Re: Polymorphic arguments and composite types

On Sat, 6 Oct 2007, Simon Riggs wrote:

> On Sat, 2007-10-06 at 10:15 -0700, Stephan Szabo wrote:
>
> > > Yeh, it does, but you're forgetting that my original complaint was that
> > > you couldn't use it in an ANY clause, which 4.2 does not exclude.
> > > Bearing in mind you can use a scalar subquery in lots of places, I
> > > thought it worth reporting.
> >
> > Well, but I'd argue that we're now talking about separate issues.
>
> It's simpler than that. I asked a question because the manual isn't
> specific on my original point. I'll do a doc patch to make sure nobody
> makes the same mistake I did and we record all the good points people
> have made.
>
> > > Section 9.20.3 mentions ANY (array expression). The term "array
> > > expression" is not defined nor is there a link to where it is defined,
> > > nor is the term indexed.
> >
> > I'm not sure why we're using a separate term for that.
>
> The term "array expression" is used in the manual, but not defined.

Right. I meant, if those are the only uses, why did we use a specific term
"array expression" rather than relying on saying that the expression given
must have array type.