A user reported encoding errors in their logs when attempting to log in

a new user when connected to LDAP. The issue here was that data returned

from ldap was bytes, but we were attempting to split it using a

unicode splitter. According to the ldap documentation, field names are

returned as the text type, but attribute values will be bytestrings. The

documentation is a unclear in that it also claims that "text values"

will be the unicode type. In order to accomodate potentially different

behavior across Python versions, all values have been wrapped in

force_text. The unit tests have been updated to supply the expected

(bytes) types for values.

Ran unit tests on Python 2.7 and 3.8.