Time Nick Message 14:20 repetitivestrain incidentally has anyone observed that the forums have been offline awfully frequently as of late 14:20 MTDiscord the forums are always terrible. the level of terrible comes and goes in waves 14:22 sfan5 didn't we have someone monitor the forum uptime 14:22 [MatrxMT] cheapie: you can use a C like language with beduino 14:22 [MatrxMT] or do i need rto use tell 14:22 MTDiscord buckaroo monitors pretty much everything luanti related 14:23 MTDiscord https://monitoring.luanti.ch/ 14:24 [MatrxMT] 8.37 million clients 14:24 MTDiscord hm, seeing https://monitoring.luanti.ch/d/xUp66EYGz/minetest-forum-user-stats?orgId=1&from=now-1y&to=now&timezone=browser&refresh=10s, i think there used to be a uptime one. maybe removed it due to c55 complaining 14:24 sfan5 it's on https://monitoring.luanti.ch/d/xtkCtBkiz/minetest-websites-status? 14:25 MTDiscord ah, there it is 14:25 sfan5 however it only goes back like 5 days 14:25 repetitivestrain why does it state the forum to be up when it still yields a 504 bad gateway for me 14:25 repetitivestrain ah it must've recovered during the past 10 minutes 14:26 MTDiscord anyways, the solution to the forum is always one of two. convince c55 to actually do something, or get him to give it up. 14:28 MTDiscord at least docs have amazing uptime now :p 14:50 cheapie [09:22:29] <[MatrxMT]> cheapie: you can use a C like language with beduino 14:51 cheapie That's the language I was talking about not being very familiar with 15:45 [MatrxMT] I see 16:55 kyle56 is there a way to migrate the database format? like i DONT mean database type like sqlite to postgres. i mean changing from the pos,data columns to the new x,y,x,data columns without making a new world? 16:56 MTDiscord not sure if there is a better method, but the old trick was migrate it to another format and then migrate it back 16:56 MTDiscord also not sure if that trick works for the x/y/z format 16:57 kyle56 hmm i guess i should try that. but it kept telling me that leveldb was not a valid database type. i dont really wanna set up postgres. so i would like to maybe convert from sqlite to leveldb and back to sqlite 16:58 MTDiscord probably because you didnt compile leveldb in? 16:58 kyle56 hmm dunno. i just used the flatpak. does it not come with leveldb? 16:58 MTDiscord also leveldb is unreliable af, but as a temp holder should be fine 16:59 MTDiscord uh, --help or one of the command line options has the build config 16:59 kyle56 thats my rationale too heh. just need a easy temporary medium 17:00 sfan5 the trick works yes 17:00 sfan5 but you need a second format for it to work 17:00 MTDiscord --version 17:00 sfan5 --version doesn't show the database backends unfortunately 17:00 sfan5 neither does help 17:01 MTDiscord rip 17:01 MTDiscord probably should make issues for showing db backends in --version and a straight up --migrate 17:02 kyle56 i might have been typing it wrong. i tried these parameters: `--server --worldname example --gameid minetest_game --migrate level_db` but i did also try 'leveldb' 17:04 kyle56 i agree that would be nice. 17:04 MTDiscord https://github.com/luanti-org/luanti/issues/16439 17:07 MTDiscord https://github.com/luanti-org/luanti/issues/16440 17:52 sfan5 the flatpak build probably only supports dummy and sqlite3 17:52 sfan5 which means you're out of luck 18:50 kyle56 oh. thats unfortunate. but thanks for letting me know. 18:52 sfan5 out of curiosity: why do you want to update your world format? 18:57 kyle56 i wanna play more with sqlite queries. like using special queries to delete certain areas. and while i havent made the feature request yet, it gave me an idea for how core.delete_area could be optimized. 19:02 kyle56 i think someone on the lunati team had touted this as a benefit. basically running a command like this `DELETE FROM blocks where (x < %d or x > %d) or (y < %d or y > %d) or (z < %d or z > %d);` in this case, this is made to delete all blocks outside of a specified zone. but could be tweaked to delete all blocks INSIDE of a zone and i think core.delete_area could use it to work faster. 20:03 MinetestBot erle: Aug-22 00:02 UTC AFAICT, there is no magic "if you have this GPU, set this near plane". check logs 20:03 erle oki 20:31 erle luatic i don't understand anything from the logs. but “a proper solution is to let the game devs control the value” is a good joke! 21:07 erle luatic so what does the near plane setting depend on and why can't it be checked programattically? 