Add `Diagnostic` object for lint emissions

[xFrednet]

The error messages of Rust are remarkable, creating them can be challenging, though. Rustc uses a Diagnostic struct and a DiagnosticBuilder to create these messages. For this project, I'd like to have something similar.

Things to keep in mind:

1. The Diagnostic object has to be FFI safe (Maybe it can first use Vec during construction and then swap to an FFI safe diagnostic)

2. The information of the Diagnostic backend should be the common denominator off all backends as it kind of dictates, what each backend has to provide as a lint emission mechanism.

   Rust-analyzer supports rustc's diagnostic structure AFAIK, so it's most likely safe to take that as a rough example.

And probably plenty of other things I can't think of rn.

[xFrednet]

I suggest binding each diagnostic object/lint emission to a node. Here is the background:

In rustc, diagnostic objects are emitted at a specific span, while the lint level and type context is determined by the node, that is currently processed by the LintPass. This can sometimes cause a separation of the span and the node that is currently used for the emission.

It can make sense, to let the user decide the span of the lint emission. However, I'd suggest making the emission node explicit. This can avoid some mistakes especially when it comes to lint expectations (RFC 2383).

We can then also use the node, to determine if it originated from a macro, therefore implementing the MacroReport option, which can be set in the declare_lint! macro. It basically gives us more metadata about the lint emission, which might also be useful in the future. :)

[xFrednet]

A reddit post recently referenced some crates to create rustc like diagnostics. These might be good references as well, what the interface could look like:

• ariadne
• miette
• codespan-reporting