Re: Testbed for predtest.c ... and some arguable bugs therein

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Thu, Mar 8, 2018 at 12:33 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> I'm not sure that that's worth fixing right now. Instead I'm tempted
>> to revert the addition of the clause_is_check argument to
>> predicate_refuted_by, on the grounds that it's both broken and currently
>> unnecessary.

> Hmm, I think you were the one who pushed for adding that argument in
> the first place: http://postgr.es/m/31878.1497389320@sss.pgh.pa.us

I'm kind of disappointed that you failed to take the *other* advice
in that message, as I still think that clause_is_check is a poor
choice of name for the flag. It could have been salvaged with a
clear comment defining the semantics, but that's not there either.

> I have no problem with taking it back out, although I'm disappointed
> that I failed to find whatever was broken about it during review.

Maybe I'll spend a few minutes trying to isolate why the current
results are wrong. However, it's certainly arguable that we shouldn't
spend much time on this with no use-case in sight.

regards, tom lane