I wrote:
> Andrew Gierth <andrew@tao11.riddles.org.uk> writes:
>> The hstore_compat one is arguable as to what is the best approach; the
>> assert that was there was just wrong, but I have been unable after
>> considerable searching to find any architectures that would fail the
>> check.
> [ scratches head... ] It looks like that ought to be an immediate
> core-dump for old data, given an assert-enabled build. Are you
> saying it isn't? How?
I tried this, and indeed an assert-enabled hstore dumps core instantly
on a pg_upgraded table. So that upgrade path obviously hasn't been
tested very well. But I don't see why we don't just fix the Assert.
regards, tom lane