Time Nick Message 13:53 shaft Can somebody explain to me why the stack count number is only affected by GUI scaling and the HUD it's displayed on by HUD scaling? If I increase GUI scaling the GUI will be GIANT while the numbers are still very small compared to my screen. It doesn't pose a readablity issue for me but it looks kinda off 14:06 rubenwardy Sounds like a bug 16:41 shaft Stupid question: why are MTG cactus blocks generated upside down in the world? The V6 cacti are not affected btw! 16:41 shaft With my high res texture packs the thorns point in the wrong direction. 16:44 shaft Is it an issue in the Minetest engine or in Minetest Game? 16:45 shaft I was so proud of my dedicated cactus top texture I applied with an override only to find out that some cacti have their nodes generated upside down and some do not 16:52 ireallyhateirc I believe there's a mapgen glitch on mapchunk edges that causes that 17:01 shaft Why would that affect cacti with arms but not cacti without arms? It's BOTH cactus grown with the sappling and generated ones. 17:02 shaft armless cactus not affected in V7 mapgen! 17:02 shaft I just found one 17:03 shaft Do I have to report the bug on minetest or minetest_game github? 17:04 MTDiscord could you check whether this is an issue with the https://github.com/minetest/minetest_game/blob/master/mods/default/schematics/large_cactus.mts schematic? 17:05 ireallyhateirc we get a rotation bug with naturalslopeslib in exile 17:05 Blockhead256 shaft: report it the minetest_game. If it's an engine bug, it can be escalated 17:06 ireallyhateirc on mapchunk y-edges slabs/stairs are usually rotated upside down 17:06 MTDiscord this need not be an engine bug 17:07 MTDiscord for example a modder could easily set a node content ID in voxelmanip but forget to update the corresponding param2 17:07 ireallyhateirc yeah I guess, just the cactus issue sounded familiar to the slab issue 17:07 MTDiscord so first report to the games; once you (perhaps together with the game maintainers) have confirmed that their engine usage (mostly param2s, i'd assume) is correct, escalate to the engine, as blockhead said 17:11 shaft How do I load an MTS file? Worldedit mod doesn't accept it. 17:12 MTDiscord worldedit should be able to work with schematics. make sure you're using the right command - worldedit knows two types of schematics (.we and .mts) 17:17 shaft /mtschemplace works the GUI doesn't seem to offer it 17:17 shaft / 17:25 shaft It's in the large_cactus.mts file and I filed a bug report: https://github.com/minetest/minetest_game/issues/3122 19:41 ireallyhateirc I'm writing a road system/city generation algorithm and I'm looking for an efficient way to make a 2D representation of an area called citychunk that's 10x10 mapchunks. So each citychunk is a 800x800 area. To store that data I made an array that has all 100 mapchunks of a citychunk, some of that mapchunks are shared by adjacent citychunks for overgeneration. To write data into that data structure I need to calculate 19:41 ireallyhateirc indices of 1. the mapchunk 2. node position in the mapchunk. My implementation however is slow as it takes ~1000ms to generate a citychunk. Any ideas, hints? 19:42 ireallyhateirc I store mapchunk data in an array like this: local chunk[x][z] = {...} 19:43 MTDiscord Generating 100 mapchunks in 1000 ms doesn't sound slow to me - that'd be 10 ms per mapchunk. i think that is hard to beat. are you already using the async env? 19:44 MTDiscord the async mapgen env that is 19:44 ireallyhateirc yes 19:44 MTDiscord IIRC even just filling a mapchunk with nodes can take something like 20 - 50 ms (these numbers may be wrong, but they should be the right orders of magnitude) 19:44 MTDiscord so i doubt that there is much optimization you can do. i'd recommend profiling with jitprofiler (https://content.minetest.net/packages/jwmhjwmh/jitprofiler/) though. 19:45 ireallyhateirc well, first I generate citychunk data - so the central citychunk and all 8 neighbors, this takes about 7 seconds 19:46 ireallyhateirc after that writing mapchunks takes 10-20ms 19:46 MTDiscord Why generate all the data ahead of time rather than lazily? 19:48 ireallyhateirc the citychunk is the smallest unit of city generation in my game, I need to generate the road layout first to place buildings later 19:49 ireallyhateirc though I could make an algorithm that doesn't need to start generating at the citychunk edge 19:49 ireallyhateirc then I could perhaps start in the middle and generate it lazily as you say 19:50 ireallyhateirc The goal of the algorithm is generating a random, yet map-wide road system 19:51 ireallyhateirc and then placing buildings also randomly and connecting them to the roads with streets 21:07 MTDiscord how do you enter the mapgen environment? are on_generated functions in it automatically? I read the docs but I still don't quite understand 21:08 ireallyhateirc gimmie a second 21:09 ireallyhateirc so here I call the mapgen script: https://codeberg.org/lord_of_the_dumpster/perfect_city/src/commit/92fc7a2f89ceb7d126b0af39d8380a5c660dc773/mods/pcity_mapgen/init.lua#L38 21:10 ireallyhateirc and then register_on_generated 21:10 ireallyhateirc https://codeberg.org/lord_of_the_dumpster/perfect_city/src/commit/92fc7a2f89ceb7d126b0af39d8380a5c660dc773/mods/pcity_mapgen/mapgen.lua#L150 21:14 MTDiscord idealist: https://github.com/MisterE123/luamap/tree/mglua2 21:19 ireallyhateirc what's the use case? 21:36 MTDiscord ah I see so it's minetest.register_mapgen_script 21:38 MTDiscord hmmm my minetest says that's a nil value 21:39 MTDiscord it would be very useful since right now my custom mapgen makes my entities lag 21:42 MTDiscord it requires minetest 5.9-dev 21:48 MTDiscord ah okay 21:48 MTDiscord is 5.9 when gltf support will arrive? 21:49 MTDiscord maybe