From: Alvaro Herrera <alvherre@alvh.no-ip.org>
> I'm pretty sure I named the flag PQTRACE_SUPPRESS_TIMESTAMP (and I
> prefer SUPPRESS to NOOUTPUT), because the idea is that the timestamp is
> printed by default. I think that's the sensible decision: applications
> prefer to have timestamps, even if there's a tiny bit of overhead. We
> don't want to force them to pass a flag for that. We only want the
> no-timestamp behavior in order to be able to use it for libpq internal
> testing.
Agreed. It makes sense to print timestamps by default because this feature will be used to diagnose slowness outside
thedatabase server. (I misunderstood the motivation to introduce the flag).
Regards
Takayuki Tsunakawa