Re: Keeping CURRENT_DATE and similar constructs in original format

"David G. Johnston" <david.g.johnston@gmail.com> writes:
> On Thursday, May 12, 2016, Tom Lane <tgl@sss.pgh.pa.us
> <javascript:_e(%7B%7D,'cvml','tgl@sss.pgh.pa.us');>> wrote:
>> (I'm not particularly in love with the node type name
>> ValueFunction; anybody got a better idea?)

> SQL99DateTimeFunction (or roughly whenever they were introduced)?

Some of them aren't datetime-related, though.  I thought about
NiladicFunction but it seemed maybe too technical.

> I agree with the premise.  I took notice of it recently in explain output
> on these lists using current_date.  That example read like
> ('now'::cstring)::date which was really odd since I was at least expecting
> text as the intermediate cast...

Yeah, that's another fun thing: the reverse listing currently differs
depending on whether you're looking at an expression tree that's been
through const-folding.  It didn't use to --- looks like the mention
of cstring started in 9.2.
        regards, tom lane