Time Nick Message 00:00 PilzAdam it seems to happen when a sapling hangs on the edge of a node (see screenshot) 00:00 PilzAdam you can litteraly watch the stack growing up 00:08 hmmmm mmm.. 00:09 hmmmm just realized, papyrus doesn't generate in mapgen v6 if water_level != 0 00:09 hmmmm hardcoding and assuming all these things, jeez 00:10 PilzAdam we are switching to register_decoration() anyway 00:11 hmmmm one question; what do you think about decorating the bottom of the ocean? 00:11 Exio with what? 00:11 hmmmm *shrug* 00:11 Exio :P 00:12 hmmmm maybe somebody came up with a mod that places starfish or kelp or something 00:12 kahrl sunken ships 00:13 kahrl should underwater plants be made in a way that makes them not air pockets? 00:13 hmmmm obviously 00:14 khonkhortisan and not just underwater, but any liquid 00:14 Exio i wonder how without hacks 00:14 hmmmm you say that like the solution is obvious 00:14 hmmmm is the content at that node water, or is it the decoration? 00:15 khonkhortisan Fluid properties could be added to the decoration, and render both, but tiles are needed for both. 00:15 hmmmm btw if you can solve this in an elegant manner, we'd have another one up on minecraft. they don't have anything of the sort either. 00:17 diemartin hmmmm, http://forum.minetest.net/viewtopic.php?id=4627 00:19 hmmmm i'm assuming he did that with a nodebox, and making it an alternative liquid? 00:19 hmmmm either way, if we can make what that guy is doing in the mod not-glitchy, there'd be a solution 00:22 PilzAdam this mod has other 1000 lines of nodeboxes 00:23 hmmmm yeah but nevermind that, how does it merge with the water nodes? 00:23 hmmmm i am not seeing anything special 00:25 PilzAdam ok, its a normal node where the nodeboxes is higher than the actual node 00:26 PilzAdam the plant is above the actual node 00:26 Exio offtopic-about-minetest_game, am i the only here who thinks the grass, flowers and dry shrub should be dig_immediate? 00:26 Exio (buildable_to nodes) 00:27 PilzAdam hmmmm, its a hack 00:27 hmmmm is this feature wanted enough to make it not a hack? 00:28 PilzAdam I dont need it 01:50 VanessaE Exio: very little improvement for me, at least not in that one "bad" area on my server's map. +4 FPS perhaps 01:50 VanessaE (running latest git with your patch already merged per kahrl) 01:50 Exio aw 01:51 VanessaE CEnv: num of objects....... 477 !? 01:51 VanessaE where?? 01:53 hmmmm :D i found a nyancat 01:54 VanessaE CM: blocks occlusion culled: 0. um ? 01:54 hmmmm vanessa's seed, 305, -1544 01:55 VanessaE oops. 01:55 VanessaE hmmmm: kacey found one at about -329,-188,-571 01:56 Exio aw damn 01:56 Exio got trapped with sand 01:59 VanessaE Exio: your change did help marginally, like I said about +4fps 02:00 hmmmm https://github.com/minetest/minetest/commit/f18b4872ea7e46b50c9ea0e5e6484baae41b0a44 02:02 Exio i just did what the comment said 02:03 * VanessaE watches num of objects creep up.. +200 since I mentioned it a few mins ago, and I'm just standing here. 02:03 VanessaE (now 734) 02:04 Exio is it just client-side? 02:04 VanessaE beat me. 02:05 VanessaE beats* 02:05 VanessaE 882 now. 02:05 VanessaE and my fps is dropping accordingly. 02:06 Exio any local-mod for reproduce it? 02:06 Exio (you know, my network is shit) 02:06 VanessaE Exio: there are only four mods on my server that make entities, that I know of - simple mobs, item_drop (technic), pipeworks, and signs 02:07 VanessaE only the signs entities are visible from where I'm standing. 02:08 VanessaE well O tjoml ot 02:09 VanessaE er.. 02:09 kahrl default makes entities (items) 02:09 * VanessaE moves her hands to the right keys 02:09 VanessaE kahrl: I mean actively, as part of normal play 02:09 VanessaE well I think the cause of my low fps is pretty obvious: too many entities and lack of effective hidden block culling 02:10 VanessaE with a view range of 50, it manages to cull exactly one "block" (I assume a mapblock) 02:10 Exio i would like a way to reproduce it locally :/ 02:10 VanessaE Exio: download my server's map and game from the form page. 02:10 VanessaE http://forum.minetest.net/viewtopic.php?id=4057 02:11 Exio size? 02:11 Exio 4TB? ;P 02:11 VanessaE then teleport to -310,19,-421 02:11 VanessaE heh no about 990 MB for the map and about 14MB for the game 02:11 Exio 1gb 02:11 Exio that is like 2 days with this network, sadly 02:11 Exio s/like/about/ 02:12 Exio (yes, i like offtopic) 02:12 VanessaE number of objects is now 1029 and I haven't moved from this spot since I signed onto the server. 