| Time |
Nick |
Message |
| 04:00 |
|
MTDiscord joined #luanti-docs |
| 08:17 |
|
[MatrxMT] joined #luanti-docs |
| 13:37 |
|
[MatrxMT] joined #luanti-docs |
| 22:54 |
|
GreenXenith joined #luanti-docs |
| 23:00 |
|
MTDiscord joined #luanti-docs |
| 23:28 |
MTDiscord |
<wsor4035> so im not toe stepping, https://github.com/luanti-org/flint/issues/5 should flint stick to its own lsp for now, or wait (rip me) on lsp support for the engine? and should it be the engine lsp, or that be its own separate thing? cc @greenxenith @mark.wiemer |
| 23:30 |
MTDiscord |
<greenxenith> "Engine LSP"? |
| 23:30 |
MTDiscord |
<wsor4035> vscode lsp for minetest/core, and yalls wanting a machine parsable docs |
| 23:30 |
MTDiscord |
<greenxenith> And flint's approach is what, as opposed to docs? |
| 23:31 |
MTDiscord |
<wsor4035> write now nothing, but similar to biome js, i want to eventually show lint results in editor |
| 23:31 |
MTDiscord |
<wsor4035> so question is just that, or build in engine api support as well |
| 23:33 |
MTDiscord |
<wsor4035> for context https://biomejs.dev/reference/language-server/, https://biomejs.dev/reference/vscode/ |
| 23:38 |
MTDiscord |
<greenxenith> We should be using LuaLS as the LSP, no reason to reinvent the wheel. Which means generating LuaCATS annotations from the docs |
| 23:39 |
MTDiscord |
<greenxenith> I believe that was Mark's conclusion as well |
| 23:40 |
MTDiscord |
<greenxenith> Correction: That was one of Mark's previous conclusions, until he pivoted to a custom LSP |
| 23:40 |
MTDiscord |
<wsor4035> so ill just stick to linting lsp then |
| 23:40 |
MTDiscord |
<wsor4035> ...or maybe? |
| 23:41 |
MTDiscord |
<wsor4035> ....or maybe |
| 23:42 |
MTDiscord |
<greenxenith> I dont actually understand why Mark abandoned LuaLS as an LSP. I understand avoiding using raw LuaCATS as documentation, but it would be a lot less work to transpile MD into LuaCATS IMO |
| 23:43 |
MTDiscord |
<greenxenith> What are you doing beyond luacheck? |
| 23:44 |
MTDiscord |
<wsor4035> https://docs.luanti.org/for-creators/flint/lint/ |
| 23:44 |
MTDiscord |
<greenxenith> Understood |
| 23:45 |
MTDiscord |
<greenxenith> So in theory, Flint could include both Selene and LuaLS |
| 23:45 |
MTDiscord |
<greenxenith> Which is the biomejs approach, I suppose? |
| 23:45 |
MTDiscord |
<wsor4035> biomejs sticks to just linting, its not a js lsp |
| 23:46 |
MTDiscord |
<wsor4035> i think |
| 23:46 |
MTDiscord |
<greenxenith> I think it does have an LSP |
| 23:47 |
MTDiscord |
<greenxenith> Either way, nothing stops Flint from having both |
| 23:47 |
MTDiscord |
<greenxenith> Assuming they dont overlap |
| 23:47 |
MTDiscord |
<wsor4035> it does have an lsp, its just for linting, not the js library of functions. whereas here im sorta proposing linting + engine library of functions |
| 23:48 |
MTDiscord |
<greenxenith> I think it would be beneficial to include LuaLS alongside Selene |
| 23:48 |
MTDiscord |
<greenxenith> Then Luanti Tools can leverage Flint |
| 23:49 |
MTDiscord |
<greenxenith> Docs' job is to make the API ref consistent enough to parse to LuaLS definitions |
| 23:49 |
MTDiscord |
<wsor4035> LuaLS is in lua, while everything in flint is rust. guess i technically could run lua via rust |
| 23:52 |
MTDiscord |
<greenxenith> EmmyLua is Rust, but it isnt quite as complete as LuaLS, especially for C++ Lua API stuff |
| 23:54 |
MTDiscord |
<wsor4035> it does appear to support Luacats, so if your using that as your "compile" target.... |
| 23:54 |
MTDiscord |
<mark.wiemer> Heyo, glad there's energy here. I'm still recovering from major illness but I've poked around with "my" work a bit last month. I didn't like LuaCATS because of its low expressiveness, but I can't remember examples. Next goal for me is just building my own mods and tinkering with functions to see how they work. Basically gather/interpret data before trying to structure it. I have stepped back from multiple other obligations to |
| 23:54 |
MTDiscord |
recover faster, but I don't think I'll be able to do much this month. Should have more energy my mid-next month |
| 23:55 |
MTDiscord |
<mark.wiemer> You can find my work on github at mark-wiemer/hello-hello, there is a Luanti package and you can browse the history for notes and findings |
| 23:55 |
MTDiscord |
<greenxenith> It supports LuaCATS as definitions, but there are features it doesnt support like meta tags, environment extensions, etc. |
| 23:55 |
MTDiscord |
<mark.wiemer> Gtg now, @me if you need me and I'll respond sometime tomorrow |
| 23:55 |
MTDiscord |
<greenxenith> Thank you sir |
| 23:55 |
MTDiscord |
<wsor4035> be a bit odd to use it for lsp, and it has a lint and formatter when im already using stylua+Selene |
| 23:56 |
MTDiscord |
<greenxenith> Id just have to be convinced that it is worth the technical debt to implement our own or use something that obviously misses some targets when a comprehensive standard is available |
| 23:57 |
MTDiscord |
<wsor4035> i mean, fair |
| 23:57 |
MTDiscord |
<wsor4035> its what flint does, uses the standards and then builds on top of then for luanti specific support + bundling everything in one tool |
| 23:58 |
MTDiscord |
<wsor4035> anyways, happy to just do a lint lsp, but also happy to do lint+engine lsp |
| 23:58 |
MTDiscord |
<greenxenith> Running Lua from Rust isnt the worst thing in the world, but the duplication of responsibility is worth considering |
| 23:59 |
MTDiscord |
<wsor4035> i think from a users perspective, its probably better that they have just one thing that does both, but 🤷 |
