I frequently use Quick Fix in Rust Analyzer for common refactorings like extracting functions, renaming variables, or rewriting let
/match
expressions. It’s fast, reliable, and context-aware.
I think Cursor could benefit from a similar AI-powered Quick Fix feature. After pressing the Quick Fix button, a few intelligent suggestions—generated by the LLM based on the code context—could appear. This would provide a helpful middle ground between Cursor’s LLM autocomplete and the full chat interface.
It would be more focused and stable than autocomplete, yet less involved and disruptive than starting a chat or invoking an agent—ideal for quick, context-sensitive edits.
For example, given the following code piece:
match item2_data {
VdSemExprData::TrivialDelimited {
left_delimiter,
inner,
right_delimiter,
} => todo!(),
_ => todo!(),
}
If I press quick fix, rust analyzer will provide the option of converting match to if let stmt.
However, what I actually want to perform is converting match to let else stmt.
This features are very slow to develop in rust-analyzer (software engineering is hard, especially around compilers). But I can see llms as very quick to get these done. I can start a chat and let agent do it, but it can be slow. Currently, my habit is to first use Rust-analyzer quick fix to convert it to if let, and then hand modify it to let else.