Add a continuously updated (not necessarily incremental) code model to the compiler API

[novusnota]

To position itself well for the future in terms of IDE support, I think it would be nice if the compiler gets the basic data model for IDE use-case. That is, there should be an API to create a persistent analyzer process, which ingests a stream of code edits, and produces a continuously updated model of the code without explicit compilation requests. The model can be very simple, just “give me an AST of this file at this point in time” would do — all the fancy IDE features can be filled in later. What matters is a shape of data flow through the compiler — not an edit-compile cycle, but rather a continuously updated view of the world.
— from https://matklad.github.io/2023/03/26/zig-and-rust.html

As of today, this should be possible using Ohm's incremental parsing capabilities or at least providing the parsed AST, but without analyzed semantics. Ohm is not error-resilient or even error-recovering, so until we have #286, IDE-related tools like Language Server are likely to augment the compiler's/Ohm's parsing capabilities by tree-sitter-tact.

If you want to get something working quickly, I think today the best answer is “just use Tree-sitter”, so you’d better read its docs rather than this tutorial.
— from https://matklad.github.io/2023/05/21/resilient-ll-parsing-tutorial.html

Related to #285.