hey just popping in to see if cbindgen's usecase is at all interesting to this project. It needs to slurp up a crate and its dependencies to generate a C header for all the
pub extern "C" fns, including resolving the definitions of any type referenced in that function's signatures. It currently does path resolution in a "basically-works but ultimately breakable and adhoc" way. As a result I(?) need to put in some work to teach it how to resolve Rust 2018 modules, which is a bummer since, ideally, something like this project should be solving that problem for us.
Long term, yes. Short/medium term, likely no: we need to implement a ton of stuff before path resolution works reliably (macros being a significant contributor to the stuff).