Time Nick Message 14:32 Zughy[m] as non core dev, I think we should move with 5.7, if we keep adding bugfixes we won't ever release. There are 18 PRs with one approval and two with 2, stuck because of feature freeze. Only the particle spawner PR is missing 14:33 nrz Zughy, you don't think quality is important ? 14:33 nrz the 18 PR with approval can be merged in batch after the release 🙂 14:35 Zughy[m] I think we should release often: it's been 8 months since 5.6.0 14:35 sfan5 this blocks work somewhat though with different PRs depending on eachother 14:35 proller 19 with https://github.com/minetest/minetest/pull/12142 ! 14:35 sfan5 anyway it's not that we want to add more and more fixes to 5.7, we're waiting for #13200 vs #13374 14:35 ShadowBot https://github.com/minetest/minetest/issues/13200 -- Fix missing particlespawner values by velartrill 14:35 ShadowBot https://github.com/minetest/minetest/issues/13374 -- Particle spawner: Move definitions and fix missing params by SmallJoker 14:35 sfan5 I guess I could go and review the latter one right now 14:36 Zughy[m] oh my god proller, stop with that PR 14:37 proller i will stop after merge 14:51 sfan5 do we have a changelog written yet? 14:51 nrz proller, i note you said the PR have bugs 14:57 proller nrz, it have no bugs with disabled pos32 feature 14:57 proller and it shouldbe merged in disabled state 14:59 proller with enabled i know only one minor problem with sneaking or autojump code 15:04 sfan5 refactoring the types sounds useful in principle but I don't remember whether we decided last time that there's no interest in ever enlarging the worlds 15:09 proller i can prepare pr with type changes only (if here is chance to merge) 15:10 nrz sfan5: i don't remind why we decided it, and if the PR permits to think about it why not after all. The enforced typing is cool, the main issue i see is, as it's c++, typedef are not real types, it means if you start to reuse v3f in some places you may break your code 15:17 proller typedefs can be changed to class ... : public realtype { } with disabled copying/creation from real type 15:20 proller also v3f can be converted to/from v3d without problems in s16 space 15:41 nrz yeah having strong types is better i think, for preventing coding errors, our codebase is huge 17:31 Desour why did "s16" link to in the logs?! 17:32 Desour https://irc.minetest.net/minetest-dev/2023-04-04 17:46 Pexin would the breaking world limits thinger make it more complicated / practically impossible to later implement workable world-wrapping? 17:47 Pexin I had wanted to finally get off my butt and adapt spheretest into something server-configurable. but I probably won't. 17:57 appguru Desour: Logs are powered by https://github.com/moritz/ilbot/ - an ancient thingymabob written in Perl. Of course a Perl dev would consider it a convenient feature to auto-link some Perl stuff (which I suppose perlcabal used to be). See "sub synopsis_links { [...] if ($s =~ m/^S(\d\d)$/i){ return [qq{}, $s, q{}]; }" 17:58 appguru I don't want to know which other, potentially exploitable features this thing has ;) 17:58 Desour 👀 18:05 sfan5 I suppose I could edit that out 18:07 MTDiscord seem like it really used to be perl stuff, there's wayback machine copy from 2014 where that link was still perl stuff - http://web.archive.org/web/20140904181802/https://perlcabal.org/syn/S16.html 18:09 MTDiscord 2015 copy had it redirects to http://design.perl6.org/S16.html in which the link currently redirects again to https://design.raku.org/S16.html 18:09 sfan5 should be gone now 18:11 Desour s16 18:11 Desour yep 23:34 Guest5497 celeron55 (erle here) so i figured out that if you limit downflow according to viscosity you can do great things! https://mister-muffin.de/p/u3-K.png 23:35 Guest5497 i constructed this wall using some lava and water. i.e. first make a lavafall, then put water above it so it flows down. 23:35 Guest5497 it is called ”lavacasting” 23:36 Guest5497 so far it works surprisingly well to construct giant walls, but lavacasting needs to be an option of course 23:36 Guest5497 otherwise everyone with a lava bucket and a water bucket becomes a freelance builder 23:36 Guest5497 so i think it should not be default 23:43 Guest546 https://github.com/facebook/zstd/releases/tag/v1.5.5 23:44 Guest546 > While the probability might be very small, corruption issues are nonetheless very serious, so an update to this version is highly recommended, especially if you employ high compression modes (levels 16+). 23:47 Guest546 https://github.com/facebook/zstd/pull/3517 23:48 Guest546 > TLDR: This is a very rare bug affecting the block splitter. Only levels using the optimal parser (levels 13 and above depending on the source size), or users who explicitly enable the block splitter, are affected. 23:48 Guest546 minetest should maybe limit zstd compression to level 12 as long as the zstd library version is <1.5.5?