On Thu, Mar 17, 2016 at 12:59 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> One idea that might be worth considering is to define the function
> as generate_series(date,date,interval) returns timestamp (without
> time zone). The point here would be only to move the behavior for
> date inputs as far as getting timestamp without tz rather than
> timestamp with tz; which would at least save some timezone rotations
> in typical use, as well as rather debatable semantics. (The fact
> that timestamptz is the preferred type in this hierarchy isn't
> really doing us any favors here.)
That's a fairly tenuous benefit, though, and a substantially different
patch. I think it's time to give up here and move on. We can revisit
this for another release after we've had more time to think about it,
if that seems like a smart thing to do when the time comes.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company