02:12 VanessaE I seem to lose about 1 fps per 100 objects 02:15 Exio VanessaE: can i have your debug.txt 'of that'? 02:16 VanessaE the server's debug log, you mean? 02:16 Exio i want to check the infostream, so yes :P 02:17 VanessaE sure, I 02:17 VanessaE I 02:17 VanessaE damn it. 02:17 VanessaE I'll upload it in a sec. 02:18 VanessaE 1177 objects now, and my fps is down by 10 now. 02:19 VanessaE holy...it's a god-awful big file 02:20 VanessaE 4.1 GB for the log file alone 02:20 Exio rm 02:20 Exio and reproduce it other time :P 02:20 VanessaE lol 02:22 VanessaE deleted the log and restarted the server. 02:22 VanessaE client shows 541 objects now. 02:23 Exio wait a bit, until it gets 600 or 650 or so 02:23 Exio then post it :P 02:23 VanessaE sure. 02:24 Exio thanks 02:25 VanessaE flew around a bit and loaded up all of the map that would be visible from where I am standing now (but no more than that). fps is peaking 31 where it used to do about 23-ish 02:26 Exio i would call that part of your map the FPS killer 02:26 VanessaE yes, indeed so 02:26 Exio if you survive 10 minutes inside that zone, you win a prize :D 02:26 VanessaE standing at the spawn, looking down the north road from the left side of the road. 02:28 Exio what about https://github.com/minetest/minetest/pull/539? 02:29 VanessaE I walked forward a bit to examine a sign up ahead, then returned to the spawn. 816 objects now. 02:29 Exio kahrl / hmmmm: ^ 02:29 hmmmm no idea exio, not really my area... 02:29 Exio but, what do you think? 02:30 hmmmm why not 02:31 VanessaE damn it. bad ownership on the log file after 'touch'. had to start over. 02:31 VanessaE (it didn't create a log for that +300 object jump) 02:31 VanessaE Exio: I have no objection to that. 02:32 Exio aw 02:33 VanessaE ok, it's creating a log now :-) 02:33 VanessaE loaded up all of the relevant area of the map, started with 555 objects, and pushing about 30 fps. 02:35 VanessaE walked over to the same area to check out the same signs, then a bit beyond, and returned to the spawn. 705 objects now. 02:35 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/debug-server.txt 02:35 Exio downloading 02:36 VanessaE this time it's "only" 5MB 02:36 VanessaE wow, 1MB per minute of uptime. :D 02:38 VanessaE obviously the signs entities are the ones being duplicated 02:39 VanessaE (there are a couple dozen such signs in the area) 02:40 Exio i don't get how the entities will be duplicating only client-sidely 02:40 VanessaE no clue at all. 02:41 Exio as that is in the serverthread... 02:42 VanessaE want the client's log too? 02:42 Exio unless it is 40 gb, yes :P 02:42 VanessaE heh 02:42 VanessaE I'll edit out all but the last session :-) 02:43 VanessaE holy.. 02:43 VanessaE 2GB for that bastard 02:43 VanessaE perhaps..I should redo that one also 02:43 Exio ha 02:45 VanessaE http://digitalaudioconcepts.com/vanessa/hobbies/minetest/debug-client.txt 02:45 VanessaE enough to see it rise from 550 to 700-and-change 02:47 VanessaE 22:44:50: INFO[main]: ClientEnvironment::addActiveObject(): added (id=5441) <--- WHAT!? 02:48 VanessaE surely that doesn't mean it created and/or destroyed that many objects? 02:50 * VanessaE updates the downloads of her game and world files (just in case the existing ones are too old) 05:31 VanessaE kahrl: after clearing objects, ad signing out/in, I'm getting 40-47 fps at the spawn, using the modified game (and already two sheep have re-spawned and are walking around) 05:31 VanessaE and* 05:38 VanessaE kahrl: started a new world as requested, still with the all-nodeboxes-replaced-with-cubed modified game, view range of 50, flew around a bit to get a good amount of generated land, blocks occlusion culled is still 0. If I turn the view range up, it starts exceeding 0. Exceeds 10 at a view range of 90, exceeds 100 at a view range of 220. 05:40 VanessaE there is only 1 object in the loaded region, and with a view range of 100 I can still get 25 fps even with lots of large trees in view (if I turn away from them, fps shoots up into the high 50's) 05:42 * VanessaE heads off to bed. 05:43 hmmmm wow this is freaking amazing 05:43 hmmmm exio, you're the one who fixed that re-calculation, right? 