twoelk *startup 18:29 rubenwardy The mainmenu of the client runs lua but stops running lua when it starts connecting 18:29 rubenwardy It doesn't block anything 18:31 twoelk so if I flush the homedecor toilet or hear the villages church bell all the calculating of checking and comparing the positions is done by the server that then tells the client how to play the sound from the cache? 18:32 rubenwardy Yes. The server runs a lua command which sends a packet to the client telling it which sound and where 18:33 twoelk I had thought it would be just - soundhere + rangethat - and the client figures out the rest 18:33 rubenwardy I don't understand what you're saying 18:35 twoelk sounds can have a range, right? 18:35 sofar sounds are rendered just like graphics 18:35 sofar the client does all the rendering 18:35 sofar the server says "a sound is here, this loud" 18:35 sofar the client tells openAL to render the sound in that position 18:35 sofar the sound rendering is done in C++ 18:36 sofar what my suggestion is, is to add ambience sounds in client-side lua 18:36 sofar e.g. you take damage - on_client_damage() handler in lua gets called and plays sound and displays special effects 18:36 sofar that's several less packets to the client 18:37 sofar since now the server only needs to send a "damage" event 18:37 twoelk so sounds get handled like client side textures? 18:37 sofar and not a "damage" event, and a "sound" event, and a "effect" event 18:37 sofar sounds get handled just like a node change, or an object move 18:38 twoelk wouldn't that be too general? 18:39 sofar I don't get your question 18:40 twoelk maybe the game has more detailed sounds for special effects 18:42 twoelk might be on another siedtrack but I was just playing with positioning certain sounds in certain places that mingle with each other 18:43 sofar twoelk: obviously the subgame would have to be able to tell the client how and what sounds to play and when 18:43 sofar twoelk: in other words, the client-side lua would be sent to the client by the server when you connect to a server 18:44 twoelk for example I don't just use one -running water- sound but experiment with different ones to give fountains a unique sound character 18:44 sofar twoelk: also, sound assets are already coming from the server, so the client has access to per-subgame sounds 18:44 twoelk or as in sokomines villages use different bells for each church 18:54 twoelk sofar: I fear I don't get where you want to split the computing 18:54 twoelk the server sends -this damage to that- and the client looks up the sound and effects? 18:55 sofar twoelk: I have a better example 18:55 sofar for instance, the player walks into a water node 18:56 sofar if the player is only feet-deep in water, play a splashy water foorstep sound 18:56 sofar if the player is head-deep, play a "player submerges" sound 18:56 sofar that doesn't need any server side processing 18:56 sofar can all be done client side 18:56 sofar but 18:56 sofar let's take that 1 step further 18:56 sofar suppose you're standing NEAR the water 18:57 sofar and a different player walks into the water 18:57 sofar the client knows a player object walked into water 18:57 sofar it knows how deep the other player is 18:57 sofar so it can play the right water sound effect for the other player 18:57 sofar also, it can put splashing particles on the water 18:58 red-001 how would this be used by mods? 18:58 twoelk you mean it can calculate the right sound (loudness, direction) 18:58 sofar twoelk: no, loudness and direction is calculated by OpenAL 18:58 twoelk wich is clientside 18:58 sofar the difference is that the server doesn't send any sound events to the client 18:59 red-001 doesn't this need client sided lua? 18:59 sofar the client lua code does all the detection of sound events and generate them 18:59 sofar red-001: yes 18:59 sofar red-001: design wise, you'd have a small set of builtin lua client-side code to initialize and run handlers 19:00 sofar red-001: then the subgames would provide most of the code that generates