On Dec 4, 2010, at 11:46 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Robert Treat <rob@xzilla.net> writes:
>> Actually I think I'd even be comfortable with A, either you must name the
>> constraint after the index, or you can leave the constraint name out, and
>> we'll use the index name.
>
> Or we could omit the "CONSTRAINT name" clause from the syntax
> altogether.
>
> I think that allowing the names to be different is a bad idea. That
> hasn't been possible in the past and there's no apparent reason why
> this feature should suddenly make it possible. We will have problems
> with it, for instance failures on name collisions because generated
> names are only checked against one catalog or the other.
So maybe we should start by deciding what the semantics should be, and then decide what syntax would convey those
semantics.
What would make sense to me is: create a pk constraint with the sane name as the existing unique index. If that
constraintname already exists, error.
...Robert