Time Nick Message 03:16 paradust sfan5: https://github.com/minetest/minetest/pull/12409 and https://github.com/minetest/irrlicht/pull/111 will fix Mac 03:17 paradust Homebrew is pulling from 5.5.1.tar.gz and 1.9.0mt4.tar.gz: https://github.com/Homebrew/homebrew-core/blob/master/Formula/minetest.rb#L7 03:17 paradust Not sure the right way to update homebrew. Could backport and add a special patch release for mac? 03:18 paradust Or could add these patches directly into the minetest.rb file 09:51 sfan5 paradust: great, thanks! 09:51 sfan5 while we can pull in the minetest fix for 5.5.2 (soon) there won't be a point release for irrlicht 09:52 sfan5 it also depends on the preferences of the homebrew maintainers but I think patching is the best option 11:03 appguru Zughy: Can you take care of https://github.com/minetest/minetest/issues/5618#issuecomment-1147325318? 11:07 appguru BTW #3733 might have been fixed by x2048's massive node depth sorting PR 11:07 ShadowBot https://github.com/minetest/minetest/issues/3733 -- Waving water not rendered correctly through glass underwater 11:07 appguru Fixer: Could you see whether you can still reproduce it? 11:12 Zughy[m] I trust your judgement appguru, can't really check in depth right now 11:22 sfan5 not sure which bug I'm supposed to be seeing here https://0x0.st/oMB8.png https://0x0.st/oMBK.png 11:23 sfan5 though that might be because of framed glass 11:26 Wuzzy I wrote a Dev wiki page for the Spawn algorithm. Please someone read over it if there are any mistakes: https://dev.minetest.net/Spawn_Algorithm 11:27 Wuzzy In other news, the file upload on DevWiki is currently broken :( 11:28 Wuzzy After analyzing this algorithm, i see several flaws: 11:28 Wuzzy 1) It happily may spawn players into Ignore nodes, which is a gamble. This happens if the area was not generated yet. 11:29 Wuzzy This works fine as long you add ANY decorations on the surface. if there are decorations, the player might spawn in a decoration becaus the spawn algorithm does not knowthem yet (since its ignore) 11:32 Wuzzy 2) (more esoteric) Due to the way positions are picked, if there are only a few valid positions between (-4001,-4001) and (4001,4001), the algorithm will likely miss them all and falls back to (0,0,0) 11:34 Wuzzy for example, you could have an empty world with only a straight line between X=-4000 to X=4000 and the algorithm will completely fail to see it because of bad luck 11:41 Zughy[m] why MT has to have all these information scattered around 3-4 places? 11:42 sfan5 well for one putting it in lua_api.txt would mean we have to support it staying the way it is 11:42 sfan5 perhaps doc/ would be a good place 11:42 sfan5 perhaps not because nobody ever looks there 11:47 Zughy[m] yeah, but there are two wikis, lua_api.txt, the "unofficial" modding book AND the (stale?) project to provide some code snippets/docs in the minetest repo 11:59 MTDiscord is the large amount of outdated junk copied from lua_api.txt into the dev wiki ever going to be removed 12:12 Fixer appguru: i doubt i will reproduce it, bug is 5 years old and is rare 12:28 Wuzzy Zughy[m]: The DevWiki is !!!NOT!!! the Lua API documentation and never was 12:29 Wuzzy all the pages replicating lua_api.txt have been marked for deletion by now 12:29 Wuzzy we now only need someone who actually deletes them 12:30 Wuzzy Sadly, the peopole who can delete pages are never available :( 12:30 Wuzzy btw, if someone wonders what needs to be done to clean up the messy DevWiki: https://dev.minetest.net/MinetestWiki:Lua_API_Wiki_Documentation_Cleanup 12:30 Wuzzy I have done my part of the homework 12:31 Wuzzy I have done everything that is currently within my power to clean up as much of the mess as I am allowed on DevWiki 12:31 Wuzzy meaning marked all offending pages for deletion 12:31 Wuzzy so the ball is now in the courtyard of the privileged DevWiki users. everything now depends on them 12:32 Wuzzy sfan5, was putting the spawn algorithm into the DevWiki OK in your opinion or would you rather have it in the repo? 12:33 sfan5 both ways are fine by me 12:33 Wuzzy At leats I mentioned the version number in the wiki article so it should be fine 12:42 Wuzzy Thoughts about #5863? 