Time Nick Message 00:32 erle oh lol, stripping color codes 00:32 erle like seriously 00:32 erle they are stealing the rainbow from us 00:32 erle coredevs hate FUN 00:32 user2_ yea 00:32 cheapie Hey, at least this time they're leaving a setting for it instead of going full GNOME "feature? better remove it" 00:33 cheapie (yes, I'm still grumpy about digging particles) 00:33 erle “If you told most server owners right now that this is possible they would consider it cheating.” – i know quite a few server owners and i have yet to see anyone say that writing colored text is cheating 00:33 erle cheapie at least that's a thing you can change by having your own client 00:34 erle maybe i turn digging particles off because they are visual clutter, but you know, giving users options is apparently wrong these days 00:34 erle the server owner should decide! 00:34 erle if the web went that way we wouldn't have ad blockers 00:35 cheapie erle: It's worse than that, for digging particles the server owner (or mod devs) can't turn them off either 00:35 cheapie (but yes, I do patch them out of my client) 00:35 erle cheapie that is unfortunate 00:35 user2_ we should send digging CRACKS to all players, but that makes it feel to much like MC imo 00:36 erle cheapie i think that is basically the main difference between the direction the vanilla minetest/luanti is going and where almost all of the forks go. one is like “less features, less performance on hardware the devs don't have, more control to server owners”, everything else opposes at least one of these aspects 00:37 cheapie erle: FWIW I did at least have some luck in -dev recently convincing a core dev or two that no, you cannot just set the default view range to 3000 and expect most people to have a good experience on all servers 00:38 erle LMAO 00:38 erle cheapie i think you got taken for a ride, no one is that stupid 00:38 user2_ there's a fork that is literally the exact same, but has the engine's name changed in README.md. that's it, no executable renaming, no rebranding, no nothing except that ONE rename. 00:38 user2_ 3000 viewrange?... if you have a supercomputer, yeah sure 00:38 erle not even tthen 00:39 user2_ i would set it to 100 and leave it for the player to adjust 00:39 cheapie erle: https://irc.luanti.org/luanti-dev/2025-09-20#i_6285281 00:39 cheapie 3000 was exaggerating a bit but 1000 was mentioned 00:41 cheapie If you scroll down a bit, I managed to get lhofhansl to join VE-C and see what the performance is like near the spawn there 00:41 user2_ i never get more than 60FPS, i think i have a hardware cap 00:41 cheapie sfan5 has tested there too (quite a lot actually) and has a general understanding of what the problem likely is 00:43 user2_ i will say, 5.1x.x clients get more FPS on my RPi5 than 5.x.x clients 00:43 cheapie Yeah, I've heard that from other players too, 5.10 and up (or thereabouts) being somewhat less slow 00:44 cheapie sfan5: This reminds me - would having a local copy of that world be helpful for your testing? The downloads aren't publicly available but I can probably make the case for sending you one if it would help 00:44 erle you know the funny thing here is that things only get tested when people object 00:44 erle not like, by default 00:44 cheapie Same goes for the other VE servers, except nodecore 00:44 user2_ i was stuck on the 5.6.1 release from APT on debian 12 for a while... i can say it got ~30-45 FPS and 5.10+ get 45-60FPS 00:45 user2_ after APT i used the flatpak releases, then switched to compiling which tends to get slightly higher FPS 00:46 user2_ (for some reason) 00:46 erle user2_ question: does enabling/disabling shaders on 5.10 make any difference for you on the RPI? i know low-spec machines where shaderless rendering is faster and machines where the shader rendering is faster. 00:46 user2_ i can test it rn, does it need to be 5.10 in specific? 00:47 erle since 5.11 has removed that option (leading to 30% to 60% fps drops on a lot of older or weaker GPUs), it needs to be 5.10 or lower. 00:47 user2_ allright, i have every 5.x version precompiled and ready to run so it's not an issue 00:47 erle neat 00:48 user2_ unmodded 5.10 getting 60FPS with 100 viewrange 00:49 user2_ shaders off 00:49 erle increase view range to 150? 00:49 erle also which game? 00:49 user2_ lightly modded MTG, server is 37.27.252.202:30000 00:50 user2_ 150 getting 50FPS 00:50 user2_ 49 to be precise 00:50 erle okay, so then with shaders on. is it more or less? 00:50 erle shaders on only, no effects. 