Re: spurious "UNIQUE constraint matching given keys for referenced table" error

"Mathew Frank" <mathewfrank@qushi.com> writes:
> I'm always a bit nervous about calling anything a bug... 9/10 is caused by a
> typo or something on behalf of the user ;)

Well, it's a two-way street: typos and other sorts of mistakes tend to
stress paths that the developers would've never thought to test.
Besides, I'm a firm believer in the notion that software should deliver
useful error messages, and PG definitely wasn't getting that done in
this case.  So complain away.  I'm glad to be able to say "we already
fixed that for the next release", but that won't always be true...

            regards, tom lane