Tom Lane wrote:
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > You mean dumping and reloading pg_dump output in 7.3 generates errors?
> > I didn't know. Can you give an example? I wasn't aware of that.
>
> Well, looking at the SETs already currently emitted:
>
> $ pg_dump -s regression | grep ^SET
> SET client_encoding = 'SQL_ASCII'; fails before 7.1
> SET check_function_bodies = false; fails before 7.4
> SET client_min_messages = warning; fails before 7.3
> SET search_path = public, pg_catalog; fails before 7.3
> SET default_tablespace = ''; fails before 8.0
> SET default_with_oids = false; fails before 8.0
>
> so I'm not at all clear what you've got against this one.
Very clear. The issue is that I can't find any of these emitted by a
pg_dump version who's native backend doesn't understand them.
I assume that it is expected that a cross-db dump/reload will generate
errors, and it is done rarely for upgrades, but I assume same-version
dump/restore is done more frequently and people don't expect errors.
Is that not a significant distinction?
-- Bruce Momjian http://candle.pha.pa.us EnterpriseDB http://www.enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +