Stephan Szabo <sszabo@megazone23.bigpanda.com> writes:
> On Tue, 20 Aug 2002, Tom Lane wrote:
>> Less obviously, void_in should succeed (and return nothing
>> interesting, probably just a zero datum; it can ignore its input). This
>> allows plpgsql functions to be defined to return VOID.
> Does this require additional work to the plpgsql grammar?
I suspect you'd need to say "return 0" (or return anything-at-all,
pretty much) to make it fly with the current plpgsql sources. This
is a tad ugly but I think we can live with it until someone wants to
fix it. If we have type void then for sure people will want to use
it for plpgsql functions; there are plenty of cases where you run a
plpgsql function just for side-effects.
> I think we should throw the notices right away, although this makes me
> wonder in general about upgrade path. Are we ever planning to make that
> an error, and if so, how are we going to handle functions that are coming
> from previous versions where it was okay?
We can't make it an error until sufficiently far down the road that we
don't care about forward compatibility from 7.2-or-before dump files.
That'll be a long while, probably.
Throwing a notice right away is okay with me personally, but I wanted to
see what other people thought...
regards, tom lane