Time  Nick       Message
14:39 rubenwardy I suggest that the aim of 5.4.0 be reducing code rot - both in PRs and in current code
14:40 sfan5      how does that look in practice
14:46 rubenwardy I'm not sure how to acceleration PR review other than brute effort
14:49 sfan5      me neither
14:49 sfan5      PR review is definitely an issue but I don't see a way to fix it other than throwing more people at the problem
14:49 rubenwardy I probably only have time for a few hours a week on Minetest
15:22 rubenwardy web#215
15:22 ShadowBot  https://github.com/minetest/minetest.github.io/issues/215 -- Remove customize, link directly to ContentDB by rubenwardy
15:32 rubenwardy I think the meetings are helpful in encouraging PR review
15:47 rubenwardy there's a healthy balance of chat to action. A planned meeting discussing PRs/issues? Healthy.   Bikeshedding over whatever? Not health
16:03 sfan5      there seem to be a large amount of PRs that just require someone to say "yeah this looks sane"
16:03 sfan5      (if you put the required trust into the submitter that is)
16:09 sfan5      <Krock> https://github.com/minetest/minetest/pull/7924/files#diff-b5b39c3a29359759ac6f0293dc37a453R277
16:09 sfan5      <Krock> this should be joystick.getWasKeyDown(KeyType::MOUSE_L)
16:09 sfan5      ^ that's not fixed yet is it?
16:23 lhofhansl  Hi all... There's a (perhaps) interesting discussion in #10422.
16:23 ShadowBot  https://github.com/minetest/minetest/issues/10422 -- Mods for in-world-games are not displayed correctly in the Configuration tab
16:24 lhofhansl  Basically MT allows to place a game/mods folder into any world... i.e. .../worlds/<world_id>/game/mods. An in that case all other mods and configs, etc, are ignored. This is a nice feature, IMHO, as it allows shipping complete self-contained worlds. On the other hand it's confusing and not used much AFAIK.
16:26 lhofhansl  Sorry, meant: #10439
16:26 ShadowBot  https://github.com/minetest/minetest/issues/10439 -- Cannot select a world with game mods in the UI
16:29 rubenwardy updated #10173
16:29 ShadowBot  https://github.com/minetest/minetest/issues/10173 -- Deprecate get_player_velocity and add_player_velocity by rubenwardy
16:30 rubenwardy lhofhansl: I had no idea that this feature existed, but cool, I guess
16:30 rubenwardy It would be good to have a single source of truth for mod and game locations, in src/content somewhere
16:32 lhofhansl  I didn't either. I was looking for a way to do that, and found this in the documentation :)
16:33 rubenwardy the usual convention is to include worlds in the game, using schematics
16:35 lhofhansl  Yep. This is so much more powerful, though. You can just include the map and all the mods. Together with the read-only maps stuff, this is really flexible.
16:59 rubenwardy #10444
16:59 ShadowBot  https://github.com/minetest/minetest/issues/10444 -- Prevent games from setting secure settings by rubenwardy
17:12 Krock      whyIsItCamelCase
17:23 Krock      good job cleaning up old PRs btw
17:24 Krock      will merge #10441 and #10407 in 15 minutes
17:24 ShadowBot  https://github.com/minetest/minetest/issues/10441 -- Update wording of game descriptions in installer by LoneWolfHT
17:24 ShadowBot  https://github.com/minetest/minetest/issues/10407 -- Improve MSVC cmake and update vcpkg instruction by Adriankhl
17:39 Krock      merging
17:42 Sokomine   just out of curiosity: does avx/avx2 play a role regarding mt?
17:42 sfan5      not directly
17:43 sfan5      in theory the compiler can optimize C++ code we have and emit avx(2) instructions
17:44 Krock      instruction stuff is up to the compiler. applications shouldn't need to deal with that
17:45 Krock      though the compiler options sometimes limit the instruction set to produce an executable that runs on almost all modern systems
17:45 Sokomine   seems reasonable. was just wondering. there doesn't seem to be much in mt that could profit from it i think
17:48 Krock      well, you could compare a binary with a special one that passes the "-mavx2" (or so) option to GCC
17:48 Krock      list here: https://gcc.gnu.org/onlinedocs/gcc/x86-Options.html
17:49 Krock      but I doubt it would change much
17:49 Krock      or "-march=tigerlake" if you happen to have a tigerlake CPU
17:55 Sokomine   no, just (still) wondering which cpu to buy. avx seems to produce more heat, so it's good if mt doesn't really need it
17:57 Sokomine   with my build no longer beeing a debug build and mostly just mtg with my mg_villages mod, it's nice how fast worlds start up now
19:37 MTDiscord  <J​onathon> @GreenXenith ☝️
19:37 MTDiscord  <G​reenXenith> already back up
20:35 pgimeno    <Krock> instruction stuff is up to the compiler. applications shouldn't need to deal with that  <--  I disagree. If you want to take advantage of vectorization, you need to write vectorizable code. You need an understanding on what the compiler can optimize and what can't. Compilers don't do magic; e.g. they can't generally transform a slow O(n³) algorithm into a fast O(log log n) one. If you want to take advantage of the machine you're
20:35 pgimeno    running on, often you need to reflect that in the code.
20:41 pgimeno    example: https://github.com/vnmakarov/mum-hash/blob/master/mum-prng.h#L64
20:46 pgimeno    so a more correct answer here is: "if Minetest uses AVX[2], it's only by chance"
20:46 pgimeno    or maybe Irrlicht does, but I have my reservations :)
22:35 paramat    merging #10366
22:35 ShadowBot  https://github.com/minetest/minetest/issues/10366 -- Improve layout of main menu local tab by paramat
22:38 paramat    here's another simple PR #10247 . well tested, perhaps only needs a quick code review
22:38 ShadowBot  https://github.com/minetest/minetest/issues/10247 -- Give unnamed world names incremental numbers. Format 'world<number>' by paramat