events 19:00 twoelk well sounds pretty much as I had hoped was allready done 19:00 twoelk :-P 19:00 sofar red-001: e.g. it would run each tick to detect things like players walking into a wall 19:00 sofar twoelk: it's all possible now, but not client side - e.g. the server needs to sent all the clients all the events 19:00 sofar which just adds lag, and processing time 19:01 red-001 so we will finally have client sided lua? 19:01 sofar red-001: I'm considering it unavoidable 19:01 red-001 I heard celeron55 started working on it 19:01 red-001 so he might have some code you could use 19:01 twoelk red-001: of course - just nobody knows when 19:02 sofar At this point I have a few other things I want to finish first 19:02 sofar I've got 13 open pull requests 19:03 sofar I'd like to see most of them merged ;) 19:03 * twoelk hopes for the next merging spree 19:03 rubenwardy hm_mm started making a prototype, then c55 said he would 19:04 celeron55 ...then i did the farmesh reboot instead 19:04 celeron55 but we'll see 19:05 sofar celeron55: hehe. as said I've got 1-2 things that I want to work on first after my current waterfall of changes, but client-side lua may be after ;) 19:05 rubenwardy This is cool: https://www.youtube.com/watch?v=nHH8N_lNZzI 19:05 rubenwardy ~web title 19:05 ShadowBot rubenwardy: web title [--no-filter] 19:05 twoelk waterfall with sounds? 19:05 rubenwardy ~web title https://www.youtube.com/watch?v=nHH8N_lNZzI 19:05 ShadowBot rubenwardy: HTTP Error 429: Too Many Requests 19:05 Krock lol 19:05 rubenwardy ShadowBot is the most useless IRC that I've ever interacted with 19:05 ShadowBot rubenwardy: Error: You must be registered to use this command. If you are already registered, you must either identify (using the identify command) or add a hostmask matching your current hostmask (using the "hostmask add" command). 19:05 rubenwardy point proved 19:06 rubenwardy "Rigid-Body Fracture Sound (ACM SIGGRAPH 2010)" 19:06 twoelk lol 19:06 sofar twoelk: hehehe 19:07 * twoelk is dreaming of walking accross a market with lots of sounds mingeling from the right directions and all individually adjustable 19:09 Fixer and farmap is buried in PR... 21:54 paramat will merge 3578 3581 very soon 22:01 nrzkt paramat, wait for 3578 22:01 nrzkt i have some comments 22:01 paramat ok 22:02 paramat i'll wait so i can merge the 2 PRs at once 22:37 RealBadAngel sofar, good idea to move defaults 22:37 RealBadAngel that looks ok now 23:13 rubenwardy celeron55, 500 Internal Server Error 23:13 rubenwardy on forums.minetest.net 23:14 sofar he was rebooting some stuff 23:14 sofar apparently he had some load issues 23:14 rubenwardy and is now working O_o 23:14 RealBadAngel luckily irrlicht forums are up now 23:15 sofar RealBadAngel: did you edit your github comment? 23:16 sofar RealBadAngel: special_tiles and tiles use the same method, so I think that works now similarly 23:18 RealBadAngel noticed that? lol 23:18 RealBadAngel ive deleted that comment because of water 23:19 RealBadAngel in its case content_mapblock expects hardcoded values in special tiles 23:19 RealBadAngel that should be changed imho, its mess 23:19 RealBadAngel and special tiles shall carry the same flags as regular ones 23:20 RealBadAngel at least by default 23:20 RealBadAngel then we can add that switching tile sets to reduce number of nodedefs 23:22 RealBadAngel as far as i remember only thing that should be added is an alternative light source value 23:23 sofar this change should help at least 23:23 RealBadAngel furnaces active become light sources 23:23 sofar moving more of the tile settings to the lua parsing code makes it more readable 23:23 sofar hmmm that's nice 23:24 RealBadAngel everybody here is crying about high ram usage 23:24 RealBadAngel and things that can reduce it simply are not in 23:24 sofar you should try to split your PRs better 23:25 sofar I know it's hard 23:25 sofar but it's so worth it 23:25 RealBadAngel if a few liner can half the number of for example mesecons nodedefs is definitely worth it 23:25 sofar especially