Re: Inconsistency between try_mergejoin_path and create_mergejoin_plan

On Tue, Sep 3, 2024 at 5:51 PM Richard Guo <guofenglinux(at)gmail(dot)com> wrote:
> The new test case fails starting from adf97c156, and we have to
> install a hash opfamily and a hash function for the hacked int8alias1
> type to make the test case work again.
>
> Now, I'm more dubious about whether we really need to add a test case
> for this change.

I pushed this patch with the test case remaining, as it adds only a
minimal number of test cycles. I explained in the commit message why
the test case is included in equivclass.sql rather than in join.sql.

I did not do backpatch because this bug cannot be reproduced without
installing an incomplete opclass, which is unlikely to happen in
practice.

Thanks for the report and review.

Thanks
Richard