Thinking in the long term, how could be the best way to clone a schema into another?
Depends on why you are cloning schemas.
Generally not cloning is the best bet - instead place the reference schema into version control and simply install it whenever necessary. Parameterize the installation script.
Or, clone into a different database and leave the namespace the same.
pg_dump, as currently written, is tasked with recreating the original database exactly and safely. If you need it to do something different you could teach it the desired behavior and add some relevant command line switches to enable said behavior.
