Re: introduce "default_use_oids"

Bruce Momjian  writes:
> We need a solution to this.

I'm really not sure we do: the SQL produced by pg_dump was totally
non-portable before, and it is slightly less portable now. ISTM that
you will almost always need to do some post-processing of pg_dump's
output to have any hope of loading it into another RDBMS in any case.

A flag for pg_dump that told it to produce standards-compliant SQL
would be a cool thing to have, but that would provide a solution to a
problem that has always existed, not one that is introduced by this
patch.

> One idea is to use SET to change the default_with_oids setting when
> a table changes characteristics.

It is easy to do this if people think it would improve things. I
didn't do it myself because I didn't see how it helped: SET is not in
the standard either. It is just as trivial to use sed/perl/etc. to
remove "SET default_with_oids ..." as it is to remove "WITH/WITHOUT
OIDS".

-Neil