Default string data types used for Unicode connection

ODBC-185 fixed that in the resultset returned by SQLColumns, but that is not enough. Different WCHAR types should be default for char columns in case of unicode connection(i.e. Unicode part of API is used). That implies type stored in IRD, returned in case default type is requested, returned by SQLColAttribute and if decriptor field value requested, etc. This is better to do in terms of new version. If to do.