Peter Eisentraut <peter_e@gmx.net> writes:
> On 1/6/14, 10:29 AM, Tom Lane wrote:
>> This will break some of the function sanity checks in opr_sanity,
> Then the tests can be changed.
That will weaken their ability to detect actual mistakes, no?
If there were a large benefit to merging the pseudotype I/O functions,
I'd think this would be acceptable; but merging them seems of mighty
marginal value.
regards, tom lane