Refactor dataset clone

[tokland]

Currently, when you clone a dataset, the name is the same than the original dataset. Furthermore, if the field name is not focused, the uniqueness validation is not run (since asyncValidations only run on field change), so we end up with datasets with the same name.

Proposal:

• 1) On clone, set name ${oldName} Clone${n}, where n is calculated (with a request to the DB) so the proposed name is guaranteed to be unique.