21:30 luatic erle: it depends, among other things, on the game. you could try to have some "magic" that infers from scene geometry what a sensible value might be but generally it's best if game devs are able to control it. 21:30 luatic proper camera control includes control of the entire view frustum, including near and far planes. 21:31 luatic to give an example from something i'm a bit familiar with, the glTF "model" (more like "scene"?) file format lets you control these, see https://registry.khronos.org/glTF/specs/2.0/glTF-2.0.html#cameras. 21:32 luatic godot, like any other 3d game engine, unsurprisingly also lets you control the near plane: https://docs.godotengine.org/en/stable/classes/class_camera3d.html#class-camera3d-property-near 21:33 erle luatic i am pretty sure that “let someone who has even less diverse hardware than the coredevs” works even less well than letting the engine handle it through e.g. depth buffer bit depth or whatever in practice 21:33 erle also overdraw is not a problem, is it? 21:33 erle underdraw is xray 21:33 erle well, unless you overdraw and see 2 triangles blocking your entire view 21:34 sfan5 you seem to be expecting a game engine to serve the player instead of the gamedev? 21:35 erle oh indeed. stuff like “every dev has the *option* to implement a graphics settings menu” (several engines do this) results in games being less-than-friendly to players 21:36 erle i actually think this is a core part of why luanti has so many (mostly short-lived) forks. almost all of them fill the space “what if the engine lets the players adjust more”. the extreme ones are e.g. cheat clients and halons “i make a custom client so mineclonia works better in it” fork. 21:36 sfan5 your latter example is literally the opposite 21:37 sfan5 oh you said "almost all" 21:38 erle well, i am showing a spectrum. and it's pretty clear that luanti will neither be customizable enough to people who want accessability/scripting features (i.e. ultimate player-friendly) nor take patches to benefit only a single name (i.e. ultimate dev-friendly) 21:39 erle single game 21:39 erle typo 21:39 erle but yes, i do believe that a game engine (and all software) should have settings and that server-sent things must be “hints” like in the browser where you can override all CSS 21:39 erle and i am pretty convinced that the coredevs do believe that a lot of stuff should never be overridden or user-settable, correct? 21:40 erle in any case, the question of settings is usually an aesthetic one. unintentional xray is not aesthetics, but functionality. and if it can be configured automatically, it should, but i have no idea how. 21:41 erle it's too important to leave it to a game dev who probably has only 1 or 2 GPUs 21:42 erle if you leave functional settings to gamedevs, you end up with a per-game equivalent of the “no text rendering on my tablet” luanti bug 21:43 sfan5 compared to "real" game engines luanti is extremely limited, a part of that is a lot of things being left up to user settings 21:44 sfan5 and I think it's pretty clear what the consensus on direction is 21:44 erle well, say it? 21:44 sfan5 become more flexible like a real game engine? 21:45 erle that does not address anything regarding aesthetics or functional settings 21:46 sfan5 it will be necessary to some extent, exactly how much is debatable 21:47 erle btw, what was the rationale on specific values of near_plane? i mean you removed it and set a value, so where does that come from? experimentation? 21:48 sfan5 I wouldn't know 21:48 erle you forgot about it or you never knew? 21:48 sfan5 why do you think I know in the first place? 21:50 erle because you approved PR #8749 21:50 ShadowBot https://github.com/luanti-org/luanti/issues/8749 -- Ignore near_plane setting on non-Android platforms by ClobberXD 21:50 erle so i assumed you knew what you were doing there, but it's years ago, so 21:52 sfan5 if you read the PR you will see that someone did some experimentation on different values and based on that we arrived at a conclusion 21:53 sfan5 and that someone isn't me 21:56 erle you had some musings about lower end GPUs though and you did immediately suggest adjusting the near_plane, so i thought you know. but i get it, we both don't. if i find out what exactly determines how this setting should be, i'll tell you. 22:49 kyle56 I dont have the full context but on the subject of near planes, is that something that the server gonna be able to set? I hope so because I have a mod that can make players tiny. And when they are tiny, they can get up close and look through walls.