Re: When an index and a constraint have the same name...

"Jason L. Buberel" <jason(at)buberel(dot)org> writes:
> Thanks for taking a look Tom:
> I am running postgres 8.1.4 on RedHet (CentOS) v4.0. Here is the
> description of the purchase_record table (somewhat abbreviated with
> uninvolved columns omitted):

Well, I was hoping I could duplicate the problem, but I can't; which
means either that it's been fixed since 8.1.4 (but I don't see anything
very relevant-looking in the CVS logs) or that there's some critical
factor we haven't identified about your table definitions.

Could you send me the full "pg_dump -s" output for the two tables?
And/or a stack trace from the point of the errfinish() call? (Note
there will be two errfinish calls, one to print the NOTICE and one
to print the ERROR --- we want to know where the second one comes from.)

regards, tom lane