On Wed, Mar 11, 2015 at 6:19 PM, Kevin Grittner <kgrittn@ymail.com> wrote:
> Either way it is like leaving the barn door open so that horses are
> capable of running out. We have an alarm that lets you know when
> something is going through the barn door; the question is whether
> to default that alarm on or off.
What we have an alarm that lets you know that your perfectly
legitimate code might have meant something different in a release you
aren't running. If that helps you catch a bug in code you are porting
from a previous version, great. But otherwise it's simply a nuisance.
I don't think this is like leaving the barn door open so that horses
are capable of running out. I think it's more like insisting on
wiring an alarm to every barn door in the county whether there are any
animals currently housed in that barn or not. Now there is nothing
wrong with giving away free alarm systems, but insisting on turning
them all on except for the people who explicitly turn them off seems a
little pushy.
Also, given Tom's remarks downthread, we seem to still lack a
plausible use case where the same code is legal in both versions but
works differently. We should really keep trying to find one of those,
because I think it would shed a lot of light on this debate.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
