Luanti logo

IRC log for #minetest, 2022-09-25

| Channels | #minetest index | Today | | Google Search | Plaintext

All times shown according to UTC.

Enable nick filtering
Time Nick Message
26 more elements. Show/hide.
10:12 MinetestBot TurkeyMcMac: Sep-21 12:00 UTC <appguru> Thank you very much for the error handler PR.
3 more elements. Show/hide.
11:46 MinetestBot [git] aerkiaga -> minetest/minetest: Briefly explain how facedir rotations work (#11934) 5ced5c9 https://github.com/minetest/minetest/commit/5ced5c9b27a3c534bcbd7d095affe2f65d07e9b1 (2022-09-25T11:45:02Z)
11:46 MinetestBot [git] coldtobi -> minetest/minetest: Fix two spelling mistakes 7069d99 https://github.com/minetest/minetest/commit/7069d99aa6d38afb4f0b8862d9c0e28328ecfd41 (2022-09-25T11:45:27Z)
11:46 MinetestBot [git] appgurueu -> minetest/minetest: Fix double escape in update checker dialog 8dec3a5 https://github.com/minetest/minetest/commit/8dec3a5ecbc6f74e2ca5d59f7bd905b8b136a591 (2022-09-25T11:45:37Z)
3 more elements. Show/hide.
12:46 sfan5 test message please ignore
20 more elements. Show/hide.
17:02 Duvalon Hello. What is the vertical_blend attribute for biome generation supposed to do? lua_api says "Vertical distance in nodes above 'y_max' over which the biome will blend with the biome above."
17:02 Duvalon But in my tests I just get a straight horizontal line separation no matter what I put in there.
17:06 MTDiscord <MisterE> Why are the minetest forums 10x slower than their normal slowness?
17:10 MTDiscord <Jonathon> https://discord.com/channels/369122544273588224/749727888659447960/1023264017814405220
17:11 MTDiscord <Jonathon> https://irc.minetest.net/minetest/2022-09-24#i_6018302 for the irc folk
17:12 MTDiscord <MisterE> Thx
13 more elements. Show/hide.
20:58 hmmmm so when are you gona REWRITE IT IN RUST?
21:00 MTDiscord <savilli> Never, doesn't worth it
21:02 MTDiscord <GreenXenith> too late, already working on it
21:02 hmmmm <Duvalon> Hello. What is the vertical_blend attribute for biome generation supposed to do? lua_api says "Vertical distance in nodes above 'y_max' over which the biome will blend with the biome above."
21:02 hmmmm wat
21:02 hmmmm did somebody change biome generation?
21:03 hmmmm or did i add this feature?  i can't remember
21:04 hmmmm sounds to me like y_max is the threshold at which biomes will begin linearly interpolating with the biome at the same x,z calculated with the new y
21:04 calcul0n_ joined #minetest
21:04 hmmmm vertical_blend sounds to me like the rate at which such interpolation occurs
21:05 hmmmm i'm confused because this sounds like a feature i was working on at some point but i don't remember ever making a PR for it
21:05 hmmmm ugh I've been wanting to do a clean break from MT to try out everything i've learned about procedural generation
21:07 hmmmm that approach to having two distinct biomes at different heights that blend into each other sounds logically appealing but might be too overcomplicated
21:07 hmmmm but how else are you gonna get snow caps on mountains
21:08 hmmmm ultimately, biome noise should be 3d or else that'd get boring fast
21:08 Desour 3d biomes sounds more useful in underground
21:09 hmmmm because most air is unpopulated, sure
21:09 Desour (mountains in minetest are usually too small to have different climate or vegetation on the top)
21:10 hmmmm you can "optimize" noise for above-ground by only generating it on demand (when there is a non-air node that exists in said chunk)
21:10 hmmmm are chunks still 5x5x5 by default?
21:11 hmmmm well that depends on the mapgen and the specific mapgen parameters.  you can't make assumptions about what kind of terrain you'll have for things independent of the mapgen itself
21:11 hmmmm biomes are supposed to technically be an attribute imbued on top of the map itself, like decorations and other misc. "objects"
21:12 hmmmm or at least that was the original concept.  i haven't looked at the current code but you guys probably butchered it lol
21:13 hmmmm so i guess "mapgen" is a bit of a misnomer for what i was aiming for... more like "landgen", and then the "biomegen" and "decoration gen" and "dungeongen" come together to make said "mapgen"
21:20 Duvalon interesting...
21:21 Dacmot joined #minetest
21:22 Duvalon hmmmm: that's just it, I don't think there's vertical interpolation at all
21:23 hmmmm ugh i can't believe you're getting me to look at minetest code
21:24 Duvalon hmmmm: you don't actually have to :D
21:26 hmmmm ahh no it's not a parameter for a lerp you're right, it seems to be the width of the region where biome dithering begins
21:27 hmmmm same as with the x and z coords
