Minetest logo

IRC log for #minetest-dev, 2022-05-10

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

All times shown according to UTC.

Time Nick Message
04:00 MTDiscord joined #minetest-dev
05:02 CoolJar10 joined #minetest-dev
05:53 calcul0n joined #minetest-dev
09:00 proller joined #minetest-dev
09:21 nrz_ cool optimization sfan5, than should reduce lookup
09:25 proller joined #minetest-dev
09:42 HuguesRoss joined #minetest-dev
10:01 Alias joined #minetest-dev
13:16 Fixer joined #minetest-dev
15:38 rubenwardy #12314 and #12315
15:38 ShadowBot https://github.com/minetest/minetest/issues/12314 -- Add warning when writing to mod directory by rubenwardy
15:38 ShadowBot https://github.com/minetest/minetest/issues/12315 -- Add world-independent storage directory for mods by rubenwardy
15:40 rubenwardy may I could merge them, that probably makes sense
15:40 rubenwardy yeah
15:41 rubenwardy merge into one PR that is
15:43 MTDiscord <Warr1024> Interesting.  Seems to imply there is already some existing use-case in the wild for mods writing stuff cross-world.  Any known examples you can share?
15:43 rubenwardy Schematics
15:44 rubenwardy localscoreboards
15:44 rubenwardy the issue is #4821
15:44 ShadowBot https://github.com/minetest/minetest/issues/4821 -- Add world-independent storage directory for mods
15:44 MTDiscord <Warr1024> Oh, nat
15:44 MTDiscord <Warr1024> neat*
15:57 rubenwardy Ok I guess one reason to keep them separate is to stop people mixing them up
15:57 erle wdym mixing them up
15:58 erle i think its pretty obvious where stuff needs to go
16:00 rubenwardy Them being those PRs
16:09 sfan5 rubenwardy: are there any mods that store global data in the mod dir?
16:13 MTDiscord <Bastrabun> Wasn't it rhotator tries to create a new lua file each server start?
16:49 erle hmm, where do the skin mods store their skins?
16:54 MTDiscord <Jordach> Usually in their textures folder
16:58 erle Jordach, that means data in the mod dir, right?
16:58 MTDiscord <Jordach> It means the textures folder like any other mod
16:59 erle well, then any mod that downloads textures or lets users edit them is a candidate for “this needs to store global data, i.e. outside of world folders”
16:59 MTDiscord <Warr1024> "Traditional" skins mods don't really "store" their textures anywhere.  The server admin installs the textures and then restarts the server, and then the mod can load them.
17:00 MTDiscord <Warr1024> global dynamic textures could probably just as easily be using world storage though.  For more "universal" storage, cross-world might make some sense, but you're getting much closer to the "http API" point by then.
17:10 rubenwardy Yeah exactly
17:12 erle i guess i should not try to speak for people who make mods like skinsdb and epidermis. my own mods only ever write to world folder ;)
17:15 MTDiscord <Warr1024> I've only ever made the kind of skin mod where players submit skins to me, I approve them, and then my server update scripts pick them up and install them on the server.  I haven't messed with some of the HTTP API stuff yet.  It does look pretty fancy, but then there's also more moderation risk involved...
17:16 MTDiscord <luatic> Warr1024: in the strict modlib PNG & JSON parsers trust you must
17:17 MTDiscord <Warr1024> Yeah, parsers are not really going to help ensure that players don't wear skins that make their players look like giant dicks.  I think you might have the wrong threat model in mind.
17:18 MTDiscord <luatic> anyways, epidermis persistence only writes to the world folder - see https://github.com/appgurueu/epidermis/blob/b7cd469284c2f682c57eefa9185e1456a2deab02/persistence.lua#L4-L11
17:19 MTDiscord <Warr1024> Well now as of an upcoming release, assuming rubenwardy's change is in, you'll have the option for your mod to maintain a cache shared across worlds, which would probably be nice for stuff like reducing your skinsdb API traffic.  Assuming, of course, you can dynamic media out of it, I guess.
17:23 sfan5 could anyone quickly look at #12289
17:23 ShadowBot https://github.com/minetest/minetest/issues/12289 -- Support packing arbitrary graphs by TurkeyMcMac
18:18 hlqkj joined #minetest-dev
19:27 proller joined #minetest-dev
20:26 sfan5 thanks
20:26 sfan5 merging #12297, #12152, #12310, #12289, #12309 in 10m
20:26 ShadowBot https://github.com/minetest/minetest/issues/12297 -- Fix possible unreliable behavior due to uninitialized variables by Endi0n
