On Sat, 2026-04-18 at 15:11 -0400, Robert Haas wrote:
> Returning to this topic after some time, I have realized that both of
> these rules are inadequate.
I agree with others that perfection isn't required.
I am slightly worried about how hard this will be to document. But if
we solve a good portion of the problem, we can at least document some
best practices, even if the underlying mechanisms are complex.
Regards,
Jeff Davis
