A class is used to resolve the normalizer mapping hash used for normalization.
This interface is implemented and passed into several kinds of parsers that
need to handle the normalization of LDAP name strings.
Why you may ask are we doing this? Why not just pass in the map of
normalizers to these parsers and let them use that? First off this mapping
will not be static when dynamic updates are enabled to schema. So if
we just passed in the map then there would be no way to set a new map or
trigger the change of the map when schema changes. Secondly we cannot just
pass server side objects that return this mapping because these parsers may
and will be used in client side applications. They will not have access to
these server side objects that generate these mappings. Instead when a
resolver is used we can create mock or almost right implementations.