Time Nick Message 05:27 paramat #5512 trivial mapgen stuff 05:27 ShadowBot https://github.com/minetest/minetest/issues/5512 -- MapgenBasic: Add lava source as commonly used content by paramat 05:39 nerzhul okay for me 05:39 nerzhul paramat, can look at trivial #5510 05:39 ShadowBot https://github.com/minetest/minetest/issues/5510 -- Client handlers: Remove useless stringstream usage in two handlers by nerzhul 05:41 nerzhul merging #5508 05:41 ShadowBot https://github.com/minetest/minetest/issues/5508 -- [CSM] Add local node meta reference. by red-001 05:42 paramat #5513 more trivial mapgen stuff, will merge later with 5512 05:42 ShadowBot https://github.com/minetest/minetest/issues/5513 -- Mapgen: Remove '#include treegen.h' from non-mgv6 mapgens by paramat 05:43 paramat i don't understand 5510 but if it's trivial go ahead 05:46 nerzhul in fact we use a stringstream to deserialize long string instead of asking packet to do it, this copies the memory space of this packet just for deserialize, i remove the copy and use native packet feature :) 05:47 nerzhul merging then 05:51 paramat ok :] this is stuff i'm not good with 05:51 nerzhul merging 5512 05:54 paramat thanks 06:17 VanessaE paramat: I wonder if you shouldn't leave #5107 open until that issue is actually resolved? 06:17 ShadowBot https://github.com/minetest/minetest/issues/5107 -- Make nodes coloured by param2 distinct 07:15 paramat nah it's ok, work on coloured itemstacks is well underway :] 07:16 VanessaE ok. 10:37 kilbith nrzkt: https://hastebin.com/kedesamaqi.cpp 10:38 nrzkt kilbith: it's a problem when you compile with a C++11 compiler and use jsoncpp for non C++11 build, just ignore it or update jsoncpp to 1.8.0 to use C++11 instead of C++03 10:39 kilbith jp@r2d2:~$ yaourt -Qi jsoncpp 10:39 kilbith Nom : jsoncpp 10:39 kilbith Version : 1.8.0-1 10:42 nrzkt here you use embedded jsoncpp instead of system 10:42 nrzkt it's why you compile it and have warns 10:44 kilbith how do I use the system one ? 10:47 nrzkt open CMakeCache.txt and set ENABLE_SYSTEM_JSONCPP to "on" 12:32 burli hi 12:35 burli ported the pathfinder to get_node() successfully. Still fast. Now it should be possible to port it to C++ much easier 12:56 celeron55 sounds good 15:08 glut32 Hi all 15:13 glut32 I'm working on some new features for the FormSpec. I added a new parameter for the textarea and inside the method GUIFormSpecMenu::parseTextarea(...). The problem is that I have a warning that I would like to remove. The warning message is "field/textarea: use field_close_on_enter[name, enabled]..." and there a comment "// TODO: remove after 2016-11-03". Can I remove this warning? 15:38 nerzhul can you give us the commit in blame ? 15:41 sfan5 nerzhul: https://github.com/minetest/minetest/commit/067766eec213918b6cb5b2533d0c78eceb3949ec 15:43 nerzhul sfan5: ty 15:44 nerzhul ruben ! 16:39 kilbith there's no option to change the font color globally ? 16:40 DS-minetest kilbith: can't you make the font colored? 18:36 burli https://youtu.be/GU7VeVgsGOs 18:36 burli Pathfinder in real terrain. Sometimes the mobs get confused. Not sure why 18:37 burli Average time per path is around 1ms 18:39 burli And I use stepheight 1.1 because jumping in mobs redo is buggy 18:39 red-001 could you make it take the hight of the mob into account? 18:40 burli Yes 18:40 burli I know that's still a bug 18:40 red-001 is this written in lua or c++? 18:40 burli There are a few things I have to change. 18:41 burli This is lua 18:42 burli red-001 this should be ported to C++ later. 18:43 red-001 nice work btw 18:43 burli First I want to test different algorithms 18:43 burli The code is still rough and straigt forward 18:44 red-001 how does it compare to the bultin pathfinder right now? 18:45 burli In speed or quality? 18:45 red-001 both I suppose 18:45 red-001 I guess it's a bit slower since it's written in lua? 18:46 burli the builtin is much faster. But the results are horrible. Let me show you something 18:46 burli This is the builtin pathfinder https://www.youtube.com/watch?v=OUsW2M85oYg 18:47 burli and this is my version from yesterday https://www.youtube.com/watch?v=kMqiE2frmeg 18:48 burli jumping is disabled and only flat terrain is scanned 18:49 red-001 neat 18:51 sofar I'm pretty much convinced that a lua pathfinder would be far more flexible than any C++ one and fast enough that it doesn't matter 18:51 sofar once you have read all the map data into lua, lua is nearly as fast as c++ 18:51 sofar at least in the same order 18:51 burli I use get_node() 18:52 sofar and you can do much more with the data you have available 18:53 burli I used the voxelmanip before. But it takes time to load it. And you can't cache it because the data is invalid if the map has changed 18:54 burli And it was not faster than my current version with get_node() 18:55 burli I can speed up the algorithm. But first the A* algorithm has to be stable 18:58 burli I highly recommend to port the pathfinder to C++. Why should a Lua version more flexible? 