"Brendan Jurd" <direvus@gmail.com> writes:
> I'm all for the prevalance of sanity, but I'm not really clear on what
> about the above scenario is not sane.
Well, a situation like that just calls into question whether there's
been a mistake --- in particular whether the underlying function is
really null-safe or not.
We could remove the opr_sanity test, but to me that'd be akin to
disabling a compiler warning. Our project policy has generally been
to write warning-free code, instead.
regards, tom lane