On 2020-09-29 18:37, Tom Lane wrote:
> Unfortunately this has still got a problem: it will mishandle the case of
> a child column that is GENERATED while its parent is not. Peter opined
> way upthread that we should not allow that, but according to my testing
> we do.
Did I opine that? Commit 086ffddf3656fb3d24d9a73ce36cb1102e42cc90
explicitly allowed that case. What we don't want is the other way around.
--
Peter Eisentraut http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services