Is a ``zig lsp`` feasible?

Heyo, this might aswell be under the category “Brainstorming” and not just “Explain”.

I was thinking about lsps and zig. To my knowledge, zls isnt capable of executing code at comptime and the only real way to deal with zig types and make a really good lsp is to basically just make a complete Sema.zig-based lsp - or just use the zig compiler itself.

Is this something the zig team is already working towards? I feel that incremental compilation may be a strong step towards this. If not, are there any good pros and/or cons for this?

Id be happy to both learn and hear some opinions.

1 Like

The relevant issue is

Other related links:

8 Likes

Thank you very much!