00:51 erle so it looks the same 00:51 user2_ which setting in specific- i dont mess with shaders too often 00:52 erle change the line "enable_shaders = false" to "enable_shaders = true" in minetest.conf 00:54 erle also: if you knew that shaders were off, how did you not know how to turn them on or off? like how did you check even? 00:54 user2_ shaders off seems to get ~57 FPS, on gets ~49 00:54 user2_ FPS 00:54 erle see, that's what i am talking about 00:54 erle but it looks the same, doesn't it? 00:54 user2_ i assumed shaders are off by default, i redid the tests to be sure 00:55 user2_ yeah, why would you add shaders that dont do anything visible... 00:55 erle now want to try with mineclonia on oysterity anarchy? it has a very … colorful. spawn. 00:55 user2_ i've been there.... 00:55 erle i mean for performance testing 00:55 user2_ very terrible no matter what settings i use 00:55 user2_ i'l do it anyway tho 00:56 erle if you do it, try to join without shaders and lower viewing range until it becomes playable subjectively. then report fps, rejoin with shaders, report fps again. 00:56 erle oh and viewing range at which it is considered playable 00:56 user2_ all right, loading now 00:57 erle user2_ fun fact: there is a hole in the nether roof near spawn. i did not make it. but what i did make was the end portal on the nether roof, using a bug that allowed me to acquire end portal nodes. ;) 00:57 erle the rails are most likely made by coras railbot CSM 00:58 user2_ 43 FPS with shaders off @150 viewrange 00:59 erle and with shaders on? 00:59 erle also i hope you are above ground and not looking at a wall lol 00:59 user2_ i am above ground... 01:00 erle although i must admit that looking at a wall is a great way to figure out if something is still rendered even if you can't see it 01:00 erle :P 01:00 user2_ 41 FPS with them on 01:00 user2_ not a huge amount 01:00 erle yeah, interesting how small it is on the RPI 01:01 user2_ ima try another area 01:01 erle on all intel machines i have tested except one and on my underpowered arm laptop the difference is much more pronounced (usually it is the difference between playable and not playable) 01:01 user2_ 34 FPS with them on here 01:01 erle and off? 01:02 erle ofc you can always reduce the view range to make stuff playable, but then you walk around with view range 50 :( 01:03 erle user2_ some time back oysterity used to have a lag machine at spawn. someone had made a menger sponge out of crying obsidian which generated particles that lagged players that were near it so much the game became unplayable. 01:04 user2_ 32ish FPS on, 36 off 01:04 erle now crying obby got fixed to no longer cause as much lag 01:04 erle user2_ guess RPI users are lucky! 01:04 erle user2_ have fun exploring it. if you find any automated farms, look for signs like “ælla was here” :P 01:05 user2_ RPis are underrated and treated as underpowered machines from 2005 01:05 user2_ they are really powerful >:-( 01:05 user2_ erle: i'm not going to play there much, just was testing smth 01:05 user2_ (aka the shader settings) 01:05 erle user2_ well graphics performance depends on a lot of things. 01:06 erle if you want to do one more test, turn on mipmaps and tell me if that gives you some fps 01:06 erle should be 1 or 3 01:06 user2_ on 5.10 as well? 01:06 erle yeah 01:06 erle ofc it will only do that if you enter an area where mipmaps are rendered 01:07 user2_ on the same server or on the MTG server i was on earlier? 01:07 erle nah, on oysterity anarchy 01:08 erle since the scaling algorithm is naive/bad you can see where the mipmaps start easily if you look at nodes in the distance. you can test how bad exactly using my checkerboard test nodes. 01:09 erle i believe hybrid_dog (?) once offered an improvement years ago (it's not a secret how to scale images in a way that you don't get such a bad result), but that was rejected IIRC 01:09 erle so, does it have an effect on the RPI or not? 01:09 user2_ 37 FPS in a different spot with shaders on, mipmaps on 01:10 erle just make sure it is a spot where you can actually see the mipmaps being used 01:10 erle when i tested it, i got no difference at all when they weren't 01:11 user2_ ~27-30 FPS just walking around with them off 01:13 erle shaders or mipmaps off? or both? 01:13 user2_ ~30-33 FPS with them on, walking around 01:13 user2_ (shaders on) 01:14 erle i see! 01:14 user2_ so it is causing a performance increase, just a slight one 01:14 user2_ speaking of shaders.... my friend is working on a proper SSR shader 01:15 erle SSR? 01:15 erle user2_ well obviously it is only a slight performance increase. the issue about such settings is that usually you have like 5 or 10 in a game and every one gives you a few frames. 