Hannu Krosing <hannu(at)skype(dot)net> writes:
> I remember that discussion, but I'm surprised that something got
> implemented and accepted into core with so many unsolvable
> problems/logical inconsistencies/new pitfalls.
The current behavior of ALTER ADD COLUMN & SET DEFAULT is per SQL spec.
If you feel it's inconsistent, take it up with the standards committee.
regards, tom lane