On Tue, Jun 3, 2014 at 4:02 PM, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
> Andres Freund wrote:
>> On June 3, 2014 9:40:27 PM CEST, Robert Haas <robertmhaas@gmail.com> wrote:
>
>> >Maybe we could make it be controlled by TIMING. Seems like it fits
>> >well-enough there.
>>
>> Don't think that fits well - TIMING imo is about reducing the timing
>> overhead. But the server side total time is still interesting. I only
>> thought about tacking it onto COST because that already is pretty much
>> tailored for regression test usage. C.F. disabling the planning time.
>
> Pah. So what we need is a new mode, REGRESSTEST ON or something.
Well, we could invent that. But I personally think piggybacking on
COSTS makes more sense.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company