On 11.10.2010 04:25, Tom Lane wrote:
> It's true for scalar input datatypes, though.
>
> I had been wary of this idea because I didn't see any suitably cheap
> place to insert the necessary processing, but after some reflection
> and rejiggering of eval_const_expression's responsibilities, it's
> done.
>
> regards, tom lane
>
Awesome, thanks a lot! :-)