05:44 hmmmm minetest is running a solid 7-13 fps faster for me, i'm hitting 50 fps no problem 05:44 hmmmm this has gotta be the single biggest enhancement in a while 05:47 hmmmm this is awesome.. people ought to be able to experience this right away 05:48 hmmmm i'm going to finish up decorationdef at least to the schematic, and then move trees and what not over to that, and i think we need to make an 0.4.8 release 05:49 hmmmm anything else big currently cooking? 05:54 ShadowNinja hmmmm: thexyz is working on rebasing db_backends_2. 05:54 hmmmm i'd wait for that, but i really feel like this is enough of a reason alone for people to update 05:55 hmmmm (and then on top of that is other performance boosters and bugfixes) 05:59 thexyz ShadowNinja: sadly, I don't have much time those days 06:57 lrh9 Hello. 12:55 Exio the line i removed was just doing what a comment said.. 12:55 Exio :P 12:58 PilzAdam ok, this duplaction of saplings and decayed apples seems to be something new 12:59 PilzAdam I found two "stacks" in my singleplayer world too 13:10 rubenwardy This would be awesome: http://irrlicht.sourceforge.net/forum//viewtopic.php?t=16037 13:10 rubenwardy It is a portal scene node 13:11 rubenwardy you could add a draw type "portal" and a portal is drawn instead of transparency 13:11 rubenwardy portal location is defined by setmeta("portal","0,0,0") 13:11 rubenwardy any thoughts? 16:50 sapier https://github.com/minetest/minetest/pull/774 MirceaKitsune (indirectly) suggested changeing set_animation prototype to avoid having a new function this obviously requires changeing protocol too. Any oppinions to doing it this way? Except of compatibility problems I don't see any issue with this but this is a decision that should be based on more than 2 oppinions 16:55 Taoki sapier: I think it shouldn't require changing protocol, since we wouldn't add any new parameters. We just have to make sure that leaving the other parameters to nil works 16:55 Taoki eg: set_animation( , 10, ) 16:55 sapier we do add a new parameter 16:55 sapier base_velocity 16:55 Taoki Ah 16:55 Taoki What is that for? 16:56 Taoki There already is one for animation speed 16:56 sapier yes but this is the speed the animation is played 16:56 Taoki Sure. What would the other one do though? 16:56 sapier compare it to slow motion normal fast forward ... 16:57 Taoki I still don't understand 16:57 sapier the base velocity specifies which object speed matches the animation at (default) animation rate 16:57 Taoki We have a parameter for animation speed. Why would we need 2? 16:57 Taoki ah... 16:57 Taoki A builtin way to change animation speed vased on object velocity? 16:57 Taoki eg: For walk animations to match the walking speed? 16:58 sapier yes but not only builtin but client does already predict the new animation rate while objects velocity changes 16:58 Taoki I see 16:59 sapier server doesn't have to update animation speed if velocity changes 16:59 Taoki Well firstly, how much is that needer right now? There's not much acceleration /. deceleration so I'm not sure if such a change would be very visible 16:59 sapier this doesn't work reasonably well due to network latency 16:59 Taoki Sure, doing that over network would be sotyl 16:59 Taoki **costly 16:59 Taoki But how much does this improve anything visually? 16:59 sapier it's one of main graphical glitches for mobf mobs which do smooth speed changes 17:00 Taoki ok 17:00 sapier improovement is drastical as mobs don't seem to slide through landscape 17:00 Taoki Yeah, sounds like a new parameter would be needed then... specifying how much velocity is applied to the animation. But I still agree it should be part of set_animation and not a new function 17:00 sapier of course this depends on animation quality 17:00 Taoki And such would require bumping the protocol version 17:00 Taoki There would be another way, but rather hacky and I doubt anyone would agree 17:01 sapier sending two messages for one setting isn't an option thats not even a hack ;-) 17:02 Taoki I was thinking animation speed could have a negative value to indicate it's influenced by velocity. But that would suck so no 17:02 sapier that wouldn't tell you how animation is related to objects speed 17:03 sapier e.g. if someone specifies very high detailed animations 1block/s could be 100 frames 17:03 Taoki Yah. A new parameter should be added then I think 17:03 Taoki It should be ok to simply bump the protocol version up 17:03 Taoki I think. Best to ask PilzAdam to be safe 17:04 sapier that's why I post it here to get some more oppinions to this issue :-) 17:06 Taoki That's good. Sadly I don't know exactly what would need to be done, other than the protocol version must be bumped and documented 17:07 Taoki hmmmm: Are you there? 