Hi folks. I went through the last 2 years of rustc upgrades in a project I maintain and found quite a bit of breakage due to clippy lints with high blast radius or false positives.
#![warn(clippy:all, clippy::pedantic)] to all of my crates. These are some of the not fun bugs I've encountered:
I've suggested in a few of these issues that new clippy lints be tested with crater, but haven't gotten much engagement. I'd like new lints to go through crater runs to determine:
- What blast radius they have.
- How much churn they suggest.
- Whether there are false positives.
- Whether the suggestions compile.
I think a theme of these linked tickets is that there tend to not be tests for code that should not be triggered by a new lint. In a world where the project has the infrastructure to test compiler changes on the entire ecosystem, I'd hope there would be fewer false positives in clippy that make it to stable.
Is investment in this area something that would be in scope for the clippy 2021 roadmap? https://github.com/rust-lang/rust-clippy/blob/master/doc/roadmap-2021.md