19:00 burli *be more flexible 19:01 burli damn 19:02 burli sofar, what kind of data do you mean? 19:32 sofar burli: node information like whether something is a door or not (and thus can be opened) 19:34 burli sofar, such things are on my todo list, but I'm not sure if a pathfinder is the right place for this 19:35 sofar well, if you want to path to e.g. a bed inside a house 19:35 sofar then searching for a path to a door is pretty senseless 19:35 sofar so what is needed is a list of "consider these nodes passable" that can be provided to the path finder 19:36 sofar or better, "nodes with these properties" 19:36 burli I understand what you mean 19:54 burli doors are a big problem in general. Mobs can't see me even if the door is open 19:55 burli mobs see doors as solid node, no matter if open or closed 19:55 benrob0329 Perhaps a walkable identifier 19:56 burli Currently I just read all nodes and store all walkable nodes in a list 19:57 burli walkable nodes are blocking nodes for the pathfinder. You can walk ON them, but not THROUGH them 19:57 benrob0329 Then a passivle identifier? 19:58 benrob0329 Make them seen as air 19:58 benrob0329 *passible 20:02 celeron55 the pathfinder should definitely have a way to specify a list of nodes that are passable 20:02 burli not sure how to solve that. I assume that I have to rewrite the code a few times 20:03 burli celeron55, I aggree. I have this on my list, but it is to early for this to think about 20:04 celeron55 yeah, keep doing whatever is importnat 20:04 celeron55 important* 20:04 celeron55 is the code available somewhere for testing? 20:04 burli maybe using groups, at least for something like doors 20:05 burli celeron55, https://github.com/MarkuBu/pathfinder 20:06 burli but it is not the current version and it needs some modifications in mobs_redo 20:06 burli so it is not ready for testing 20:09 burli I'm sorry for the long code lines. I just ignore that for the moment 20:15 burli just FYI: this is the first version with voxelmanip. I'm surprised that it is slower https://github.com/MarkuBu/pathfinder/blob/voxelmanip/init.lua 20:58 Fixer fps drop bug at hometown server, tried clients 0.4.15+ and up to 0.4.14, i guess it is time to reproduce it in singleplayer somehow :( 20:59 Fixer will try 0.4.13 client 21:04 Fixer interesting, when I have those large fps drops some people also moan about lag, weird, coincidence? don't think so (c) 21:14 Megaf Fixer, MegafServer has major FPS drops around spawn 21:14 Megaf but that's due to minetest not culling caves and things at all 21:15 Fixer Megaf: no, that is different problem 21:29 red-001 #5516 21:29 ShadowBot https://github.com/minetest/minetest/issues/5516 -- [CSM] Add support for positional audio. by red-001 21:30 red-001 not sure if gain is working correctly 22:07 sofar red-001: so, hum, good start, what's the problem with max_hear_distance though? 22:18 red-001 it seems to be server-sided right now so it will be harder to add 22:20 red-001 not sure if it even needs to be added 22:21 nerzhul red-001, what do you mean ? 22:23 red-001 It seems that all it does is check is the player within the range before sending the sound packet 22:23 red-001 which could easily be done by lua 22:24 nerzhul but for CSM you just check is player is near sound, no ? 22:24 red-001 ? 22:25 nerzhul you talked about implementing positional sound in CSM ? 22:25 red-001 yes 22:35 VanessaE will someone PLEASE fix the rollback command? it's horribly slow (as in several minutes just to go back 45 seconds in a huge log) 22:53 Megaf VanessaE, not gonna happen 22:54 twoelk devs don't need the rollback command, so why enhance it? :-) 22:55 Megaf I've been running my server for two years now without rollback 22:55 Megaf best thing I've ever done on it 22:57 VanessaE Megaf: I have it turned on so that I can at least identify and ban griefers, but the rollback (repair/undo) part itself is useless 22:57 Megaf VanessaE> Megaf: I have it turned on so that I can at least identify and ban griefers, 22:57 Megaf that's a very useful functionality of it indeed. 22:58 Megaf that function could be moved to it's own tool IMHO 22:58 Megaf something like /inspect (punches node) 23:00 VanessaE nerzhul I believe suggested adding an index to the log, but I never got the chance to do so 23:38 VanessaE can I make a feature request please? I'd like to define node a:b, but have it use node c:d as the object to be displayed in the inventory 23:39 VanessaE that is, not an inventory *image*, but an actual node. 23:41 VanessaE for example, pipeworks' tubes are pretty small when displayed in the inventory. I'd like to define some oversized versions of them to be used for display, because inventory images get distorted when they're presented as wields 23:41 VanessaE (e.g. if the image is rendered in 3d perspective) 23:43 VanessaE in the case of something like technic, I would choose its 1m-long cable to be displayed (rather than defining an oversized node), since its default form is a tiny little ~0.25m cube.