David E. Wheeler wrote:
> On Aug 6, 2009, at 7:28 AM, Tom Lane wrote:
>
> > That would cover the problem for OIDs needed during CREATE TABLE, but
> > what about types and enum values?
>
> I haven't been following this discussion very closely, but wanted to
> ask: is someone writing regression tests for these cases that
> pg_migrator keeps bumping into?
Yes, I have regression tests I run but they are not in CVS, partly
because they are tied to other scripts I have to manage server settings.
Here are my scripts:
http://momjian.us/tmp/pg_migrator_test.tgz
One big problem is that pg_migrator fails as soon as it hits one of
these so there isn't much to automate.
-- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB
http://enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +