Thread: 8.0.0beta1 small inconsistency in pg_aggregate listing

8.0.0beta1 small inconsistency in pg_aggregate listing

From
Fabien COELHO
Date:
I've noticed a minor inconsistency when listing pg_aggregate in
current 8.0.0beta1 version.

psql> SELECT aggfnoid FROM pg_aggregate;

4 functions are not schema qualified:
 count
 bool_and
 bool_or
 every

although all others are shown with a schema name:
 pg_catalog.avg
 pg_catalog.sum
 pg_catalog.max
 pg_catalog.min
 ...

However psql \da show all of them in pg_catalog, as expected.

I cannot see the difference in the initialization script to explain that.
Maybe something wrong with the regproc type?

--
Fabien Coelho - coelho@cri.ensmp.fr

Re: 8.0.0beta1 small inconsistency in pg_aggregate listing

From
Tom Lane
Date:
Fabien COELHO <coelho@cri.ensmp.fr> writes:
> Maybe something wrong with the regproc type?

regprocout doesn't schema-qualify if the function name is unique without
it.

Of course, the function name isn't necessarily unique *with* it, either,
but regproc can't do anything about that.  Cast to regprocedure if you
want to see a more helpful display of the function.

            regards, tom lane

Re: 8.0.0beta1 small inconsistency in pg_aggregate listing

From
Fabien COELHO
Date:
> > Maybe something wrong with the regproc type?
>
> regprocout doesn't schema-qualify if the function name is unique without
> it.

Ok. Maybe too clever, as I couldn't guess by looking at it.

> Of course, the function name isn't necessarily unique *with* it, either,
> but regproc can't do anything about that.  Cast to regprocedure if you
> want to see a more helpful display of the function.

Much better indeed.

Thanks a lot for the light. Have a nice day.

--
Fabien Coelho - coelho@cri.ensmp.fr