> > Come to think of it, I wonder if I could have the disk copy of
> > pg_attribute use the pg_type length, and use the pg_attribute length
> > only when creating/updating entries? I wonder if that is what it does
> > already. Looks like that may be true.
> >
> > Comments?
>
> Is what you are trying to do related to what could be used to implement
> other (SQL92) data types like numeric(precision,scale) where there are one
> or two additional parameters which are assigned when a column/class/type is
> defined and which must be available when working with column/class/type
> instances? We probably don't want to do anything about the latter for v6.3
> (spread pretty thin with the work we've already picked up) but I'd like to
> do something for v6.4...
>
> Oh, while I'm thinking about it, this kind of thing is probably also
> necessary to get arrays working as expected (enforcing dimensions specified
> in the declaration).
Yes, I had the numeric in mind.
--
Bruce Momjian
maillist@candle.pha.pa.us