1.0 will be a huge milestone for Rustfmt. As part of it’s stability guarantees, it’s formatting will be frozen (at least until 2.0). That means any sub-optimal formatting still around will be around for a while.
What formatting changes would be found if rustfmt were to be run on every crate in crates.io, and could they be categorised (by instrumenting rustfmt)?
If a strong pattern emerges then that could indicate a mismatch between rustfmt (or the guidelines) and current practice. This could also be used to measure how common custom formatting settings are used in the wild.