17:08 sapier I know what needs to be done it's only typework but I don't want to change it just to realize the old version is the one to add ;-) 17:16 hmmmm yes? 17:16 Taoki hmmmm: Hi. 17:17 Taoki hmmmm: I wanted to ask you about something regarding mapgen v7. There's a feature that would be really awesome and useful if you wish to consider adding it... 17:18 Taoki http://forum.minetest.net/viewtopic.php?pid=95397#p95397 17:19 Taoki hmmmm: Basically, I'm wondering if you could add support for rendering multiple layers of ground levels vertically. So that we could divide the world and add more realms past a certain height (like how Minecraft has the Nether and The End, as different realms in the same world) 17:19 hmmmm we decided against realms already 17:19 Taoki When and why? 17:19 sapier "The world size of Minetest is painfully small I agree" ???? 17:19 hmmmm divisions in height are the way to go, because they're much easier to implement and don't require rewriting everything 17:19 sapier 32k blocks? 17:20 Taoki hmmmm: Yes, that's what I'm suggesting in the last post. Thought that's what you meant the team decided against 17:20 Taoki I suggested that in the first post, but like everyone said it would be hard to do 17:20 hmmmm i already am going to do this 17:20 Taoki So now I'm suggesting just that; Dividing past a certain height and generating a new world there 17:20 Taoki Awesome! 17:20 hmmmm we'll call a large region of height a realm 17:20 hmmmm each realm has a collection of biomes 17:21 sapier hmmmm you should add a "get_ground_level(pos)" function in that case 17:21 hmmmm i mentioned early on that get_ground_level is going to be more difficult to implement well 17:22 hmmmm i think we're sticking with passing an approximate range to look through 17:22 sapier yes but if there are multiple layers you need that function 17:22 Taoki My thought was that we could have a Lua parameter specifying how many realms you have. When that parameter is larger than 1 (therefore there are more worlds), the world boundaries are divided vertically and each squeezed / treated like a different realm. 17:22 sapier lua parameter? 17:22 sapier I guess this should be a world parameter 17:22 hmmmm it'd have to be a lua parameter 17:22 Taoki sapier: So each game (eg: minetest_game) can define how many realms they have and what biomes each use 17:22 Jordach you can divide 31000 into 10 different realms with 3100on y 17:23 sapier why lua? is number of worlds adjustable while a game is running? 17:23 hmmmm taoki, i'm not worrying over details at the moment, but they are definitely not going to be uniformly divided like that 17:23 Taoki Alright. That's ok if they do work :) 17:24 hmmmm anyway 17:24 Taoki hmmmm: Another suggestion for this feature (which I'm not sure the mapgen can do but asking either way) is to block seeing and walking between realms. So if you either fly all the way up or dog too much down, you don't pop into the other realm. And instead see everything stop loading like you do currently when you reach world bounds 17:24 Taoki But that might have to be implemented separately 17:24 hmmmm a "realm" will be defined as a range in height 17:25 Taoki Since well, we wouldn't want the risk of digging down so much or flying up so high you get to the other realm :P 17:25 Taoki nice 17:25 hmmmm each "realm" gets its own generateTerrain(), and its own collection of biomes 17:26 hmmmm so there'll be the overworld "realm", the aether "realm" which will be positive 3d noise (the floating islands), probably occuring around 7000 or so 17:26 hmmmm below the overworld will be the "mantle" realm which is basically just a bunch of large sheets of iron blocks and lava sources and things like that 17:27 hmmmm then underneath that is the nether 17:27 hmmmm so you can see the need for separate generateTerrain() functions, since they each do something quite different 17:27 Taoki hmmmm: Are all those going to be hard coded however? I was hoping that Lua could define if there are multiple realms, the height of this one and the biome set 17:27 hmmmm they're gonna have to be hard coded. 