Time  Nick       Message
05:42 jas_       #game1894
05:43 jas_       oh. lol
06:20 jas_       game#1895
06:20 ShadowBot  https://github.com/minetest/minetest_game/issues/1895 -- Door sound volume up close is too loud compared.
14:37 paramat    can anyone review game#1886 ? seems good
14:37 ShadowBot  https://github.com/minetest/minetest_game/issues/1886 -- Fix locked chest protection against explosions (fixes #1885) by danielmeek32
15:53 dharkael   Hello! I'd like to talk about the pathfinding for Minetest before I start implementing what I think is the good approach
15:55 dharkael   please, read this: https://pastebin.com/aBNyq3Ss
15:58 sfan5      entity ids are not exposed to lua
15:58 sfan5      also i don't see why modders would need to manually free some cache
15:59 dharkael   they shouldn't need it, but sometimes entities can change properties and that changes how the pathfinding should be calculated
15:59 rubenwardy it's a better API to pass in an entity, anyway
16:00 dharkael   but maybe they could overwrite the latest cache
16:00 dharkael   with the register function
16:01 dharkael   is the free_path_cache function needed for some special cases? I don't know, that why I came here to talk about the design
16:01 dharkael   I'm not an expert minetest modder
16:04 rubenwardy when I started reading that, I thought you were talking about the path finding algorithm the generates a graph for the terrain which can be used by multiple entities for long distance fast path finding. Is this just caching A* for a single entity at a low range?
16:04 dharkael   I thought about the case of pets, they seek the owner and if they don't die and if they use cache the cached data would increase mor and more
16:05 dharkael   we should have, or a way to free cache under centain condiciones, or an automatich mechanism to control the level of cacged data
16:05 dharkael   cached data*
16:06 dharkael   shared cache is hard because entities can be configured with very different properties
16:08 dharkael   actually I thought about implementing the Moving Target D* Lite algorithm as sugested by raymoo here https://github.com/minetest/minetest/issues/5479#issuecomment-329682709
16:09 dharkael   well he said G-FRA* but it is suboptimal for moving target problems in dynamic environments
16:18 dharkael   I think the best approach is to have a configurable load facor for automatic cleanup of cache if it becomes quite big. That would make the free_path_cache function unneeded
16:21 dharkael   of course we could use a pahtfinding without cache but it can be from 6 to 20 times slower depending on the case. I could explore variations of Jump point search but they are inconvient when we have to consider size of the mob and many movement conditions
17:39 sofar      speaking of ABM's. If we search a mapblock for ABM stuff to do, and we find {array} nodes in there. Do we at least cache this information so that we're not searching the entire mapblock again every ABM tick if the mapblock didn't change?
17:53 paramat    good question
17:57 Krock      Merging game#1884 and game#1886 in 10 mins
17:57 ShadowBot  https://github.com/minetest/minetest_game/issues/1884 -- chest: fix open chest after player leave by tenplus1
17:57 ShadowBot  https://github.com/minetest/minetest_game/issues/1886 -- Fix locked chest protection against explosions (fixes #1885) by danielmeek32
18:00 paramat    thanks
18:01 paramat    game#1813 would be nice to have reviewed soon, it's delaying other work now
18:01 ShadowBot  https://github.com/minetest/minetest_game/issues/1813 -- Ores (non-mgv6 mapgens): New distributions by paramat
18:03 paramat    maybe game#1880 too?
18:03 ShadowBot  https://github.com/minetest/minetest_game/issues/1880 -- Fix some small sfinv issues by rubenwardy
18:03 Krock      ^ a balance PR. You have the most knowledge about the ore types and distribution, so I think that's good if you compared the different mapgen
18:04 paramat    ok thanks, i spent hours thinking it over
18:55 Krock      paramat, I believe we should change the local code to playSoundAt as well
18:56 Krock      for the consistency
19:05 paramat    hm
19:05 paramat    but that's unnecessary extra calculation
19:05 paramat    when we could just add a factor as in my PR
19:06 Krock      no, that makes it more realistic, as you don't make footstep noise with your ears
19:06 paramat    lol
19:06 Krock      you can't tell me I'm wrong with that, eh!
19:06 * paramat  walks on his ears
19:06 * Krock    runs away
19:07 paramat    directional sound only bothers with yaw, so the direction will not seem to come from below, it will make no difference it seems, except be more intensive
19:08 paramat    anyway, my PR is fine after all that study
19:09 Krock      mhm okay.
19:10 * Krock    goes to bed on his ears
19:10 paramat    lol
19:31 paramat    ok i'm happy with #6440 now, just seems to need a volume tuning factor added
19:31 ShadowBot  https://github.com/minetest/minetest/issues/6440 -- CAO footstep sounds: Reduce gain to balance volume by paramat
19:31 paramat    which is what it does
21:29 nerzhul    merging #6430
21:29 ShadowBot  https://github.com/minetest/minetest/issues/6430 -- Fix compile error in openbsd by jcalve
22:16 Megaf      Interesting, KSP depends on Mono, so it ships Mono with itself, both 32 and 64 bit versions
22:16 Megaf      we could include Irrlicht too, couldn't it?
22:16 Megaf      Im mean, bundle
22:18 sfan5      why would we do that