Update the process that SASL mechanisms use to set authentication info for the
bind operation.  Previously, they were setting it directly in the client
connection, which introduced the possibility that a failure in bind processing
which occurred after the SASL credentials were validated could result in a
case in which a failure was returned to the client but the associated connection
could still be authenticated in the server.  SASL mechanism handlers will now
set the authentication info in the bind operation and it will only be updated in
the client connection if all processing associated with the bind operation is
successful.

OpenDS Issue Number: 1203