17:28 Taoki ok. That's not really the feature I was thinking of them :/ 17:28 sapier why do they have to be hardcoded? 17:28 hmmmm because they're components of the mapgen 17:28 hmmmm so unless you want to define the mapgen in lua ... 17:29 Taoki hmmmm: But couldn't Lua define where each type of mapgen is used (between which heights)? And allow defining any set? 17:29 hmmmm i honestly cannot see the point in that 17:29 hmmmm i mean, sure, i guess you can define where they are and if they occur or whatever 17:29 Taoki hmmmm: It would allow any game to define its own set of realms and setup it sees fit 17:30 hmmmm but you won't be able to actually change it 17:30 Taoki I'm actually thinking of using that 17:30 sapier but mapgen could use a generation range couldn't it? 17:30 Taoki That's what I'm thinking 17:30 hmmmm i'm saying that you wouldn't be able to change the method of generation 17:31 hmmmm that means the aether is still going to be floating islands generated in the same way, the most you'll be able to do is define between which heights it spawns and if it spawns at all 17:31 Taoki hmmmm: Main feature I'm suggesting is the ability to define diggerent ground levels and ranges at different heights. So the mapgen can eg: generate a land at height 1000 just like it does at height 0. Using the same noise function etc 17:31 hmmmm and of course the blocks and whatever, but that's a given with the current infrastructure 17:31 sapier yes hmmmm but you could specify the generation mode as world parameter 17:32 Taoki Same way you explore the normal existng world. Just replicate the same generation rules at some very big heights 17:32 hmmmm sapier, if you mean you want to define another "nether-like" realm, then of course 17:32 hmmmm you can define as many of those as you'd like 17:32 sapier no I was more thinking about games wanting different realms or even only a single one 17:33 hmmmm you can do that too 17:33 sapier if it's hardcoded how? 17:33 hmmmm that detail wouldn't be hardcoded 17:33 hmmmm what would be hardcoded is the generation technique 17:34 hmmmm you'd have a limited selection of things like "nether-like" and "aether-like" and "regular terrain-like" 17:34 sapier what is meant by "generation technique" maybe this is just a matter of understanding 17:34 sapier ok and specify which realm uses what technique in world settings? 17:34 hmmmm yes 17:34 Taoki hmmmm: So imagining the existing world. The ground level is usually at height 15 - 20, and water level at height 0. As you explore, new hills mountains etc. are generated at this height. My idea is a Lua parameter to allow doing the same high in the sky. So at height 1000 there would be a new water level, the ground level would usually be around height 1020, and so on 17:35 sapier and then specify different biomes for different techniques? 17:35 hmmmm yeah 17:35 hmmmm here's the complicated factor 17:35 hmmmm uninstall a mod, the terrain becomes inconsistent 17:35 hmmmm there's a reason why i did what i did with putting noise parameters in map_meta.txt 17:35 sapier hmmmm that happens atm too 17:36 hmmmm and people are okay with this???? 17:36 Taoki If you uninstall a mod that contained terrain block definitions, it's logical you have to delete that world too :P 17:36 sapier uninstalling mods always gives negative results sometimes bigger sometimes less big 17:36 hmmmm you can redefine them just as easily 17:36 Taoki hmmmm: The mod uninstallation issue? Sure! It's the server admin's fault if he used block definitions for teerrain he now has to take down 17:36 Taoki That's good too 17:37 sapier i don't think that problem can be fixed by engine at all 17:37 hmmmm my problem is that once it's generated, it's generated and the damage is done 17:37 hmmmm you can re-enable a mod and have the unknown nodes back again 17:37 PilzAdam hmmmm, since we have this per world mod system now it would be ok to let mods define it 