Time Nick Message 06:07 user333_ only 3 PRs to go, can we get to 100? 06:12 mrcheese :o 17:29 MinetestBot 02[git] 04sfan5 -> 03luanti-org/luanti: Add new setting to enable/disable dark mode in main menu (#16787) 133877b93 https://github.com/luanti-org/luanti/commit/3877b93f961b3e8ca7f9080d49127f428c0473a9 (152026-01-16T17:27:44Z) 17:31 crazylad :D 17:31 mrcheese YAY DARK MODE GOT MERGED 17:54 MTDiscord Yeah and "before GTA 6" 😅 17:54 MTDiscord finally 18:02 MTDiscord eventually someone's going to mess up and mention that after GTA6's release 18:53 user333_ now we need raytracing/coloured lighting before half-life 3... 20:14 cheapie This is how to use a bike rack, right? https://cheapiesystems.com/media/images/screenshot_20260116_141327.png 20:21 user333_ those are some odd bikes 20:24 mrcheese damn those things look like a lot of work to pedal 20:25 user333_ though, assuming those are bikes, you are using it correctly 20:25 Krock cheapie: strange bike 20:26 Krock I think there's actually a bike mod somewhere 20:26 Krock !mod bike 20:26 MinetestBot Krock: Mod: Bike [bike] by Hume2 - https://content.luanti.org/packages/Hume2/bike/ 20:26 user333_ though, assuming those are bikes, you are using it correctly 20:27 user333_ why did that get sent twice 20:27 mrcheese what did? 20:28 user333_ my message "though, assuming those are bikes, you are using it correctly" 20:28 user333_ thats what you get for using an IRC client from like 2015 ig 20:30 user333_ oh it's even older ...Copyright 2010 by..." 20:36 Krock user333_: at least it's not archived like HexChat (not sure whether that makes it any better) 20:39 user333_ it might be archived for all i know 20:39 Krock https://github.com/leedo/alice 20:39 user333_ yay it's not 20:39 Krock (source according to Debian) 20:40 user333_ just hasn't had an update in 11 years 20:40 user333_ no open PRs or issues 20:40 user333_ well at least it st 20:40 user333_ still works 20:41 Krock obviously it has no issues 20:41 Krock just look at the tabs on the project page 20:44 user333_ no up to date forks either 22:01 cheapie Totally normal checkout lane: https://cheapiesystems.com/media/2026-01-16%2015-59-38.webm 22:01 cheapie Inspired by https://www.reddit.com/r/schkreckl/comments/1qcqkpb/inspo_von_uattemptweird4209_werken/ 22:09 cheapie Also this is downwind of one of the nuclear power plants in Bonnthal... it's probably fine? https://cheapiesystems.com/media/images/screenshot_20260116_160744.png 22:18 mrcheese how to crash a minetest server: `local t=0;core.register_globalstep(function()t=t+1 core.settings:set(tostring(t)..math.random(0,999999999),string.rep("data",1024):rep(1024))end)` 22:18 cheapie How to crash one, but easier: core.register_globalstep(error) 22:19 mrcheese lol 22:19 mrcheese mine kills a server the slow and painful way..... aka filling RAM with 'datadatadatadatadatadatadatadatadatadatadatadatadatadata' etc 22:23 mrcheese how to crash a whole client (assuming singleplayer): `core=nil` 22:31 cheapie I wonder now what the shortest thing is that you can type into something like worldedit's //lua that executes successfully but then crashes the server 22:33 mrcheese i think its just `core=nil`... idk how this could be made shorter xd 22:34 cheapie core={} is already shorter :P 22:35 cheapie I don't know if either one counts though, I don't think they manage to return the success message before the server crashes 22:36 mrcheese true- 22:36 mrcheese `core.after(1,function()core={}end)` 22:36 mrcheese lol 22:36 cheapie If you want to use core.after(), then core.after(1,error) is shorter 22:37 mrcheese oh yea 22:37 cheapie (or 0, whatever) 22:37 mrcheese 12 characters shorter than core.register_globalstep(error) 22:39 MTDiscord io=1 or os=1 feel like good candidates, assuming these end up being used somewhere 22:40 mrcheese xD 22:41 mrcheese doesnt work on MTG with no mods other than worldedit :P 22:43 user333_ output of //l core=nil: 22:43 user333_ 2026-01-16 18:42:01: ACTION[Server]: singleplayer executed core=nil 22:43 user333_ 2026-01-16 18:42:01: ERROR[Main]: ServerError: AsyncErr: Lua: Double fault error from mod 'worldedit_shortcommands' in callback on_chat_message(): error in error handling 22:43 user333_ 2026-01-16 18:42:02: ACTION[ServerStop]: Server: Shutting down 22:43 user333_ terminate called without an active exception 22:43 user333_ Aborted (core dumped) 22:43 user333_ only 5 lines, pushing the limit on using a pastebin... 