I'm currently prototyping, and I'm rarely interested in the specifics of an error presented by the library.
Would there be any interesting in constraining the traits' associated Error types to core::fmt::Debug so they can be printed and forgotten, or passed along to anyhow! or whatever?
This would be a breaking change, obviously, but I'm curious whether you think this would be valuable? Or is flexibility in the interface more important to you?
I'm currently prototyping, and I'm rarely interested in the specifics of an error presented by the library.
Would there be any interesting in constraining the traits' associated
Error
types tocore::fmt::Debug
so they can be printed and forgotten, or passed along toanyhow!
or whatever?This would be a breaking change, obviously, but I'm curious whether you think this would be valuable? Or is flexibility in the interface more important to you?