17:37 Taoki yeah 17:37 sapier yes and where's the problem? if an artist cuts his stone block wrong damage is done 17:38 sapier no way to glue it back without lots of additional work ;-) 17:38 hmmmm i would be more okay with this if we had a good way to forcefully regenerate blocks 17:38 sapier hmm what about world_drop(min,max) 17:38 Taoki A system to delete and re-generate specific parts of the world would be great. But can be a different feature for later on 17:38 Taoki sapier: Yep, that 17:39 hmmmm sapier, what's that, trim the world to a specified bounding box? 17:39 sapier no that should drop every block within min,max 17:39 Taoki That's a common issue in Minecraft. There are those servers that have huge cities people worked on. Then they all get deleted because a new version of MC came which adds new stuff to the mapgen, which the existing world can't get 17:40 hmmmm I think there should be both behaviors 17:40 sapier return value would be exact coordinates of block boundaries beeing dropped as it's quite unrealistic user will exactly hit those 17:40 hmmmm taoki, i believe there is a way to have new ores added to old blocks 17:40 Taoki With such a function, epople wouldn't have to delete their cities... just the unexplored areas and let those regenerate :) 17:40 PilzAdam hmmmm, what about hashing the most relevant variables for world creation and check if it matchs the current setup at each startup 17:40 Taoki Nice, didn't know that 17:41 sapier this problem isn't limited to ores 17:41 hmmmm pilzadam, i think it might be okay if there's a way to have blocks regenerated 17:41 sapier every mod using on_generated has same problem 17:41 PilzAdam but where do you know wich blocks you have regenerate? 17:42 hmmmm easy 17:42 hmmmm generate the whole chunk 17:42 hmmmm only blit to blocks full of content ignore 17:42 hmmmm or isDummy 17:42 PilzAdam I guess this problem would be trivial to fix if we had this "only save modified blocks" feature 17:42 hmmmm we should do that too sometime 17:43 sapier hmm isn't threre a dirty flag used for network transmission? could this be used for saving too? 17:43 hmmmm i wouldn't know. 17:44 Taoki hmmmm: Anyway, here's a better example as to why I'm suggesting de-hardcoding realms to Lua: Let's say I make a game type where I want players to go between Earth, Mars and a Moon realm. In all 3 cases, the ground would be generated by the same mapgen type, only different biomes. But you'd need to define it to run at 3 heights for each 17:44 PilzAdam anyone objects: https://github.com/minetest/minetest/pull/743 ? 17:44 hmmmm i think the only 'dirty' flag is being not present in the sent blocks set 17:44 hmmmm pilzadam, i thought that was merged a while ago 17:45 PilzAdam it was blocked because 0.4.7 freeze 17:46 Taoki But yeah, I hope you consider this for v7. Would be really useful for some ideas :) 17:46 hmmmm i think for a moon mod it'd be cool if there were multiple aether-like realms defined 17:46 hmmmm and they're just huge blobs of 3d noise with a *very* large spread factor 17:47 Jordach at least we have an *actual* sky biome before the leading product 17:47 Taoki Aether = floatlands? 17:47 hmmmm i guess 17:47 hmmmm it's going to look a lot nicer than the floatlands thing 17:48 Taoki Ok. Wouldn't work for a moon or mars realm in that case, since those would have to be constant ground just like the normaal world's surface (not floating islands) 17:48 Taoki Hence why mapgen would need to allow re-running the ground generator at another height 17:48 hmmmm then in that case 17:48 hmmmm nether, with the air blocks replaced with solid blocks, and solid blocks == air 17:49 Jordach common sense right there 17:49 Taoki I... don't know how that would look either 17:49 hmmmm how about you not worry about how something looks when it's far, far from completion 17:50 Taoki I tend to imagine something like it's already there a lot of the time :P 17:50 hmmmm i know people have big ideas but i'd rather just focus on getting things done instead of yapping away about some far off features 17:51 Taoki True... there's more important things to finish on v7 now. Although it might help to discuss such features early before the code gets more complex 17:51 hmmmm sort of reminds me of something a futurologist would do.. you know, the kind of people fascinated with home automation, flying cars and the like 17:51 Taoki heh, that sounds like me too :) 17:51 hmmmm and they don't shut up about THE SINGULARITY 17:51 hmmmm kurzweil basically 17:52 random_esrdxctfg hi. any specific reason why boost is never used? 17:53 PilzAdam random_esrdxctfg, what boost? 