01:15 user2_ ScreenSpace Reflections (i have no idea what it means, just means that water and stuff will reflect nodes/entities/players) 01:18 erle user2_ it means that reflections will only be rendered from what the user can see anyway. 01:19 user2_ i got an early access screenshot- but i can't share it until the final product is done 01:21 erle user2_ imagine you have a frame in which some water is reflecting other objects. you first paint everything else except the water. then you fill in the place where the water is with the reflected images of whatever it should reflect. 01:21 erle you can of course only reflect things that are in the frame 01:22 user2_ yeah, apparently he had a lot of trouble with the last part... 01:22 mrcheese oh, i happen to be the person working on this shader :P 01:22 user2_ oh hi 01:23 mrcheese user2_: yeah the water normal was in the wrong coordinate space and i did not notice that for a while -_- 01:23 erle user2_ so the main drawback is that you can not reflect things the camera does not see. 01:23 user2_ does that really matter? 01:23 erle user2_ think of it as a post-processing effect (because that is what it is) 01:23 user2_ if you can't see it, it doesn't really matter 01:24 erle user2_ depends. imagine a desk and below it is a puddle of water. you do not see the underside of the desk. 01:24 erle user2_ imagine looking down at water. you do not see the clouds. 01:25 user2_ that could technically be a cheat tbh 01:25 mrcheese yeah, another limitation is that if you try to reflect something behind a wall of some sort, you either dont get a relfection or get a bugged reflection of the edge of that wall and you wont see what is behind the wall 01:25 erle user2_ or simply imagine rendering a bathroom mirror. 01:26 erle user2_ player character looks into the mirror. so what happens now? 01:26 user2_ they don't see the stuff behind them reflected in it- 01:27 erle preciscely. so, imagine your game has bathrooms. i give you your question back: does it really matter? 01:28 erle obviously often game developers go to lengths to hide these drawbacks 01:28 user2_ no, because mirrors haven't been implemented in MT yet :P 01:28 erle so you have games with lots of reflective surfaces, but no mirrors 01:28 erle user2_ ever played half life 2? 01:28 user2_ no, i literally just started the original a week ago- 01:29 user2_ and i have a free key for it if you want :PPP 01:29 erle nah 01:29 erle user2_ pay attention to the reflections. especially what happens when you stand in water and look down. 01:30 user2_ anyway, i understand 01:30 erle good night 01:30 user2_ gn 01:30 mrcheese gn 11:24 MinetestBot 02[git] 04SmallJoker -> 03luanti-org/luanti: IrrlichtMt: Fix Win32 device compile error (non-SDL) 134f77a91 https://github.com/luanti-org/luanti/commit/4f77a91e04951cb3f1bbfc7133bcdc46d05d05fb (152025-10-04T10:57:14Z) 19:39 shaft What about this PR to worldedit? https://github.com/Uberi/Minetest-WorldEdit/pull/255 It's been half a year now and I got no feedback at all and nothing happened. 19:52 shaft sfan5 has been active since two weeks and there is literally no one else with access to the repo who is active. I thought maybe he could take another look at it. 19:52 MinetestBot 02[git] 04109247019824@users.noreply.hosted.weblate.org -> 03luanti-org/luanti: Translated using Weblate (Bulgarian) 138b3e80f https://github.com/luanti-org/luanti/commit/8b3e80f534026c89e47c654bda55b5fa8b75c6c9 (152025-10-04T19:35:52Z) 19:52 MinetestBot 02[git] 04Wuzzy2 -> 03luanti-org/luanti: Translated using Weblate (German) 1323e208c https://github.com/luanti-org/luanti/commit/23e208c7aab15a03ab630bab59c476a503f7a16e (152025-10-04T19:35:52Z) 19:52 MinetestBot 02[git] 04LudusExe -> 03luanti-org/luanti: Translated using Weblate (Italian) 13b6fe5fe https://github.com/luanti-org/luanti/commit/b6fe5fee4315ea3ebeee9c2be6f98c996b20248a (152025-10-04T19:35:52Z) 19:52 MinetestBot 02[git] 04newrizen -> 03luanti-org/luanti: Translated using Weblate (Portuguese (Brazil)) 13bca99fc https://github.com/luanti-org/luanti/commit/bca99fc7234b11777c61335d6f22385d775e8833 (152025-10-04T19:35:52Z) 19:52 MinetestBot 02[git] (7 newer commits not shown) 19:57 MinetestBot 02[git] 04maxchen32@users.noreply.hosted.weblate.org -> 03luanti-org/luanti: Translated using Weblate (Chinese (Simplified Han script)) 133a0152c https://github.com/luanti-org/luanti/commit/3a0152c491f4dbc56444f40b8f39d784f922e51e (152025-10-04T19:55:05Z) 20:05 sfan5 shaft: you will simply have to wait 20:05 shaft You know I can't :( 23:43 user2_ get your commits in, less than 24 hours till 5.14.0 releases 23:47 MTDiscord chucks feature freeze at user2 23:48 user2_ bugfix commits*