On Tue, Mar 5, 2019 at 5:35 PM Andrew Dunstan <andrew.duns...@2ndquadrant.com> wrote: > OK, I think we have agreement on Tom's patch. Do we want to backpatch > it? It's a change in behaviour, but I find it hard to believe anyone > relies on the existence of these annoying messages, so my vote would be > to backpatch it.
I don't think it's a bug fix, so I don't think it should be back-patched. I think trying to guess which behavior changes are likely to bother users is an unwise strategy -- it's very hard to know what will actually bother people, and it's very easy to let one's own desire to get a fix out the door lead to an unduly rosy view of the situation. Plus, all patches carry some risk, because all developers make mistakes; the fewer things we back-patch, the fewer regressions we'll introduce. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company