Handling OID Changes in Regression Tests for C Extensions

Dear pg-hackers,

I am facing a challenge related to OID changes in regression tests while
working on a PostgreSQL C extension. Whenever I add or remove a function to
the SQL file, it alters OIDs and leads to cascading modifications in
multiple test files. This creates a significant overhead in maintaining the
regression tests.

I would like to understand if there are best practices or tools used in
PostgreSQL's own regression test suite to handle or mitigate the impact of
OID changes. How does PostgreSQL manage OIDs in its tests to minimize
disruption? Are there strategies, configurations, or patterns that can be
recommended for extensions?

Any guidance or reference to documentation on this topic would be greatly
appreciated.

Best regards,
Narayana