Two commits: one which fixes incorrect annotations in nacl.bindings, and then the changes from #692 which apply to nacl.pwhash. It looks like these were mostly a case of transcribing the types in docstrings to the annotation syntax.
The reason ffi.string is called that is because it assumes it's a C string (and is therefore null terminated). Confusing for str vs bytes in Python though, I agree.
Two commits: one which fixes incorrect annotations in
nacl.bindings
, and then the changes from #692 which apply tonacl.pwhash
. It looks like these were mostly a case of transcribing the types in docstrings to the annotation syntax.