Time Nick Message 08:09 nerzhul sfan5: sounds that clang is not happy with one of your optimization 08:09 nerzhul mapblock_mesh.cpp:422:64: attention: « void* memcpy(void*, const void*, size_t) » écriture dans un objet de type « v3s16 » {aka « class irr::core::vector3d »} sans affectation par copie triviale; use copy-assignment or copy-initialization instead [-Wclass-memaccess] 09:53 sfan5 nerzhul: I don't have that with clang-10 here but I noticed that gcc complains about that 09:53 sfan5 either way the optimization is definitely safe but the compiler doesn't understand that 11:18 rubenwardy A non trivial constructor implies it shouldn't be safe 12:16 sfan5 the "non-trivial" constructor in this case is vector(const vector &other) : X(other.X), Y(other.Y), Z(other.Z) {} 12:17 sfan5 which the compiler also turns into three moves even if it should be easily vectorizable 12:20 rubenwardy sounds like it should = default? 12:20 sfan5 probably 12:20 sfan5 irrlicht was written before c++11 and still doesn't use it to this day (I think?) 12:27 rubenwardy ah right 14:46 nerzhul no it doesn't use it, it's exact 16:33 sfan5 reminder that #9627 is ready and works 16:33 ShadowBot https://github.com/minetest/minetest/issues/9627 -- Make Mapgen (mostly) work multithreaded again by sfan5 16:40 nerzhul i will take a look today 17:13 sfan5 merging game#2662 in 5m 17:13 ShadowBot https://github.com/minetest/minetest_game/issues/2662 -- doors: Remove unnecessary nodebox and empty texture by sfan5 18:04 rubenwardy Krock: appguru is doing that PR 18:04 rubenwardy tbh, they should have just done it 18:04 rubenwardy referring to old_speed/velocity 18:05 rubenwardy tbh, you'll have done it in less time than to respond to this 18:06 Krock was about to fix it together with fog #9605 18:06 ShadowBot https://github.com/minetest/minetest/issues/9605 -- get_sky_color returns unknown fields 18:06 Krock well yes. that commit is already done locally 18:06 rubenwardy maybe just do it then 18:24 Krock btw #9812 18:24 ShadowBot https://github.com/minetest/minetest/issues/9812 -- [NO SQUASH] Script API: Rename fields for constency by SmallJoker 18:54 rubenwardy I've updated 18:54 rubenwardy !dev TODO 18:54 ShadowBot TODO - Minetest Developer Wiki -- http://dev.minetest.net/TODO 18:56 Krock > perhaps coinciding with the release of 0.5.0, is better than breaking compatibility all the time 18:56 rubenwardy Loool 23:44 slemonide Minetest partitions the map into chunks and stores serialized chunk data into the database. I am curious if this is the best practice in general? Would it be better to split chunk data across several records and then SELECT the ones we want? 23:48 rubenwardy a chunk is loaded in its entirety 23:48 rubenwardy a chunk is the discrete level at which the map is loaded 23:49 rubenwardy it doesn't make much sense to split stuff up more, and it makes it more complicated