Time Nick Message 07:42 MTDiscord And therefore that patch was forgotten to time and it never worked on Guix ever again 08:33 Krock [: #15189 08:33 ShadowBot https://github.com/minetest/minetest/issues/15189 -- Use std::filesystem::remove_all in RecursiveDelete by SmallJoker 08:34 MTDiscord A Halloween miracle! 08:34 MTDiscord Thanks, Kro ck 11:59 sfan5 [: see comments in the PR 12:49 MTDiscord rip 15:54 MTDiscord Since I've been working on MT perf issues for the past couple of weeks, I no longer feel that #15000 has substance to it that's not already explained better by #15125 or by exogenous factors like VPS performance. In laboratory settings I can't get 5.9 to perform worse in general than 5.8, it seems like the blame 5.9 got was due to a confound. 15:54 ShadowBot https://github.com/minetest/minetest/issues/15000 -- Major Server Performance Regression 15:54 ShadowBot https://github.com/minetest/minetest/issues/15125 -- Busy Disks Cause Lag 15:54 MTDiscord If I no longer support 15000 myself, is it okay if I close the issue myself, or is it better to comment and let the core team decide? 15:57 sfan5 fine to close yourself 16:08 MTDiscord 👍 17:05 [ Using rm is also problematic if it's used on non-absolute pathnames, since filenames beginning with "-" may be interpreted as options, although that's easily solved by prepending "./" if the path doesn't begin with "/" and I don't think minetest actually uses it on non-absolute pathnames anyway 17:07 MTDiscord dirs named ~ because some software doesn't interpret that as $HOME (while others do) is the worst 17:07 pgimeno can paths contain [] ? 17:08 pgimeno I'm appalled by the lack of escaping in the current implementation 17:09 pgimeno oh wait, it's not a shell command 17:09 pgimeno forget what I said 17:31 Krock will merge web#309 in 20 minutes. objections? 17:31 ShadowBot https://github.com/minetest/minetest.github.io/issues/309 -- Update downloads for 5.9.1 by SmallJoker 17:51 Krock merging