Not sure if it is suitable here to fully depend on pamela or not.
It seems however the current design auth(username) -> password cannot coexist with the arguably safer PAM auth(username, password) -- delayed --> Bool behavior, given that both usecases can have a second optional argument with different semantics...
So the alternative to not depend on pamela, is to restrict the design to support only a auth(username, password) -> Bool behavior, and current examples become invalid... -> new major version ?
Not sure if it is suitable here to fully depend on pamela or not.
It seems however the current design
auth(username) -> password
cannot coexist with the arguably safer PAMauth(username, password) -- delayed --> Bool
behavior, given that both usecases can have a second optional argument with different semantics...So the alternative to not depend on pamela, is to restrict the design to support only a
auth(username, password) -> Bool
behavior, and current examples become invalid... -> new major version ?