elixir-ls VS flake-utils

No. Not even close. But it's getting better.

There are currently two worth mentioning:

ElixirLSP: https://github.com/elixir-lsp/elixir-ls

Elixir tools: https://www.elixir-tools.dev/

ElixirLSP is the older project, and has been around for a while. It does a lot, but has had sporadic issues over the years. Things like the debugger are a dog to get working, and the server itself will occasionally run into issues where it just doesn't want to work. It's always sort of focused on a subset of language server features, so don't expect much in the way of inline corrections. But it's got the essentials, formatting, basic linting, type hinting, on demand documentation, and primitive reference navigation

Elixir tools is a new up and comer, written by Mitchell Hanberg. It's aiming to be a more complete lsp, and has plugins in its "ecosystem" for most editors. Features have been arriving rapidly, starting with things like inline corrections and far more reliable linting, and recently growing autocomplete. One of the main selling points is the elixir-tools backend is a self contained binary, so it can mostly work independent of system Elixir/Erlang version, which was a frequent tripping point for ElixirLSP

Personally I use both at the same time, but plan to move to tools only when it's got all the features I need

I decided I can live without elixir-ls when couching in return for having a usable editor. When the plugin ecosystem and documentation matures I can see myself switching to using Lapce for my primary editor.

Hi! I've compiled gdb from master and installed it. When I run gdb -i dap, I get JSON-RPC, so it looks like it's working, but I'm lost as to where to go from here. Does your change enable me to use a tool like https://github.com/elixir-lsp/elixir-ls with GDB now, right? How might I use them together? What would be required to then have GDB debugging over DAP from inside Emacs?

Elixir-ls provides Language Server Protocol support as well as VS Code Debug Protocol support which gives extra powers to VS Code, NeoVim, Emacs, and the like

https://github.com/elixir-lsp/elixir-ls

BUT, with the news that in the next Emacs version eglot (lsp client) is going to be in the core, I decide to modify alchemist in a different way, enhancing other capabilities and letting eglot do what he best does (which is the functionality of elixir-ls).

I think it's here: https://github.com/elixir-lsp/elixir-ls

elixir-ls powers the VSCode experience, and while it works decently, I've never really clicked with VSCode. In general, for other languages, I tend to use Jetbrain's Products. I would love to give fleet a whirl, and I know in the background it can use the Language Server Protocol to support many of the languages it currently ships with.

Now it doesn't mean that immediately reading this starter guide, you will have everything under the sun set up with Nix Flakes for your development need. But at least, you won't have to worry about setting up asdf, your weird hacks you need for your machine and the other tiny little things to get elixir started with elixir-ls.

I have a minimal lspconfig with coq_nvim with elixirls and tsserver. The problem is that whenever i code on elixir everything becomes slow.

Well I know one extension that runs your code, elixir-ls. I believe it scans your code and runs dialyzer, a static analysis tool, which runs your code and generates types based on it.

The ‘flake-utils’ readme is a pretty good jumping off point: https://github.com/numtide/flake-utils

I have this or that nitpick with FL and FLP but overall it’s very solid stuff. FLP is a little more “magical”, and that’s not always the best starting out, but you really can’t go wrong with either.

We added two inputs, the first is nixpkgs which lets us specify which version of nixpkgs we should use. There are many thousands of packages in the nixpkg repository, and they are updated often so here will use the unstable branch. We also added flake-utils which helps us generalize the flake to support multiple systems, not just Linux.

The inputs is how you can import external sources of other flakes into the flake project you have. In other words, any project you may need or tools required to get started, this is where you will define their source. Example below is using the standard nixpkgs and a tool called flake-utils, which provides a set of functions to make flake nix packages simpler to set up without external dependencies.

I'm not sure what nix-shell --pure does, but is it equivalent to using a flakes.nix in your projects? Ie i use https://github.com/numtide/flake-utils and direnv to replicate the old shell.nix with a Flakes setup. Per project i have a flakes.nix and a flakes.lock, so it feels just like my old shell.nix setup, but using flakes instead.

A Nix Flake is just an object - check out those surrounding curly braces. This object has two keys, inputs and outputs. The inputs are where we define the flake's dependencies and where to find all the tools we use. This one has two, nixpkgs and flake-utils. Each of these just points to a GitHub URL, and if you follow those links, you'll see each repo provides its own flake.nix. The outputs of each remote flake get piped into the inputs of my flake, so we can use what they provide.

You can easily transition your shell.nix (and default.nix) to a flake-based one by using flake-utils and flake-compat. The former is actually unnecessary, but I would recommend it for typical project environments. Unless you have an impure dependency, this transition would be easy.

Flakes can provide different types of things: - some flakes provide applications that you can nix run, - some flakes provide functions that you can import (e.g. https://github.com/numtide/flake-utils), - some flakes provide overlays to use with nixpkgs (e.g. that emacs-overlay you posted).