Tom Lane wrote: > Alvaro Herrera <alvhe...@alvh.no-ip.org> writes: > > Thanks, pushed. > > This has broken the selinux regression tests, evidently because it > removed ONLY from the emitted FK test queries. While we could change > the expected results, I would first like to hear a defense of why that > change is a good idea. It seems highly likely to be the wrong thing > for non-partitioned cases.
Yeah, there ain't one, because this was a reversal mistake. I restored that ONLY. (There were two ONLYs in the original query; I initially removed both, and then went over the file and included them conditionally on the table not being a partitioned one, based on review comments. In this line I restored one conditionally but failed to realize I should have been restoring the other unconditionally.) Pushed a fix blind. Let's see if it appeases rhinoceros. -- Álvaro Herrera https://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
