Removal of special case handling of non-blocked vector?

[schnellerhase]

A vector associated with a single space/form can optionally carry block data, or not. Can this be unified?

[jorgensd]

Since we have had a side discussion of the future of "block" data which is not nest, I’m not sure what we should go for here, ref:
#3728

You also get to the potential issue that people want to write generalised code that works on block systems, that should also work on systems with one block. We had this discussion for mixed element, ref #2944

[michalhabera]

Good point, in case "block" is removed this is a non-issue.