Time Nick Message 10:16 ROllerozxa I wonder if the `enable_vbo` setting could be dropped? Enabled by default, I don't know if anyone would want to disable it 12:12 MTDiscord maybe some people's GPU doesn't support it 12:12 MTDiscord (please ask erlehmann before doing this :trollface:) 13:50 MTDiscord why not detect its support and automatically enable/disable it? 13:54 sfan5 LOL 13:54 sfan5 we have minetest.registered_schematics in our documentation and nobody has ever noticed that it does not actually exist 13:55 MTDiscord nobody has complained loud enough to notice or made an issue 14:12 Zughy[m] sfan5: actually we did #7917 14:12 ShadowBot https://github.com/minetest/minetest/issues/7917 -- `minetest.registered_schematics` returns nil even after `minetest.register_schematic` 14:17 sfan5 oh 14:41 MTDiscord The way our Irrlicht build is structured is causing a lot of issues trying to add new features. I suppose the benefit is faster compile times, and with CMake 3.12 and up the issues with adding features go away as well, but we're still trying to support CMake 3.5 14:44 MTDiscord More specifically, we can't link targets to object libraries, so we have to iterate over all our object libraries and give includes etc. to each one. Adding a new library involves adding its includes to a list of includes that are thrown everywhere and the entire concept of usage requirements being local to their dependents that CMake is built on is undermined. 14:45 pgimeno why cmake 3.5? I'm on debian oldoldstable and I have cmake 3.16 14:46 MTDiscord For what it's worth I did manage to find a GitHub Action that lets me download old CMake versions so I set up CMake 3.5 on CI for probably the first time in Minetest/Irrlicht's history. 14:47 MTDiscord Which I guess was worth it because I discovered a breakage in old CMake versions due to an old CMake bug that I had to work around. Which adds to the confusion. 14:47 MTDiscord If we could eventually move up our minimum CMake version to around 3.11 it would probably help clean things up significantly. 14:48 MTDiscord (3.12 would be preferable due to the already noted object library changes) 14:52 pgimeno @josiah_wi I suggest you open a discussion issue about bumping the minimum required cmake version if you don't get a reply 14:52 MTDiscord Ok, I'll see what I can do about that. My internship starts Monday and I'm racing to finish the first glTF loader PR before then, so I might be pretty busy. 15:06 sfan5 pgimeno: ubuntu 16.04 probably which wasn't EOL at the time 15:06 pgimeno and does it support the C++ version required by Minetest? which I can't remember which one it is, by the way 15:07 sfan5 dunno 15:08 pgimeno I mean, I had to download a bunch of packages from backports to be able to compile minetest, and that includes cmake - I think it's very reasonable to require a newer version 15:11 sfan5 the current situation looks like: centos 7 via epel = 3.17, debian 10 = 3.13, ubuntu 18.04 = 3.10 15:12 pgimeno does ubuntu have an equivalent to debian backports? 15:12 sfan5 don't think so 15:13 sfan5 the idea is that if you want to compile something and have to go hunting for a more recent version of maybe not just libraries but the build system itself that is a massive annoyance 15:25 pgimeno ubuntu 18.04 normal support ends in June, by the way - probably before the next release 15:26 pgimeno ubuntu 14.04 is not EOL until April 2024, geez 16:35 MTDiscord I was just bringing it up for discussion so we're aware of the potential mess there. We can make sure to add a lot of comments documenting why things are in the build to help refactor it in a couple years. 17:19 sfan5 I just noticed you were talking about the irrlicht build script 17:19 sfan5 IMO that one is still much more modern than MT's 17:20 sfan5 if we want to improve something that should be first 17:42 MTDiscord Yeah, I like how you structured the Irrlicht build script and that's why I haven't tried to change it. I think it's a good design. 17:43 MTDiscord It's just that because of this issue with object libraries on old CMake versions the benefit of the design from a design perspective is mostly negated. 17:43 MTDiscord It should still result in fast incremental builds, though. 17:46 Zughy[m] there are 5 PRs ready to be merged (2 trivials with one approval and 3 with two approvals) 19:34 MTDiscord How should I do the copyright in Irrlicht? Jordan and I have both worked on this code. 19:35 MTDiscord I want to make sure he gets attributed for his work. But in COBJMeshFileLoader.cpp there's only one author listed in the copyright from a long time ago, with no contact information such as email.