17:54 PilzAdam hmmmm, https://github.com/minetest/minetest/pull/778 <- what about that? the format isnt "wrong", but it restores the old behaviour 17:54 PilzAdam (an earlier commmit changed it a bit) 17:54 proller libof>>>> 17:57 sweetbomberfdgdf /nickserv ghost sweetbomber 17:58 PilzAdam is the license compatible with LGPL? 18:01 ShadowNinja sweetbomber: Make sure to type identify commands and such into the status window. 18:04 sweetbomber ShadowNinja: lol, i was lazy. almost leaked my pass. ty anyway 19:45 hmmmm pilzadam, don't really have anything against it..... 19:46 VanessaE PilzAdam: that fix should be pushed, though plants_lib has already been patched to avoid it per your pull request 19:48 PilzAdam VanessaE, yes, that commit is unrelated to any "compatibility" problems; its just a comsetic fix 19:52 lrh9 Hello. 19:53 PilzAdam lrh9, this channel is not for "Hello", go to #minetest 19:55 kaeza nice way to greet newcomers 19:55 hmmmm kinda says in the topic, "chit chat goes to #minetest" 19:56 VanessaE agreed. This channel is for core development discussion but saying "hi" doesn't constitute chit-chat. 19:56 PilzAdam he did it already several times 19:56 PilzAdam without anything useful 19:56 hmmmm i saw him yesterday, true 19:56 VanessaE so? 19:56 VanessaE how does that detract from the conversation -- especially right now when the channel barely had anything going on>? 19:57 hmmmm it's just kind of irritating 19:57 hmmmm personally i'd just ignore it though 19:57 PilzAdam I get highlighted for each message here, switching away from what Im doing to just see a "hi" isnt really good 19:57 kaeza then configure your client? 19:57 VanessaE PilzAdam: then fix your IRC client not to highlight you unless someone says your name. 19:58 hmmmm that must be the single most annoying irc client ever 19:58 PilzAdam I configured it to highlight me on each message and I dont want to change that 19:58 PilzAdam because I expect messages here to be important for development 19:58 VanessaE then don't complain if you get irritated at every little message that comes across. 19:59 PilzAdam we should add ""How's it going?" = generally a /kick" back into the topic 19:59 VanessaE no, we should not. 19:59 VanessaE being an asshole is not a prerequisite for being a core dev. 20:00 hmmmm other FOSS project -dev channels aren't like that 20:00 sapier no we're way to nice compared to other foss projects ;-P 20:00 hmmmm yeah that's true. but they wouldn't kick you for saying hi 20:01 sapier true too I didn't want to support pilzadams oppinion with what I said 20:01 sapier btw I haven't asked for formspec mainmenu to be added today .... done :-) 20:02 PilzAdam sapier, is that dynamic_cast thing fixed? 20:02 sapier it doesn't occur anymore :-) 20:02 sapier fix has to be done in gentoo/irrlicht depending who you ask 20:03 Exio is that patched in archlinux? 20:03 Exio i mean, they are pretty vanilla too 20:03 PilzAdam I kinda agree what you have said some time ago: We cant find all bugs when we keep it seperated, so merge it now and test things in upstream 20:03 sapier i don't have any Idea but fedora suse redhat debian all fix it 20:05 hmmmm am i the only one who does this: 20:05 sapier pilzadam that was before someone brought up an early 0.4.8 release ... mainmenu needs some time to get fixed completely so if 0.4.8 is in e.g. 2 weeks that's not best thing to do 20:05 hmmmm when i am punching out code, i just put a placeholder for writing an error message because it takes too long to think of something helpful and informative 20:06 hmmmm for example, errorstream << "DecoSchematic::resolveNodeNames: blahblah" << std::endl; //////////////// 20:06 PilzAdam sapier, I havent heard anything of an "early 0.4.8 release" 20:06 hmmmm pilzadam, i brought up the idea yesterday. 