Re: pg_upgrade from 8.3.4 issue

[Justin Arnold <jharnold81@xxxxxxxxx>]

Thanks Tom and Merlin, I removed that logic from check.c, rebuilt, and it worked fine.

On Tue, Aug 30, 2011 at 2:47 PM, Tom Lane <tgl@xxxxxxxxxxxxx> wrote:

Merlin Moncure <mmoncure@xxxxxxxxx> writes:
> It looks like some time after 8.3 was released that function was
> changed from returning 'record'.  This is making me wonder if the
> upgrade process was ever tested/verified on 8.3.

Not lately, apparently :-(

> I absolutely do not
> advise doing this without taking a lot of precautions, but you might
> force your way past that step with:

I think it'd be a lot safer to modify (or just remove) the test in
pg_upgrade.  It looks like a one-liner:

   prep_status("Checking for prepared transactions");

   res = executeQueryOrDie(conn,
                           "SELECT * "
                           "FROM pg_catalog.pg_prepared_xact()");

   if (PQntuples(res) != 0)
       pg_log(PG_FATAL, "The %s cluster contains prepared transactions\n",
              CLUSTER_NAME(cluster));

There's no reason at all for this code to not use the published API,
which is the pg_prepared_xacts system view.

                       regards, tom lane