On Wed, Jan 07, 2026 at 08:09:35AM +0900, Michael Paquier wrote:
> Indeed, I'll clean up that. Thanks.
So, I have spent a couple of extra hours on this part, and tweaked
a couple of things before applying the result:
- Removal of atooid8(). I did not use it in the patch set here, and
one could just use strtou64() instead.
- More tests, closing gaps for related to ordering, hash (hash_func,
HashAggregate), btree, compare functions, UINT64_MAX values.
- Some tweaks to the docs.
While the final commit has sticked with oid8 for the data type and
Oid8 for the C type, we still have a couple of months to tweak the
names if necessary. I would keep these as they are, but if there are
more voices for a Oid64, feel free of course. What's on HEAD now
works fine enough for me.
--
Michael