20:07 sapier someone wrote that related to that big performance fix 20:07 PilzAdam must have missed it in the logs 20:07 hmmmm exio's patch is pretty amazing and i feel that the players ought to have that right away 20:07 Exio the big performance fix is just removing a line that was explicity-said is useless 20:07 sapier obviously it's useless and harmfull 20:07 PilzAdam Exio, you should grep through the code and search more of those ;-) 20:08 hmmmm minetest would be smooth as butter now if only that choppiness spike was fixed 20:09 sapier choppiness spike? 20:09 hmmmm i think it happens in the render thread, but i can't be sure. 20:09 PilzAdam hmmmm, if you really want to release 0.4.8 that fast then at least finish the decoration stuff (i.e. add it to indev mapgen too) and use it in minetest_game 20:09 hmmmm yeah, every half second 20:09 hmmmm PilzAdam, that's the current plan. 20:10 PilzAdam and we need to ask other devs, I am undecided 20:10 sapier but plz add the mainmenu right after release ... the formspec improvement comming along with it are very very usefull 20:10 hmmmm btw there's no such thing as an early release 20:11 hmmmm these are patch version levels we're changing here 20:11 hmmmm we can make releases and bump that number as often as we'd like 20:12 sapier we don't really use the possibilities our version scheme would provide ;-) I guess we would be at 0.10.0 if we did with some 0.x.y versions 20:21 hmmmm i had to write a generalized and simplified version of getBlockNodeIdMapping()... where should that go, you think? 20:21 hmmmm it's not part of any class 21:18 PilzAdam sapier, would you agree to merge https://github.com/minetest/minetest/pull/773 ? we can turn the object collision for falling nodes back on later 21:19 Exio i agree by now 21:19 sapier I already told my oppion about adding this new property 21:20 PilzAdam oh, I though all your comments were about the dropped items and falling nodes 21:20 PilzAdam +t 21:20 sapier as a property added once is most unlikely to be removed anytime soon don't call it a temporary fix 21:20 PilzAdam the option itself should stay there forever 21:21 sapier no imho this commit is a workaround for a problem that should be fixed correctly 21:21 PilzAdam why dont give modders the ability to decide? 21:21 PilzAdam you forve everyone to use all collision or none at all 21:21 PilzAdam *force 21:21 sapier because they already can decide this property is almost same as "physical" 21:22 PilzAdam turning physical off will also remove the object node collision 21:22 VanessaE PilzAdam: he's right in that the bugs should be fixed to begin with, but we still need this setting - think of dropping craftitems. Why should those *ever* stack up in practice? 21:22 PilzAdam and dropped items should never collide with objects, but they have to collide with nodes 21:23 VanessaE *nod* 21:24 sapier if I read "collide_with_objects" I don't have any idea what this means ... objects aren't a concept within lua 21:24 sapier there are nodes and entities 21:24 Exio collide_with_other_entities_around 21:24 Exio :D 21:25 sapier then rename physical to collide_with_everything 21:25 PilzAdam http://dev.minetest.net/Terminology 21:25 PilzAdam "Object - something in the world that is not attached to any particular node or position, can move and act on its own. Examples include mobs, dropped items, falling sand or gravel, primed TNT. Players are also objects. " 21:25 sapier or make collision handling a value not beeing bool 21:25 sapier what happens if someone sets physical to false but collide_with_objects to true? 21:26 PilzAdam thats documented 21:26 PilzAdam "collide_with_objects = true, -- collide with other objects if physical=true " 21:26 sapier we have two parameters specifying same thing but depending on each other 21:27 PilzAdam making it a string field would break compatibilty 21:27 PilzAdam +i 21:29 sapier so we add a hack ... imho this is wrong but you'lll do it no matter what I say 21:29 sapier sane solution adding this hack would declare physical deprecated 21:29 sapier and add a new parameter collision_type (none,nodes,everything) 21:30 sapier at least thats my oppinion 21:59 hmmmm hey guys, anybody know of a function that sorts v3s16 members? 22:00 hmmmm e.g. p1 = {1, 2, 3}, p2 = {-1, 5, 6} sorted would be p1 = {-1, 2, 3}, p2 = {1, 5, 6} 22:02 ShadowNinja hmmmm: Look at WorldEdit. 22:02 hmmmm no.. i mean a function in minetest that already does that 22:02 ShadowNinja No.