21:28 hmmmm seems like nerzuhl was touching my biome code but paramat fixed it
21:28 Duvalon :-|
21:28 hmmmm does nerz still add memory safety bugs in his code?  lol
21:28 Duvalon I wouldn't know...
21:34 Verticen joined #minetest
21:39 hmmmm has anybody audited minetest code at all?
21:39 proller joined #minetest
21:40 hmmmm even just the free coverity scan?
21:44 schwarzwald[m] hmmmm: Yes, we have that.
21:44 schwarzwald[m] You need access from the core devs in order to see the report, I believe.
21:44 hmmmm I am a core dev
21:44 schwarzwald[m] Then you can probably already see it?
21:44 hmmmm i haven't even looked, just chatting
21:45 hmmmm if i get too involved with mt again it's going to suck up all my spare time
21:45 schwarzwald[m] I remember you I think. You're that core dev who is very friendly and isn't very active anymore IIRC.
21:46 schwarzwald[m] I was trying to fix 2 bugs from the coverity report every day as a new habit, but I was only able to keep it up for a few days before I didn't have time and it was tiring.
21:46 hmmmm there was a time when i wasn't friendly at all lol
21:47 schwarzwald[m] Sorry, I should've said seems friendly. This is the internet after all.
21:47 hmmmm no my personality has changed for the better over time
21:48 schwarzwald[m] I see you also got wise enough not to suck up your spare time on Minetest, haha.
21:48 hmmmm yeah, my fear was that the coverity report would spot too many issues and it'd be a flood
21:48 hmmmm i feel like maybe new feature contributions should stop at some point and there should be focus on optimizing and fixing
21:48 hmmmm we sort of did this before the 0.5.0 release but seems like it's gone off the wall again
21:48 schwarzwald[m] It looked to me like a flood of minor details that nobody except someone who wanted to exploit them could exploit.
21:49 hmmmm yeah that would be my primary concern
21:49 hmmmm the last thing i'd want to see is people joining a friendly MT server for a friendly session of playing their video games and then they get malware
21:49 schwarzwald[m] Someday I want to sit down and try to dedicate myself more to learning how to seriously contribute to Minetest maintenance, but this is my first year of university and I probably won't even have time to make more CMake PRs until that is over...
21:50 hmmmm i spent most of my time in university working on MT, and then i got into the real world where i coded for a living and that made me burn out completely
21:50 hmmmm this is probably your only real chance tbh
21:51 schwarzwald[m] Well, that's a point... maybe my chance is over. I didn't even have time to write my Codingame bot this week. I can't imagine how a university student could have time for open source contribution lol.
21:51 hmmmm and it's a great community to learn how open source collaboration works without too great of a barrier to entry as with larger, mature projects
21:52 hmmmm because I never did any homework
21:52 schwarzwald[m] I want to succeed, gotta do my homework. xD
21:53 hmmmm but coding mt was more beneficial to my career than spending hours on math proofs that ended up being useless in the real world
21:53 hmmmm so just saying
21:53 schwarzwald[m] I always felt that there was a pretty big entry barrier to the project because of the very specific domain and not having anyone who would teach new people.
21:53 hmmmm yeah the teaching part is a bit rough.  celeron and kahrl helped me a lot, and i helped paramat as well as a few others
21:54 hmmmm time is a very valuable resource
21:54 schwarzwald[m] It's hard to be more ovelwhelmed than when you're thrown from toy projects into something real and you suddenly realize you know nothing, but you have to figure it out for yourself haha.
21:54 hmmmm my advice?  don't try to be a generalist
21:54 MTDiscord <Jonathon> > it's going to suck up all my spare time relatable
21:54 hmmmm if you like graphics, spend all your time on the shaders and don't even bother trying to fix or review someone else's unrelated PR
21:55 hmmmm that way you can actually get "deep" enough to make real contributions without being completely overwhelmed
21:55 schwarzwald[m] See my problem is that I like fixing bugs.
21:55 hmmmm fix bugs only in a specific subsystem
21:55 hmmmm MT as a whole is way too overwhelming unless you know it well from working on it several years
21:56 hmmmm and MT is a "small" open source project, lol
21:56 schwarzwald[m] Hmm, that is an interesting point.
21:57 schwarzwald[m] Maybe I should try to capitalize on my work with Minetest's build system so far and not worry about not getting into source code.
