On 2000-02-08, Tom Lane mentioned:
> The simplest real solution I've heard so far is to dump database objects
> in order by OID rather than doing it strictly by type.
AFAIR, it was your idea ... ;)
>
> Is anyone working on this, or does anyone want to? I haven't looked at
> pg_dump in a while, but I know some other folks have been hacking it
> recently.
I might have been putting out remarks to that end once in a while, and I'm
still interested in it, but it would be a more extensive project, like the
psql revision, because pg_dump needs a lot of love as it stands. (I think
there are some parts still in it that allow you to dump PostQUEL.)
The problem with a pure oid-based ordering concept is that (as you
yourself pointed out) it won't work if you alter some object in question
after creation. The obvious case would be an alter function (to be
implemented), but another case is (probably) alter column set default (is
implemented).
What I'd like to do first is to draw up some (semi-)formal
(dependency-based) concept on paper and either verify it or come to the
conclusion that it will never work and then give up in disgust. ;) No,
seriously, I suppose I'll bring this up again in a couple of months when
we're ready for it.
Any collaborators are welcome of course.
--
Peter Eisentraut Sernanders väg 10:115
peter_e@gmx.net 75262 Uppsala
http://yi.org/peter-e/ Sweden