Magnus Hagander wrote:
> It seems the only way to dump stuff in pg_database.datconfig ("ALTER
> DATABASE foo SET something=bar") is to do an unqualified pg_dumpall.
>
> I think this should *at least* be dumped with pg_dump when it's set to
> create the database, and I think a case could be made that it should
> *always* be dumped when a complete database is dumped by pg_dump (not
> restricted by schema or table or function or whatever).
>
> Comments?
>
> (I'll await the result of the discussion that's sure to happen before I
> write up a patch for it, but I'm willing to do said patch once agreement
> has been reached)
Hm. Ok, it seems I suck, and this discussion happened just over a month
ago at http://archives.postgresql.org/pgsql-hackers/2008-06/msg01035.php.
I agree with Richard that the behavior is not very nice though :-)
And I think
http://archives.postgresql.org/pgsql-hackers/2008-06/msg01142.php would
be the simplest way to fix it.
If there are no further comments, I'll crawl back under and try to work
up a complete proposal incl. code :-)
/Magnus
