On 29.10.24 07:51, Michael Paquier wrote:
On Tue, Oct 29, 2024 at 07:38:36AM +0100, Peter Eisentraut wrote:
I think it could be useful to set up some better test coverage for various
things overflowing signed integer maximums. For example, maybe you could
hack initdb to advance the OID counter to INT32_MAX+1 or thereabouts and run
the test suites from there. That would also catch things like inappropriate
uses of atoi(), things beyond just the format strings.
Fun. One way to be fancy here would be to force a pg_resetwal
--next-oid in some of the test paths (Cluster.pm and/or pg_regress)
with an environment variable to force the command to trigger across
the board for all the clusters created in the tests. initdb cannot be
used here as the TAP tests reuse a cluster already initdb'd to save
time. No need to touch at pg_regress, either, as we could count on the
pg_regress runs in 002_pg_upgrade.pl and 027_stream_regress.pl.
I was thinking just compiling with a patch like this:
-#define FirstNormalObjectId 16384
+#define FirstNormalObjectId ((Oid) INT_MAX + 1)
Already found one bug: pg_checksums --filenode only accepts files up to
INT_MAX.
