The variants for time and timetz had zero test coverage, the variant
for interval only very little.  This adds practically full coverage
for those functions.
Reviewed-by: Vik Fearing <vik@postgresfriends.org>
Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us>
Discussion: https://www.postgresql.org/message-id/flat/c3306ac7-fcae-a1b8-1e30-6a379d605bcb%402ndquadrant.com