Tom Lane wrote:
> Do we have consensus on doing this? Should we have the warning on
> by default, or off?
This is the tough decision, isn't it. I had thought it would default to
off and people would only turn it on in their testing procedure prior to
the actual upgrade of the production systems, but how are they going to
find out they need to turn it on in the first place? We could have a
big fat red blinking warning in the release notes and a picture of a
dancing elephant in a tutu next to it, and we can be certain that many
will miss it anyway.
I think we should have an "expires" value: it means ON unless the
system's initdb is older than one month from the current date, in which
case it turns itself off. This is of course just a silly joke, but as
you said there are probably valid constructs that are going to raise
warnings for no reason, so having it default to ON would be pointlessly
noisy in systems that have been developed with the new rules.
--
Álvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
