Time Nick Message 01:08 Mantar yeah, the link is viewable just fine 01:09 Mantar though it'll be gone later when you click it in the irc.minetest.net logs, as discord deletes that stuff after a while 01:11 MTDiscord Image links shouldnt break 🤔 01:12 MTDiscord Yeah, I can still access images from discord on the irc logs from 4 years ago 01:14 rubenwardy I think they're referring to https://www.bleepingcomputer.com/news/security/discord-will-switch-to-temporary-file-links-to-block-malware-delivery/ 01:14 Mantar :iunno alls I know is I've had discord image links die on me 01:14 rubenwardy maybe that's file only 01:14 Mantar ¯\_(ツ)_/¯ 03:47 MTDiscord Part of the glTF work being done is to turn on exceptions in Irrlicht, so hopefully we can do that. 08:29 sfan5 @paradust yes, you can turn them on 10:01 MTDiscord sfan5: thank you for being sane (re https://github.com/minetest/minetest/pull/14311#issuecomment-1914330758) 10:52 zarac o/ 10:54 zarac I'm trying to run mods 'animalia' together with '3d_armor'. Server fails to start due to animalia referencing global variable 'armor' (when 3d_armor mod exists) but gets null-pointer. I tried moving up load_mod_3d_armor before load_mod_animalia, but it didn't help. Any ideas? What determines module load order? 10:55 zarac I can see that 'armor' is indeed globally defined in '3d_armor'. 10:57 zarac Nice to see some recognizable names here.. thanks for great mods rubenwardy, fluxionary, Sokomine. :) 11:43 zarac ok. so load order is undefined, but dependency specification should be sufficient ( https://github.com/minetest/minetest/issues/8933 ). 11:43 zarac Seems 3d_armor is missing from `optional_depends` in `animalia/mod.conf` 11:44 zarac Is this ElCeejo here sometimes? (dev of animalia) 11:45 zarac (server started after adding 3d_armor to optional_depends) 11:57 zarac .. i wrote him on forums. (if anyone cares) : ) 11:58 MTDiscord Hello zarac you might be looking for the regular #minetest-irc channel 11:58 zarac is that for mod development? 11:58 zarac i assumed this was 11:59 MTDiscord Yes, this is for engine development 11:59 zarac alrighty. Thanks and excuse me ; ) 11:59 MTDiscord You're very welcome 11:59 zarac i see the title, now =P 14:05 MTDiscord FOSDEM is this weekend, right? 14:11 MTDiscord (I know it's February 3 and 4, I just need someone to confirm that it's actually this weekend. Don't ask me why...) (Also, I should have asked in #minetest instead.) 14:14 Soni how do you detect client type? 14:18 [MTMatrix] grorp: yes 14:24 Soni we would like minetest to have a feature where it can enforce client type/client features like custom rendering (portals, bloom, other fancy effects), fedilinks support, etc which are otherwise not supported by the base minetest client 14:38 MTDiscord Soni, then you fork and distribute your own client, with those shaders enabled by default (bloom) and with custom features. Then hopefully you PR the portals thing back upstream 14:40 Soni we want the client to send feature flags on connect and the server to enforce them 14:40 Soni (i.e. disconnect clients without them) 14:42 MTDiscord That's definately a feature you will have to develop and contribute 14:43 MTDiscord I mean, you can check the protocol version. 14:44 MTDiscord Also note that enabling bloom isn't the most obvious process, so you will be disconnecting most players unless you provide a button to automatically set the suggested settings 14:46 MTDiscord You can check protocol version, but all future versions are reserved by the official client for future official versions, so if you use a new one for some other meaning, you'll no longer be compatible. 14:49 sfan5 @grorp yes 15:33 Soni exactly why feature flags and namespacing are important 15:34 Soni does protocol version use semver? 15:39 Soni you can't implement astral sorcery or portal gun mod in minetest because it requires client changes to do properly, so the obvious solution is to just support third-party clients 15:47 MTDiscord Protocol version is an int16. I suppose it probably follows semver as much as you can with a data type that only supports a "minor" number ("major" would be separate protocols entirely). 16:03 Soni okay, so no way to shove feature flags into it 16:12 MTDiscord why does astral sorcery require client changes? 16:15 MTDiscord Feature flags do not exist within the protocol, they're a thing used to communicate between the engine and mods. If you want a place to inject custom packets that doesn't conflict with the existing protocol, you can check out "mod channels" that are intended for communication between server-side and client-side mods. 17:43 Soni paradust: fancy visual sequences etc? 18:36 sfan5 someone should bother to open an issue for the "msvc on game join" bug, otherwise it will be forgotten 18:42 sfan5 uh, mingw has it too 18:43 grorp and Sokomine, could you open an issue for https://irc.minetest.net/minetest-dev/2024-01-27#i_6148616 ? 18:43 grorp (if it doesn't happen with older versions) 19:20 sfan5 ok there's no meaningful backtrace, so duplicate of #14140 I guess 19:20 ShadowBot https://github.com/minetest/minetest/issues/14140 -- Crash after exiting game session: MinGW thread_local problem 19:20 sfan5 only thing that doesn't make sense is that it happens on msvc 19:20 sfan5 but I don't have a debug build of that available 21:43 MTDiscord sfan5: "msvc on game join" https://github.com/minetest/minetest/issues/14324 21:55 sfan5 msvc is probably unhappy with the forward-declaration in serveropcodes.h and clientopcodes.h 21:55 sfan5 but why? 22:02 Soni btw since minetest supports wasm, have you ever thought about repackaging the entire thing to always use wasm? 22:03 MTDiscord sfan5: Are you able to read the values of pkt, pkt->getCommand(), opHandle, etc, in the debugger? 23:41 MTDiscord sfan5: commented the issue 23:44 MTDiscord I'm curious, how did you debug this to determine that this is (likely) the issue? 23:49 MTDiscord what in the world. Microsoft just decided to "alter" the C++ standard arbitarily? 23:54 MTDiscord I looked closely at the assembly code and figured out that when compiling clientopcodes.cpp compiler thought that sizeof(struct ToClientCommandHandler) == 16. But when compiling client.cpp it was 24 for some reason. I thought it was related to forward declaration and tried to reproduce this behavior - https://godbolt.org/z/McMTbo6Tj. Try to remove __multiple_inheritance and see that static_assert fails.