Time Nick Message 01:58 ShadowNinja rubenwardy: Yeah, I just pushed a patch to address the comment sfan5 had on my fixes, should be good to merge now. 02:10 rubenwardy !title https://github.com/rubenwardy/minetest/commit/b704a280e1d74bcd8b5dc3a3df968ba650ac374d 02:10 ShadowBot rubenwardy: Add link to web version of lua_api.txt · rubenwardy/minetest@b704a28 · GitHub 02:10 rubenwardy trivial PR to remove link to dev wiki, and add link to web formatted lua_api.txt 07:08 sfan5 lgmt 07:08 sfan5 s/mt/tm/ 13:38 rubenwardy merging #12015 in 10 13:38 ShadowBot https://github.com/minetest/minetest/issues/12015 -- [no squash] Various minor fixes by ShadowNinja 13:47 rubenwardy sfan5, Krock: are you around for a meeting at 7pm UTC on Sunday? 13:47 sfan5 thats pretty late 13:47 sfan5 but sure 13:49 rubenwardy I can do earlier 13:49 rubenwardy I'm free all evening, if you a preferred time 13:50 sfan5 6pm utc? 13:52 rubenwardy sure 13:56 MTDiscord ey. there is possible to export 3d figures to minetest nodes? 13:57 MTDiscord means - from files, not to generate cybes from images 14:07 MTDiscord This IRC channel is for Minetest engine dev discussion, not for general chat. Message edits also don't work over IRC. (For MT->mesh there's https://content.minetest.net/packages/random_geek/meshport/, dunno about the other way around.) 14:26 MTDiscord ah, ok nvm 14:29 MTDiscord just that i was asking if core have a function to use 3d art... (not cubes) 14:36 sfan5 you can use meshes instead of "cubes" for a appearance of a node 14:36 MTDiscord not this... from a file.. 14:38 sfan5 and what is that file 14:38 MTDiscord from 3d editor 14:39 MTDiscord (blender) 14:41 MTDiscord .obj file 14:41 sfan5 minetest supports that format 14:41 MTDiscord ? 14:42 MTDiscord didnt knew.... 14:42 MTDiscord thought it asembles 3d objects from 2d images 14:44 rubenwardy weird assumption 14:44 MTDiscord cause why sometimes im seeing non-filled nodes far away? looks like asembling 3d art from 2d squares 14:45 MTDiscord like this xscreensavers 15:30 erle hacknorris what do you mean with “non-filled nodes”? 15:30 MTDiscord empty, made like origami 15:30 MTDiscord (yep i know a little japanese cuture) 15:30 erle please provide a screenshot 15:31 MTDiscord its irc lol. how? 15:31 erle upload it to https://mister-muffin.de/paste 15:31 erle then post the link here 15:31 erle there are some circumstances, in which the engine does render stuff wrong, but your origami thing does not ring a bell 15:33 erle sofar i recently encountered a situation on “inside the box” where i stood for a long time doing nothing and when i turned around that part of the box that i had not watched had unloaded and never loaded again. do you think this a minetest client, minetest server, or “inside the box” bug? 15:33 erle “a long time” being like 20+ minutes or so 15:35 MTDiscord means - not that lag im just asking if its possible to use .obj files (filled) while nodes are cubes made out of images, not 3 art 15:37 MTDiscord and this (no its not a bug, just a cheap laggy lap) : https://mister-muffin.de/p/ynxu.png 15:38 MTDiscord at the ends... its not a 3d cube filled with "water" just a 3d cube made out of "paper" which is just 6 images... 15:38 MTDiscord and im thinking about 3d art 15:40 erle hacknorris your screenshot is perfectly okay, i don't see the problem. where is the bug? 15:40 MTDiscord hacknorris: 3d rendering is always surface based 15:40 MTDiscord (or almost always) 15:40 erle hacknorris if you want advice on how to make nodes that have an obj model, let's discuss this in #minetest – i can help you. 15:41 MTDiscord you can specify fill colors for when the viewer is inside the node 15:41 erle there are multiple techniques 15:41 MTDiscord backface culling etc. might expose that it's implemented this way 15:41 MTDiscord its possible to make a sphere in minetest? 15:41 MTDiscord but OBJs are no different, they are based on tris as well 15:41 MTDiscord hacknorris: yes, by approximating it using triangles 15:42 MTDiscord ? 15:42 MTDiscord Technically even a sprite would work 15:42 MTDiscord hacknorris: you need a sphere 3d model 15:42 MTDiscord i can make 15:42 MTDiscord i have some apps 15:43 MTDiscord this is the way "little lady" was done? or what? 15:43 MTDiscord (from contentdb 15:43 erle hacknorris, go to #minetest 15:44 MTDiscord saw 16:25 rubenwardy #12185 16:25 ShadowBot https://github.com/minetest/minetest/issues/12185 -- Separate register and login by rubenwardy 16:30 MTDiscord that is awesome, way more intuitive than the old system 16:31 erle > One caveat is that it's possible to register from login, and login from register. This will require network protocol changes to fix 16:32 erle rubenwardy could you elaborate why it would require the network protocol to be changed? 16:32 erle (ideally in the PR description, so others can see it there, too) 16:45 rubenwardy erle: "One caveat is that it's possible to register from login, and login from register." 16:45 rubenwardy I believe network changes are needed to fix this 16:45 rubenwardy updated PR description 16:55 erle rubenwardy i still don't get what exactly you want to achieve here. this is how i imagine it: to figure out if a player name is taken, you use TOSERVER_INIT and see what you get back. 16:56 erle i am not seeing a case here where the network protocol needs to be changed, only a case where the handling of the responses needs to be changed depending on context 16:57 rubenwardy well, that's up to sfan5 really 16:58 erle still, i might have missed some side effects 16:58 erle that is why i am asking 17:09 erle rubenwardy my proposal to check if a username is available, you can look at the TOCLIENT_HELLO. if the registration mechanism is AUTH_MECHANISM_FIRST_SRP, the username is available. otherwise, it is not. however, i have not thought about this too long. 19:02 sfan5 I got bit by a known bug in our thread implementation again 19:02 sfan5 if a thread immediately exits then start() misses it and continues waiting for m_running to become true 19:17 rubenwardy would probably be worth moving to std::thread 19:17 rubenwardy oh wait, it is a wrapper over std::thread?! 19:17 sfan5 yes 20:20 sfan5 good news: the hard parts of #11131 are all done 20:20 ShadowBot https://github.com/minetest/minetest/issues/11131 -- Async environment for mods to do concurrent tasks by sfan5 21:36 Zughy[m] (don't forget to link the issue sfan5, right now it's not linked) 21:52 sfan5 okay so here's the question 21:53 sfan5 should I call the in-memory thing that transfers Lua objects "serialization" or "packer"? or something else? 22:02 erle well, what do you want to convey? 22:02 erle serialization implies deserialization, packing implies unpacking 22:03 sfan5 I want to convey that it does not pack data into a byte buffer that could be put away somewhere, it's an in-memory representation 22:12 MisterE[m] holo_trans? 22:13 MisterE[m] Cerealpack? 22:15 MisterE[m] More seriously, "luaconveyor"? 22:16 erle sfan5 in what way does it matter that it is in memory? does it contain ephemeral pointers? is there some cow magic going on? 22:17 erle suggesting something similar to serialization is probably a bad idea it can not be “put away somewhere” (i.e. written to disk or so) 22:24 sfan5 in-memory = instance of a class, e.g. you cannot take one object and write it to disk