Time Nick Message 12:50 MTDiscord not sure if relevant for alot of people but OpenAL updated its codebase to C++20 since version 1.25.0 which broke VS 2019 + vcpkg compiling because VS 2019 doesn´t fully support C++20 module scanning. possible fixes are either updating to VS 2022 or using OpenAL 1.24.3 14:07 [MatrxMT] is it normal that node formspecs don't send any `on_player_receive_fields` input? 14:14 rubenwardy it'll call the node def callback, not the global callback 14:19 [MatrxMT] I don't think it's stated anywhere in the docs 14:24 rubenwardy you're right 14:24 rubenwardy also I think node formspecs are quite lame 14:26 [MatrxMT] I'm working on a trivial docs PR. Also, I agree 14:39 [MatrxMT] #16881 14:39 ShadowBot https://github.com/luanti-org/luanti/issues/16881 -- Docs: specify `on_player_receive_fields` doesn't work for node formspecs by Zughy 14:42 [MatrxMT] so there is no way to know what a player is doing inside a node formspec, unless I'm checking from within the node? 14:46 rubenwardy ContentDB translation description bug fixed 14:46 rubenwardy also the description in MTG's game.conf does not match that on CDB 14:47 rubenwardy should be edited to be much shorter so it can actually display in the LT client 14:47 rubenwardy see also: https://content.luanti.org/packages/Luanti/minetest_game/translation/ 14:53 sfan5 thx. I edited the short description now 14:53 sfan5 wait that was a stupid idea, it's now cut off 15:49 MTDiscord Do we have some nice way how to detect in Lua server protocol version? 15:51 MTDiscord core.get_version() 15:53 MTDiscord thanks 15:57 MTDiscord you normally should prefer to use feature flags, what are you working on? 16:00 MTDiscord just wants to send notification to clients which are using old client, that they can expect crazy thinks happening. 16:02 MTDiscord more generous that most mods. either kick/block or do nothing 16:03 MTDiscord You are right, but in reality, it looks like Multicrafts users like to play on Luanti servers from iOS. 16:03 MTDiscord It does not looks like admin wants to simple block them. 16:04 MTDiscord i think we already starting to see server admins start to not care about multicraft users because they are so (5.4) outdated 16:07 MTDiscord I just was asked if it is posible to block all old clients with exception for Multicrafts users. 16:07 MTDiscord it is, but its rather stupid to do so 16:08 MTDiscord I agree. Just blocks all old clients or if not.. it does not matter if it is player on Multicraft or old Luanti client. 16:12 MTDiscord by the way, in 5.15.0 is included support for more then 32765 nodes? 16:13 MTDiscord yes 16:13 MTDiscord Did we also change proto_min in get_version able? because I assume, it is totaly incompatible with old clients. 16:13 MTDiscord sounds like you want core.get_player_information(player_name).protocol_version 16:14 MTDiscord Yea, I want to compare core.get_player_information(player_name).protocol_version and core.get_version(proto_max) 16:14 MTDiscord looking at the commit will tell you that (which you can get to from the end of the pr when merged) 16:14 MTDiscord https://cdn.discordapp.com/attachments/747163566800633906/1465017069106889019/image.png?ex=697792f7&is=69764177&hm=5a24ab1eb2ce78f7dd8e80267eaf97b1925e5c8c112288dbf2b793c3ce41ad5f& 16:19 MTDiscord well, it looks like old clients can be able to manage it. 18:02 MTDiscord Well. I did more testing. It looks like if you died, and in that situation, you get disconnected due to connection lost, or you cast end game without fulfilling the respawn formspec (like CTRL-C or cross in Windows corner) And if you after join a game that shows the joined players some formspec, it will overwrite the respawn formspec and get the player stuck in death situation. 18:02 MTDiscord So the question is... is it game bug or engine bug? 18:04 sfan5 IMO games should ensure not to show a formspec to a dead player 18:04 rubenwardy what if the send it just before the player dies? 18:05 Krock have the death screen overwrite any formspec that is attempted to show 18:05 Krock or mamilpate core.show_formspec to add such check 18:05 Krock *manipulate 18:05 sfan5 isn't the death screen server-side now? 18:06 Krock it is. and if the quit field isn't received, the player is not respawned. 18:06 MTDiscord just after connection I see respawn formspec for a half of seconds. After it get replaced by game formspec showing rules or something like that. and if I exited that, no respawn formspec. No able to interact doe to death.... 18:06 Krock builtin/game/death_screen.lua 18:09 sfan5 i think we could at least detect if any non-death formspec was closed and then resend the death formspec? 18:10 MTDiscord So, should I open bug issue for this situation? 18:10 user333_ that reminds me of a bug i encountered on the CTF server, where another formspec would close the death formspec, preventing you from respawing 18:14 Krock "if any non-death formspec was closed" - this also includes core.close_formspec to cover perhaps all edge-cases. 18:44 sfan5 thought on the discussion in #luanti: it would be good if we could publish rough, user-accessible stats on what luanti versions people are generally running 18:44 sfan5 e.g. like https://apilevels.com/ 18:45 MTDiscord iirc ruben wanted to opt in stats or something 18:45 MTDiscord dunno if applicable to this 18:45 user333_ would be cool to see the most popular versions/platforms 18:47 user333_ which reminds me of this: [2025-11-06 20:14] PSA: for anyone interested here are new statistics on what versions/platforms/etc. everyone is running Luanti on from the server list POV -->> https://kitsunemimi.pw/tmp/serverlist_stats_2025-11-06.txt 18:48 sfan5 i mean we have stats via CDB and the serverlist. would just have to make sure the data represents what we want it to 18:49 sfan5 (e.g. in case of the server list stats: make sure those weird android apps don't distort it) 18:56 user333_ "weird android apps" some player came on our server and claimed to be on "mine-build-craft", clearly there's a lot of those 18:56 user333_ their version string was a 5.13.0-dev commit 19:03 user333_ also note that 5.9.0-dev-943e0e9f9 is the web client 19:05 user333_ here's some stats from a server my friend runs: https://termbin.com/jlw4 19:17 Krock I wonder whether something would break if I inserted a few newline characters or "`rm -f`" 19:17 Krock (as the reported client version) 19:21 user333_ hmmmmm 19:21 user333_ try setting it to "1.20.4_Minecraft" and see what server owners do 21:03 lhofhansl I just installed a Luanti binary (msvc-x64) produced by a github action on my son's Windows machine. I noticed that maploading and other tasks are much (50% or more) slower on this build as with the official build? Is that expected? I looked at the Windows yaml file used to produce the build, and it seemed to be doing the right thing (a normal release build, etc). 21:06 sfan5 🤷 21:09 lhofhansl So it isn't unexpected? 21:24 [MatrxMT] It's probably a good time to think about whether we want to have a barebone implementation of v-rob's new UI API. Because author doesn't answer and I had to close the PR. #14263 21:24 ShadowBot https://github.com/luanti-org/luanti/issues/14263 -- [NO SQUASH] Create new UI API (Formspec/HUD replacement) by v-rob 21:26 sfan5 incremental changes are always better 21:27 rubenwardy yeah. Even with a new API, you need to get the smallest change in first so it's in the codebase and can be worked on in smaller commits 21:47 sfan5 ~tell lhofhansl I don't think so. But I think nobody has looked at the optimization flags in detail. also there's the whole bunch of libraries from vcpkg which we have no insight into 21:47 ShadowBot sfan5: OK. 23:12 [MatrxMT] and is their current work a good starting point? 23:25 MTDiscord Maybe, they were in the weeds and may not of tidied up the core