12:42 ShadowBot https://github.com/minetest/minetest/issues/5863 -- Chance that player spawns in decoration in ungenerated map 14:17 sfan5 theres so many easy bugs that someone with motivation could fix... 14:17 sfan5 (not me) 14:19 Zughy[m] we want you on the hard stuff :^] 14:23 Zughy[m] Can anyone confirm this? It takes 30 seconds and I can't abuse the "Bug" label :P #12403 14:23 ShadowBot https://github.com/minetest/minetest/issues/12403 -- `is_visible = false` doesn't work on players 14:27 MTDiscord after https://github.com/minetest/minetest/pull/12377 minetest.formspec_escape crashes when you pass a number value to it 14:28 MTDiscord it's quite ironic that appgurueu was always for full backward compatibility and broke it by himself 14:49 Zughy[m] anticheat is active by default, right? 15:00 sfan5 not in singleplayer 15:00 sfan5 @savilli file a bug or PR 15:02 sfan5 (so we don't forget) 15:08 MTDiscord I'd like some feedback on #12411 15:08 ShadowBot https://github.com/minetest/minetest/issues/12411 -- Proposal: Refactoring RenderingCore stack 15:09 MTDiscord #12412 don't forget, don't forgive 15:09 ShadowBot https://github.com/minetest/minetest/issues/12412 -- minetest.formspec_escape doesn't accept a number value anymore 16:28 MTDiscord savilli: I'm sorry, I had not expected that a parameter named text was supposed to support numbers. Made a PR to fix it. 18:42 appguru sfan5, would you like to discuss the LuaJIT workaround? 18:43 sfan5 do you want to gift me a better implementation? :D 18:43 appguru hehe 18:43 appguru fixing the current hack should be as simple as accounting for the referenced strings 18:43 appguru which... is actually nontrivial 18:44 sfan5 yes you run that part first, let them write into the env and all following deserializations use that env 18:45 appguru I'm trying to come up with a more general workaround 18:45 sfan5 the only general workaround is Lua interpreter implemented in Lua 18:45 sfan5 or you mean on the serialization side? 18:45 appguru wrapping stuff in functions as hecks proposed a while ago unfortunately doesn't work 18:45 sfan5 do we even know this bug still happens? 18:45 appguru nah, on the deserialization side 18:45 appguru sfan5: that's the ultima ratio 18:46 appguru but I think a tokenizer written in Lua will already get you pretty far 18:47 appguru in the end it would be a partial Lua interpreter reimplementation 18:47 sfan5 in the meantime there's this https://github.com/Uberi/Minetest-WorldEdit/commit/abc9efeeb8cccb3e23c055414941fed4a9871b9a 18:48 sfan5 (not a complete fix, no) 18:48 appguru where did yueliang go? 18:49 appguru sad what happened to luaforge 18:49 sfan5 and the real real fix is not using a damn programming language as a serialization format 18:49 sfan5 seriously terrible idea 18:49 appguru TBH that's right 18:50 appguru I have a binary serialization format that is as powerful :P except for serializing Lua funcs, duh 18:51 appguru I have even written a full-fledged reader deserializer for "Lua object notation" as I call it, but it's in JS: https://github.com/appgurueu/luon (and it doesn't support "backreferences") 20:06 appguru sfan5: anything left to do for #11427? 20:06 ShadowBot https://github.com/minetest/minetest/issues/11427 -- Redo serialize.lua by appgurueu 20:06 sfan5 no 20:07 appguru yay 20:46 sfan5 unofficial announcement sorta: I plan to work on allowing Lua to run in the EmergeThread (thereby not blocking the main thread) soon 20:55 MTDiscord nice 22:26 kilbith https://paste.ubuntu.com/p/5NpqVJCM9k/ 22:27 kilbith https://paste.ubuntu.com/p/sysfS8r9p5/ 22:27 kilbith couple of warnings 22:43 MTDiscord get 'em too