I do not understand that comment at all. If you use the data off ancestry, then ancestry is the repository.

You have jumped over into original sources vs. derivatives and that is not what we are talking about at all. We are only talking about repositories.

If you get the data off ancestry, then ancestry is the repository, and the source-cite will identify what the source is...again:
repository (ancestry)
...source (whatever)
......information/data/fact (whatever)

We clearly define things differently, so I don't see how this can go anywhere except round and round. Therefore, I bow out.