> On Wed, Apr 08, 2020 at 03:06:39PM +0900, Michael Paquier wrote:
>
> I was considering this stuff again, and I got cold feet regarding the
> fact that we may be able to hit the case of a collation not set if the
> type is collatable depending on how the restrictions now in place for
> partition bound expressions get lifted in the future, so I have
> committed what you sent as v1, after changing the first error message
> to use "partition bound expression" and after adding a comment on top
> of the new checks.
Thanks!