Re: [HACKERS] PostgreSQL 10 changes in exclusion constraints - did something change? CASE WHEN behavior oddity

Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> Interesting stuff.  Here's a small recommendation for a couple of those
> new messages.

Hm.  I don't object to folding those two messages into one, but now that
I look at it, the text needs some more work anyway, perhaps.  What we're
actually checking is not so much whether the IS DISTINCT FROM construct
returns a set as whether the underlying equality operator does.  If we
want to be pedantic about it, we'd end up writing something like
"equality operator used by %s must not return a set"

But perhaps it's okay to fuzz the distinction and just write
"%s must not return a set"

You could justify that on the reasoning that if we were to allow this
then an underlying "=" that returned a set would presumably cause
IS DISTINCT FROM or NULLIF to also return a set.

I'm kind of thinking that the second wording is preferable, but there's
room to argue that the first is more precise.  Opinions?
        regards, tom lane