Thank you

[tomoverlund]

Every so often I sojourn into a C++ project in a new virtual machine setup just for that purpose. I usually resort to grepping when using Emacs, but more recently I've been playing around with language servers. I was very pleased when I eventually figured out that, after installing clangd, all I had to do was M-x eglot on a C++ file inside the project, and it just worked, no further setup required.

[joaotavora]

Yes, you just bumped into the #1 design goal of Eglot, and it makes me happy you appreciate it!

[tomoverlund]

It took some digging, as search results were herding me to lsp-mode. But I remembered from past adventures that there was an alternative, read some Reddit and Hacker News comments, eventually found my way to the Eglot GitHub page, read the "Historical differences" section, read the linked issue, and then came to this:

"Then I navigated to a Rust file in a git project, and typed M-x lsp, expecting greatness."

I thought, "Really? I can do that in Eglot, and not have to configure anything, including build commands?" It was even better when I read that Eglot was now builtin to Emacs, and I had already installed clangd at that point, and I tried it... and it worked! Blown away! Even Eclipse CDT, which I've resorted to in the past, has never been that simple.

[joaotavora]

, including build commands?"

Keep in mind that in C++/clangd you'll usually need a compile_commands.json file (or often a symlink to such a file) to be present , usually in the root of your project. Without it, clangd doesn't know how your translations units (.cpp files) are compiled and cannot guess include directories, and many other important things.This is an intrinsic of any C++ analysis tool, and not a problem of Eglot or any other client, or even LSP.

So there is usually this bit of manual labour involved (unless your C++ project is quite trivial).

Rust and many other languages don't need this, as they have simple build systems or a standardized build system.

[tomoverlund]

Keep in mind that in C++/clangd you'll usually need a compile_commands.json file

Ah, I see. Upon closer inspection, that file was already present in the build/ directory.

[joaotavora]

Yes, and the most common practice by far is to symlink it to the root directly, so clangd can find it by default. If you can't do that, there are more options.