On Jul 14, 2008, at 07:26, Tom Lane wrote:
>> I'd like to keep these tests, since they ensure not just that the
>> functions work but that they work with citext.
>
> It might be reasonable to test a couple of them for that purpose.
> If your agenda is "test every function in the system that comes or
> might come in a bpchar variant", I think that's pointless.
Or a varchar variant, or where such a variant might be added in the
future. To my mind, it's important to have good coverage in my unit
tests to ensure that things continue to work exactly the same over time.
So, since the tests are already written, and are unlikely to add more
than a few milliseconds to test runtime, can you at least agree that
such tests are harmless?
Updated patch later today.
Thanks,
David
