Philip Warner <pjw@rhyme.com.au> writes:
> Is anybody planning to fix the problem with ALTER TABLE ADD CONSTRAINT...
> in which the constraints are not applied to child tables?
AFAIK no one is looking at it presently (although Stephan Szabo has
probably thought about it). If you want to tackle it, step right up,
but coordinate with Stephan.
I was just in the vicinity of ALTER TABLE, and noted that that routine
didn't have the same loop-over-children superstructure that most of the
other ALTER code does. Should be a relatively simple matter to graft
that logic onto it, unless there are semantic funnies that come up with
propagating the new constraint.
regards, tom lane