On Wed, Aug 1, 2018 at 7:46 AM, Etsuro Fujita <fujita.ets...@lab.ntt.co.jp> wrote: > I posted the updated patch [1].
Etsuro-san: I really think we should just go with what Ashutosh had proposed. Tom, Ashutosh, and I all seem to agree that we shouldn't try to re-jigger things at create-plan time. I think that a 3-1 consensus against your proposal is sufficient to say we shouldn't go that way. Now, here you've got a new approach which avoids that, which I have not yet reviewed. I'll try to spend some time on it this afternoon, but really, I think it's too late for this. This bug was reported in February, and we're supposed to be pushing 11 final out the door in not much more than a month. Proposing a new approach in August is not good. Can't we just do what Ashutosh proposed for now and revisit this for v12? -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