20:26 ShadowBot https://github.com/minetest/minetest/issues/12152 -- Add doc to list breaking changes for the next major release by Zughy
20:26 ShadowBot https://github.com/minetest/minetest/issues/12310 -- Fix cooking and fuel crafts with aliases by TurkeyMcMac
20:26 ShadowBot https://github.com/minetest/minetest/issues/12289 -- Support packing arbitrary graphs by TurkeyMcMac
20:26 ShadowBot https://github.com/minetest/minetest/issues/12309 -- Use native packer to transfer globals into async env(s) by sfan5
20:40 sfan5 aaand we're past 10k commits
20:41 rubenwardy need a tweet
20:42 rubenwardy maybe a screenshot of 10,000 nyancats/mese or something
20:42 MTDiscord <MNH48> don't forget toot as well
20:42 MTDiscord <luatic> just screenshot the commit graph
20:42 MTDiscord <Warr1024> nyancats doesn't seem right, what with them having been 451'd out of the game...
20:43 rubenwardy zofc
20:44 MTDiscord <Warr1024> congrats though, that's almost 4x as many commits as NodeCore ?
20:45 MTDiscord <Warr1024> I always miss the fun milestones.  It seems like you should plan to make it a particularly salient commit or something instead of just something like "fix a typo" but it almost never works out that way.
20:46 rubenwardy 10k is https://github.com/minetest/minetest/commit/d17d7eba14f6a1328a3b58ee086df5ffa56304d6
20:46 rubenwardy !title
20:46 ShadowBot Fix cooking and fuel crafts with aliases · minetest/minetest@d17d7eb · GitHub
20:47 MTDiscord <Warr1024> seems like a good enough one
20:47 MTDiscord <Warr1024> I'm sure it will never be a thorn in our side like every time somebody says something like "...but my #1 reason is..." and then the bot drags up that ancient issue...
20:47 ShadowBot https://github.com/minetest/minetest/issues/1 -- GlowStone code by anonymousAwesome
20:54 Zughy[m] I was commit 9999, so close
20:59 Zughy[m] by the way, development is on steroids in these days
21:03 MTDiscord <Warr1024> I would rather see development on the engine, tbh
21:03 MTDiscord <Warr1024> Is the disagreement about numbering based on starting at 0 vs 1 or something?
21:21 MTDiscord <Bastrabun> More than on steroids. Unrelated: Thank you, core devs ?
22:34 panwolfram joined #minetest-dev
23:15 erle sfan5 sadly i can't show you a screenshot (another miscompile), but i believe that if you set chunksize 6, then the map ends at 31055 (visible map ends at 30959, add 96 nodes to it for the outermost invisible mapblock). this leads me to believe that 31007 is probably not a safe value to avoid bugs for chunksizes other than 5 – and my original assertion “do not use MAX_MAP_GENERATION_LIMIT for bounds checks” has been correct.
23:16 erle i'm sorry for not having realized this weeks or months ago, i just didn't think of it.
23:18 erle i am very sleepy, but if you want to keep these kinds of bounds checks, i guess a safe value for bounds checks always the smallest multiple of (16 × chunksize) that is bigger or equal to MAX_MAP_GENERATION_LIMIT
23:19 erle hmm, this is wrong
23:20 erle celeron55 do you know the position of the most central mapblock, like where it starts?
23:25 erle anyways, to my knowledge mineclone5 sets a chunksize of 4, so i predict fun times
23:42 MTDiscord <Jordach> the hint is 0,0,0
23:42 erle Jordach, explain?
23:43 MTDiscord <Jordach> i don't know which corner 0,0,0 explicitly belongs to
23:43 MTDiscord <Jordach> it can be x+ and z-
23:43 MTDiscord <Jordach> ie which of the 8 corners of a mapblock it could theoretically be
23:44 erle well, with chunksize=5 (i.e. blocksize 80) your map sharply ends at x=31007.
23:44 MTDiscord <Jordach> 31006
23:44 erle however often you remove 80 from that, you are not going to end up at 0
23:44 erle classic off-by-one haha
23:44 erle regardless, you are not going to get to 0 that way
23:44 MTDiscord <Jordach> 7 is there for safety reasons
23:44 MTDiscord <Jordach> because trimming the edge off a mapblock might go wonky#
23:45 erle if i am not mistaken you can't “trim the edge off” a mapblock
23:45 MTDiscord <Jordach> and a hard trim as in MMGL would do so
23:46 erle you either stop generating exactly at mapblock boundary OR your outermost mapblock lies partially within and partially outside of MAX_MAP_GENERATION_LIMIT (which is where all the funny bugs occur)
23:46 erle the only thing i know that stops exactly at the stated boundary is biomegen

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