Re: postgesql-7.3.3 lo_create/lo_open fails

Colm Dobbs <Colm.Dobbs@aepona.com> writes:

>     Oid lobjId = lo_creat(conn, INV_READ|INV_WRITE);
>     if (lobjId != -1)

This coding is wrong --- lo_creat would return 0 (InvalidOid) on
failure, not -1.  As for *why* it's failing, PQerrorMessage might
offer some hint.  I'm a tad surprised by that myself; never heard
of it happening with a connection not already in an error state.

            regards, tom lane