program-transformer

Source-to-source transformation is achieved through semantics-preserving passes that manipulate the AST while maintaining the original code's meaning. The skill applies desugaring, normalization, and optimization passes to modify Haskell, OCaml, and Python codebases.

AST manipulation involves traversing and modifying the abstract syntax tree of code. It uses specific traversal strategies and binding management to ensure variables and scopes remain valid during semantics-preserving program transformation.

Yes, you can refactor a large Python codebase by automatically applying consistent transformations. It converts older language constructs to newer, more efficient ones through optimization passes and normalization.

Applying optimization passes may require closure conversion and CPS transformation beforehand. These dependencies prepare the intermediate representation by converting closures and transforming control flow for effective AST manipulation.

The best way to improve code efficiency is applying desugaring and normalization passes during program transformation. These optimization passes simplify complex language constructs into core forms, enabling further AST manipulation and performance improvements.