Rust Glancer: Rust LSP using 100x less RAM

The project known as Rust Glancer is an alternative Rust Language Server implementation that focuses on low memory usage. This project was developed over a period of four months, with the memory consumption staying under 100MB during testing on an older MacBook Pro M1 2020 with 8GB of RAM. While Rust Glancer is not a complete LSP, it already has many of the necessary features, including a full indexing pipeline with type inference and a trait solver (chalk), support for most of the normal Rust syntax, and common LSP actions such as goto definition, hover, inlay hints, and completions.

One of the main reasons for Rust Glancer's lower memory usage, compared to rust-analyzer, is the fact that it uses a frozen analysis result that gets invalidated on save, rather than making incremental updates. This approach allows Rust Glancer to index the workspace once and preserve results in the filesystem, with queries loading the required information for the duration of their execution.

However, this method does result in slower indexing times compared to the incremental approach used by rust-analyzer. To compensate for this, Rust Glancer employs techniques such as shallow analysis of the current body and reusing the previous complete index during typing, which helps maintain reasonable performance for features like completions.

Despite its lower memory footprint, Rust Glancer has some drawbacks compared to rust-analyzer. It is not as fast, and it may not support all features or handle complex cases as well. However, for users with weaker machines or those who are willing to make some sacrifices for the sake of reduced RAM usage, Rust Glancer offers a viable alternative to rust-analyzer.

The project's creator, who has extensive experience with Rust and its tooling, emphasizes that Rust Glancer is not intended to replace rust-analyzer entirely, but rather to cater to users with specific needs or limitations.