Currently, the crater report just sorts the results alphabetically.
It would be nice to instead sort them by relevance (crates.io downloads/github stars) because especially for more "who relies on this" kind of experiments, people don't have time to investigate 100 different crates, so it would be nice to know which ones matter the most.
For example in the one shown above, crossbeam-channel and crossterm show up, two widely used crates. These should definitely be investigated!
By not sorting them, we have a higher risk of missing important regressions and wasting time on irrelevant ones (like, no one cares if some weird crate with almost no downloads relies on an implementation detail).
Currently, the crater report just sorts the results alphabetically.
It would be nice to instead sort them by relevance (crates.io downloads/github stars) because especially for more "who relies on this" kind of experiments, people don't have time to investigate 100 different crates, so it would be nice to know which ones matter the most.
For example in the one shown above, crossbeam-channel and crossterm show up, two widely used crates. These should definitely be investigated!
By not sorting them, we have a higher risk of missing important regressions and wasting time on irrelevant ones (like, no one cares if some weird crate with almost no downloads relies on an implementation detail).