22:44 cheapie This reminds me of finding what's probably the shortest C program to compile :P 22:44 cheapie 'main;' 22:44 user333_ i mean this also works "" 22:44 mrcheese xD 22:44 cheapie Does it count as a C program with no content? 22:45 cheapie I think 'main;' invokes undefined behavior, so beware of nasal demons, etc., but in practice it generally segfaults 22:45 mrcheese lmao 22:45 user333_ oh nvm it wont compile 22:46 user333_ simply 'main;' segfaults when you try to run it 22:46 MTDiscord well, math=1 should be a pretty safe bet, but math is 2 characters more than io or os 22:46 mrcheese could do core=1 :P 22:46 mrcheese that will be a easy way to torch a server 22:47 user333_ MTG still works afaik with io and os = 1 22:47 mrcheese yea 22:47 mrcheese i tested it 22:47 MTDiscord that's possible, need to get lucky for a mod to use those 22:48 user333_ gl with core though 22:48 cheapie WorldEdit needs minetest.chat_send_player to be intact to return the success message, so just nuking core doesn't really work 22:48 MTDiscord you're just nuking the core alias though 22:48 mrcheese true 22:48 mrcheese eh- 22:48 cheapie hmm 22:48 user333_ hmm... would it still work if i set minetest to {} 22:49 user333_ nope 22:49 MTDiscord now that depends on whether they do "local minetest = minetest" 22:49 user333_ using the core namespace instead of minetest will fix that D: 22:50 mrcheese oh wow so using `minetest=0` crashes the server but doesnt fully nuke the client (`core=0` on singleplayer crashes the entire client) 22:50 cheapie luatic: I especially love seeing that pattern applied to code that runs exactly once 22:50 MTDiscord heh 22:51 cheapie Or to shave a few µs off a chat command while ignoring an ABM in the same mod that takes like 200ms 22:51 user333_ this wont instantly break the server.... but will as soon as players discover it: core.register_on_chat_message(function(message) assert(loadstring(message))() end) 22:51 mrcheese xD 22:52 mrcheese wait it does nothing xd 22:52 user333_ what did i screw up this time D: 22:53 cheapie advtrains has a lot of this sort of thing, there are a few instances of it doing something in some weird non-obvious (and usually buggy) way, with a comment showing that it's marginally faster in some micro-benchmark that ran the code 1 million times 22:53 cheapie Never mind that the code path this is in isn't particularly slow and isn't called very often anyway 22:54 user333_ oh wait, try core.register_on_chat_message(function(_, message) assert(loadstring(message))() end) 22:54 user333_ yea that works 22:54 mrcheese ah 22:55 cheapie Also I think last time I ran into one of those, the engine provided a built-in function that did the same thing and was faster anyway 22:56 user333_ isnt it fast enough already 23:01 cheapie Yes, it was plenty fast even in the obvious way 23:02 cheapie IIRC the reason I was there in the first place last time around was because the "optimized" code broke 23:02 cheapie I swear that mod breaks on every update 23:02 user333_ we need a C++ modding API now :> 23:02 mrcheese either im cursed or im stupid, but core.register_on_chat_message(function(_,message)core.chat_send_all(message)end) doesnt do anything xd 23:03 user333_ works for me 23:09 mrcheese i must be cursed then .-. 23:09 user333_ indeed 23:11 mrcheese ok it works on my other client.... weird 23:13 mrcheese lol saying anything that isnt valid lua crashes the server 23:14 mrcheese core.register_on_chat_message(function(_, message) assert(loadstring(message))() end) -- (user333_'s previous code)