21:57 hmmmm you're working on the build system???
21:57 schwarzwald[m] I have a complete rewrite of the build script in progress.
21:57 hmmmm brave dude, i think only a couple others have dared to touch cmake
21:57 hmmmm it's the thing that i avoided at all costs unless it was completely necessary because it's such a deep rabbit hole
21:57 hmmmm you have my support for sure
21:58 schwarzwald[m] TBF I am scared of breaking some random thing on some 10 year old Windows system because there's a workaround in the source code that requires setting a build flag I don't know about.
21:59 Desour it would be great if minetest compiled faster
21:59 Desour it's such a pain
21:59 hmmmm i am happy that minetest compiles, period
21:59 schwarzwald[m] You're starting to sound like erlehmann, Desour.
21:59 hmmmm omg i remember erlehmann
21:59 Desour schwarzwald[m]: huh, in what way?
21:59 hmmmm and did vanessa quit?
22:00 schwarzwald[m] Not in tone, just in desire for a faster compilation.
22:00 Desour ah
22:00 jwmhjwmh joined #minetest
22:00 schwarzwald[m] I kind of think that a faster compilation is unlikely from a realistic perspective, but I hoped that by revamping the build I would at least make it more maintainable.
22:01 schwarzwald[m] I have the client building on Linux, I believe, and I downloaded MSVC at uni for the main purpose of running my build on Windows to get that working next.
22:01 schwarzwald[m] If things go like usual, MacOSX won't get tested until Jordach sees something got merged and blasts my head off.
22:01 hmmmm it's easy to break compatibility for the more obscure platforms, i kept having to fix freebsd compilation every time i'd come back from a break and there were lots of commits added
22:02 schwarzwald[m] My OpenBSD laptop decided to quit, while I was trying to get it set up to test erlehmann's own build.
22:02 hmmmm :/
22:03 schwarzwald[m] Someone told me that the important thing for working on the build is to have access to lots of different machines.
22:03 hmmmm at least there are people still trying to maintain the bsd build, at one point some tried dropping support
22:03 hmmmm yes that is true, one of the things that *I* was supposed to do but dropped the ball on was maintaining a server of VMs for each build target with jenkins-ci installed
22:04 hmmmm the mac OSX VM was the biggest pain in the ass
22:04 Desour schwarzwald[m]: can you enable precompiled headers in minetest? https://cmake.org/cmake/help/latest/command/target_precompile_headers.html
22:04 schwarzwald[m] Actually, I have Sunday afternoon to play around... maybe I'll try to compile Minetest on Windows for the first time.
22:05 schwarzwald[m] Desour, it's certainly possible.
22:07 schwarzwald[m] Do we have a recommended developer workflow for setting up a git repo on Windows for use with Visual Studio?
22:09 schwarzwald[m] I think I'll just install git from the official website and try to use that.
22:18 Dacmot joined #minetest
22:20 rubenwardy hmmmm: VanessaE retired from MT and handed all her mods and servers to other people
22:23 wallabra joined #minetest
22:24 schwarzwald[m] I installed vcpkg in a subdirectory in my user. I'm guessing I already messed up.
22:24 schwarzwald[m] Since it clearly isn't on my PATH.
22:24 rubenwardy vcpkg shouldn't be in your path
22:24 schwarzwald[m] Ah, ok.
22:24 rubenwardy you tend to install it per project, I put it inside the project dir
22:25 rubenwardy I suppose you could share the vcpkg, I prefer not to to keep things separate
22:25 schwarzwald[m] I like to save space and bandwidth.
22:26 schwarzwald[m] It's only been a half hour, I haven't even configured the IrrlichtMt build, and I'm already feeling the pain of developing on Windows.
22:31 schwarzwald[m] > Wait until cmake have generated the cash file
22:31 schwarzwald[m] Typo in the README
22:31 schwarzwald[m] * > Wait until cmake have generated the cash file
22:31 schwarzwald[m] Typo in the README
22:31 schwarzwald[m] I would PR to fix this right now except I haven't even set up SSH keys let alone my GPG signing on Windows.
22:32 schwarzwald[m] > If there are any errors, solve them and hit Configure
22:32 schwarzwald[m] "solve them"
22:32 schwarzwald[m] Pro advice.
22:34 Desour that hint solves world hunger
2 more elements. Show/hide.
22:59 * schwarzwald[m] sent a code block: https://libera.ems.host/_matrix/media/r0/download/libera.chat/c83ad92074cda174d1d127e6d68c70226720a0fc
22:59 schwarzwald[m] I just did the `vcpkg.exe install ...` command, and it completed successfully and listed a bunch of targets it set up.
2 more elements. Show/hide.

| Channels | #minetest index | Today | | Google Search | Plaintext