Time  Nick       Message
06:57 celeron55  rubenwardy: i'll try to answer something
07:04 celeron55  https://forum.minetest.net/viewtopic.php?f=7&t=28024&p=409465#p409465
07:05 celeron55  i guess that should probably be pasted as a comment in mapsector.h or something
07:05 celeron55  as it is quite unintuitive
07:10 celeron55  edited it a bit
09:25 sfan5      neat
09:43 erle       celeron55 good idea with the comment
09:44 rubenwardy Thanks for the reply, much more detailed than expected!
10:10 celeron55  it does make me think it would be worth testing and benchmarking adding a second MapSector cache slot to Map, so that it would be efficient to alternate between two MapSectors
10:11 celeron55  the downside is of course that then, for every access, two cache slots have to be checked before the actual coordinate mapping is used, instead of the one that currently exists
10:12 celeron55  it seems to me it would be worth it
10:13 celeron55  or... could
10:13 celeron55  well, maybe not
10:13 celeron55  the beauty of the current system is that it's clearly the simplest system that performs well
10:14 celeron55  there might be a more complex system that performs a bit better, but not a simpler one
10:48 appguru    The Minetest docs currently don't match the implementation concerning named colors (rebeccapurple in particular); the link points to a dev version of the page (this was discussed a while ago). There are two fixes for this: Either updating the docs to note that only Level 3 is supported or updating the code to support level 4 by adding rebeccapurple. I've made PRs for both: #12203 and #12204
10:48 ShadowBot  https://github.com/minetest/minetest/issues/12203 -- Fix docs to reference CSS Color Module Level 3 by appgurueu
10:48 ShadowBot  https://github.com/minetest/minetest/issues/12204 -- Support CSS color module level 4 by appgurueu
17:15 erle       sfan5 i am trying to understand windows timing more in an effort to figure out why windows minetest dtime jitter is so high. do you have an idea why the delta of two adjacent QPC invocations sometimes could be off by a lot? i have a program that outputs numbers, and most of the time it outputs 1 or 2, but sometimes 384 or 8020 for the measurement. am i using QPC wrong? https://mister-muffin.de/p/CHpM.c
18:44 sfan5      normal multithreading fluctuations IMO
19:39 sfan5      merging #12203, #12205 in 5m
19:39 ShadowBot  https://github.com/minetest/minetest/issues/12203 -- Fix docs to reference CSS Color Module Level 3 by appgurueu
19:39 ShadowBot  https://github.com/minetest/minetest/issues/12205 -- Fix some debug info showing despite being disabled in the UI by appgurueu
19:40 erle       good idea to reference the correct docs instead of adding yet another named color