Robert Haas wrote:
> Hmm, ISTM that cardinality() is implemented here in the manner
> previously rejected for array_length()...
The objection was that basic functionality should not be implemented in
SQL. If we want to disallow all compatibility functions implemented in
SQL as well, we have more work to do.
>
> ...Robert
>
> On Wed, Nov 12, 2008 at 8:13 AM, Peter Eisentraut <peter_e@gmx.net> wrote:
>> Robert Haas wrote:
>>> Updated version attached, this time without the compiler warning.
>> I have committed something based on this. The issue of empty arrays will
>> need a separate solution.
>>
>