since you depend on it 23:26 RealBadAngel well, i can reanimate that PR now 23:26 RealBadAngel need a break from texture tearing anywyay, its a frustrating work 23:27 RealBadAngel the reason for it is deeply burried, not only slabs are affected 23:27 RealBadAngel its visible on each and every surface, but somehow on slabs just a bit more 23:28 sofar do you want to do some mundane C++ work? 23:29 RealBadAngel oh great, another task? ;) whats that? 23:29 sofar I need a new metadata container in ItemStack, but I hate coding up classes from scratch 23:29 sofar it needs a new lua method 23:29 RealBadAngel why would you need that? 23:30 sofar so you can attach attributes to itemstack items 23:30 RealBadAngel its already capable of storing anything you would ever need 23:30 sofar not in a coherent and c++ safely readable way 23:30 sofar the itemstack meta could also be written from lua as a plain string directly 23:30 sofar it needs separation just like nodemetadata has 23:30 RealBadAngel hmm, i was using it to store whole tables 23:31 sofar :get_meta() ; meta:set_int("name", 120) ; etc. 23:31 sofar it should really have the same design as nodemeta 23:31 RealBadAngel technically you can even store here blobs 23:31 sofar items inside items 23:32 sofar but 23:32 sofar more importantly 23:32 sofar infotext 23:32 sofar and 23:32 sofar description 23:32 sofar are the ones I want to immediately use 23:32 sofar that way we can have named items 23:32 RealBadAngel infotext you already have 23:32 sofar nodedef infotext 23:33 sofar != itemstack metadata infotext 23:33 sofar you can't modify nodedef infotext 23:33 RealBadAngel hold your horses 23:33 sofar I want to rename swords 23:33 RealBadAngel ofc you can 23:33 sofar books with titles 23:33 RealBadAngel thats the lua mod that sets it 23:33 sofar there's only 2 book craftitems 23:34 sofar you can't rename the book 23:34 sofar it's always a "book" 23:34 sofar you can attach a description field in the itemstack metadata 23:34 sofar but the client doesn't know how to display that 23:34 sofar https://github.com/minetest/minetest/pull/3449 23:34 sofar that PR changes that 23:34 sofar except, the use of itemstack metadata is weaksauce 23:35 sofar it needs a full meta backing storage 23:35 sofar a real backing storage like nodemetadata 23:35 RealBadAngel i remember that PR 23:35 sofar with that, we can enchant swords 23:35 sofar make potions with damage values 23:35 sofar all sorts of varied items coming from a single nodedef 23:35 sofar plus renaming books 23:36 sofar or keys, like my example 23:36 sofar basically, I was thinking of splitting NodeMetadata into a proto Metadata class 23:36 sofar and have NodeMetadata inherit it 23:36 sofar that would be patch #1 23:36 ShadowBot https://github.com/minetest/minetest/issues/1 -- GlowStone code by anonymousAwesome 23:36 RealBadAngel LOL 23:37 sofar patch #2 would be to add the new Metadata class to ItemStack 23:37 ShadowBot https://github.com/minetest/minetest/issues/2 -- Burned wood 23:37 RealBadAngel hehehehe 23:37 sofar and #3, tell shadowbot to go fuck himself 23:37 ShadowBot https://github.com/minetest/minetest/issues/3 -- Furnace segfault 23:37 sofar sorry :) 23:37 RealBadAngel that was really good :) 23:38 RealBadAngel it even segfaulted on that fuck ;) 23:38 RealBadAngel but yeah, that would make things easier 23:39 sofar I know what needs to be done 23:39 RealBadAngel all is actually doable atm with lua, except for displaying custom names 23:39 sofar just dont like class deign 23:39 RealBadAngel hmm 23:40 RealBadAngel actually theres a simpler way to have custom tooltips 23:40 sofar well that patch gets the display part already 23:40 sofar that PR 23:40 RealBadAngel lemme check your code and compare it to my idea 23:40 RealBadAngel hold on 23:41 sofar read the PR thread too 23:43 sofar RealBadAngel: your "display infotext of dropped inventory items" could also use this metadata store to retrieve an alternative text 23:43 sofar instead of "a sword" it could display "DragonSlayer 3000(TM)" 23:46 sofar afk