Changes 2.2.3:
abstraction layer query converts int values to string (thanks dgadmin)
CaseInsensitiveDictWithAlias doesn窶冲 raise an exception anymore if alias is set multiple times to the same key
friendly names in AttrDef are properly managed when performing commits in Writer cursors
no more errors when server returns an empty schema (thanks Glen)
range attributes in entries are properly recognized when auto_range is False
fixed random errors in auto_range searches (thanks James)
fixed checking of malformed schema
added configuration parameter IGNORE_MALFORMED_SCHEMA to not raise exception for servers that don窶冲 follow the LDAP RFCs (defaults to False)
test config moved to test/config.py
testcase_id generated randomly for each test
added ATTRIBUTES_EXCLUDED_FROM_OBJECT_DEF parmeter to exclude some attribute from automatic populate of ObjectDef in Abstract Layer (helpful for AD)
added IGNORED_MANDATORY_ATTRIBUTES_IN_OBJECT_DEF parmeter to exclude some attribute from mandatory attribute list in ObjectDef in Abstract Layer (helpful for AD)
fixed error when using implicit assigning in WritableEntry
added LDAPInvalidValueError Exception
in Python 3 byte filter are converted to unicode before parsing
RESPONSE_DN_ENCODING parameter renamed to ADDITIONAL_ENCODINGS
to_unicode(value, encoding=None, additional_encodings=False) now checks for additional encoodings in ADDITIONAL_ENCODINGS list if additional_encoding is set to True
Reusable strategy uses not lazy Restartable connections

(adam)