On Mon, 2011-06-20 at 13:43 -0400, Tom Lane wrote:
> The other viable alternative seems to be to require those two properties
> (btree opclass and collation) to be part of a specific range type
> definition. The complaint about that seemed to be that we couldn't
> infer an ANYRANGE type given only ANYELEMENT, but could we alleviate
> that by identifying one range type as the default for the base type,
> and then using that one in cases where we have no ANYRANGE input?
Yes, that sounds similar to Florian's suggestion, and I think there may
be a solution down this path. However, if we're going to have range
types with non-default orderings, then we need a way to construct them.
I suggested that, if constructors are the primary problem case, then
just generate non-polymorphic constructors at range type definition
time, named after the range type name. I'll look into that approach.
Regards,Jeff Davis