On Sat, Jul 13, 2019 at 11:17 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> If the spec is allowing this because it can be seen
> to be safe, why should we not allow other cases that the user has
> taken the trouble to prove to themselves are safe? (If their proof is
> wrong, well, it wouldn't be the first bug in anyone's SQL application.)
Well said.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company