On Thu, Mar 23, 2023 at 09:33:16AM +0100, Drouvot, Bertrand wrote:
> Thanks! LGTM and also do confirm that, with the patch, the JDBC test
> does show the correct results.
How does JDBC test that? Does it have a dependency on
pg_stat_statements?
>
> That said, not having a test (for the reasons you explained
> up-thread) associated with the patch worry me a bit.
Same impression here.
> But, I'm tempted to say that adding new tests could be addressed
> separately though (as this patch looks pretty straightforward).
Even small patches can have gotchas. I think that this should have
tests in-core rather than just depend on JDBC and hope for the best.
Even if \bind does not allow that, we could use an approach similar to
libpq_pipeline, for example, depending on pg_stat_statements for the
validation with a test module in src/test/modules/?
--
Michael
