"Albe Laurenz" <laurenz(dot)albe(at)wien(dot)gv(dot)at> writes:
> So it looks like the number of result rows is the least common multiple
> of the cardinalities of all columns in the select list.
It's always been that way. The lack of any obviously-sane way to
handle multiple SRFs in a targetlist is exactly why the feature is
looked on with disfavor.
regards, tom lane