Optimizing a single Rust Clippy lint by 3133X
This article details the optimization of a single Rust Clippy lint called clippy::nonstandard_macro_braces. This lint identifies macro calls with incorrect braces, as Rust macros are not strictly defined. To investigate, the author first checks if the current tokens come from macro expansion. They then trace the source text for the span of the macro expansion, ultimately determining the correct brace using string manipulation.
By calling hygiene data functions twice and locking the symbol interner, the author engages in a recursive loop that impacts Clippy's runtime. The author emphasizes that this issue is not the fault of a single contributor but rather a broader problem that open source maintainers must address. They propose rewriting the problematic function to a pre-expansion lint, eliminating the need for source text calculation, symbol interner locking, and session globals locking.
As a result, Clippy now has a benchmarking server and the optimization has significantly reduced runtime, saving hundreds of thousands of computing dollars.
Written by urgent.news from Lobsters's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
