Time Nick Message 01:12 MTDiscord quaso 🥐 12:34 Blockhead256[m] I don't want to self-promote too much but I have made some pretty big posts to the forums the last two days, and I know they're big and will take a while to read and respond to (and most of you are probably asleep or at work right now) but I'd really like to get some responses rather than write a wall of text that nobody... (full message at 12:34 Blockhead256[m] ) 14:42 MTDiscord I think, based on how much has already been said on the subjects, for a number of the issues you've raised like dependency/compatibility hell and for proliferation, instead of trying to solve the problems, you'd probably get more traction pursuing a formal proof that the problems are unsolvable. 14:44 MTDiscord In a way, it seems like this is what the game experience of MTG really IS. In some games you fight monsters, in some you solve puzzles, in some you build architecture or sculptures. In minetest_game, you try to build a working and enjoyable mod soup, while battling against mod incompatibility, undeclared dependencies, and trying to avoid damage from distractingly mismatched art style and gameplay design. 15:00 muurkha haha 15:04 Blockhead256[m] fun fact: if you move your mouse fast enough on a small Minetest window (at least on Linux/X11), your cursor can escape the Minetest window and focus on other windows 15:05 potatoxel[m] :o tomatoful 15:14 MTDiscord I actually run into that "escaping the mouse prison" issue on OSX quite a bit. Happens much more easily for me there than on Linux. 15:16 MTDiscord Could be because pointer acceleration is insanely high on Mac even when you use a special hack to set it below the system minimum. Could also be because you're using a computer that costs 2500 as much to get 30% better performance so it can spend 75% more effort rendering unnecessary UI eye candy. 15:16 MTDiscord 2500% ... (2500x would be a bit absurd) 15:18 MTDiscord only a bit 17:05 Peter_Lankton if i have a world that is a simple singlenode world, is there any server performance advantages or disadvantages if i set the chunksize in in map_meta.txt to 1 instead of the default 5? 17:23 sfan5 not really 17:23 sfan5 best leave it at the default 18:01 MTDiscord Well one advantage it would offer you is that if any of the mods you're using are built around the assumption that chunk size is always 5, then you'll be one of the first to know. 18:08 Peter_Lankton true lol. actually is there even a way for mods to know? im guessing they could check the minp and maxp on the minetest.register_on_generated thingy 18:10 MTDiscord You might also be able to read it out of settings. I should really experiment with it a bit more and see what happens if it's set to like 4 or 8 or 7 or something. 18:11 RhineDevil How could I write a mod that starts land generation automatically (like emerge for mg_earth) and searches for a zone satisfying my requirements? 18:12 RhineDevil Like, I start a world, write a command and it starts generating 18:12 RhineDevil Kinda like the algorythm for finding nyan cats in 0.40 versions of minetest 18:13 Peter_Lankton ive been experimenting a bit with it at lower values like 1. thats because i am also experimenting with a mod i made that spawns a structure onto each mapchunk if a certain condition is met. i have kind of made a 3d graphing calculator out of it lol. each structure is listed in a table like this: 18:13 Peter_Lankton ["schem.mts"] = function (v,seed) return v.y == math.round(math.sin(v.x)) end, 18:13 MTDiscord Emerge area and callbacks. I did one like this for finding a teleportation destination in a new dimension that meets certain requirements. It's a bit slow so you'll have to deal with the possibility that the destination isn't ready yet. 18:15 RhineDevil Could you give me public code examples? 18:15 MTDiscord Mapchunk size is balanced by default for a good trade off between latency and overhead/efficiency on, I assume, some idea of an "average" system. Users pretty rarely change it, and I've seen various mods that rely to different degrees on it being 5 (from being a bit less optimized if not, to outright giving wrong results). 18:16 MTDiscord Rhine: I don't have any public ones, myself, but maybe someone else might. 18:17 MTDiscord The only public code example I have for finding a remote teleport destination right now is NodeCore skyrealm, which uses a different method, i.e. suitable locations are just generated deterministically so it's trivial to know where they will be, even if you can't generate them yet. 18:18 RhineDevil Warr1024 what if you publish it on github 18:19 MTDiscord Yeah, I may consider doing so in the future, depending on what direction it goes. 18:19 RhineDevil Warr1024: are there reasons holding you back? 18:20 MTDiscord Most of my projects are FOSS by default, and they tend to start out public when they're something I just do by myself and I don't need anyone else's permission.