Re: CountDBSubscriptions check in dropdb - Mailing list pgsql-hackers

From Tom Lane
Subject Re: CountDBSubscriptions check in dropdb
Date
Msg-id 23050.1573315694@sss.pgh.pa.us
Whole thread Raw
In response to Re: CountDBSubscriptions check in dropdb  (Amit Kapila <amit.kapila16@gmail.com>)
Responses Re: CountDBSubscriptions check in dropdb  (Amit Kapila <amit.kapila16@gmail.com>)
List pgsql-hackers
Amit Kapila <amit.kapila16@gmail.com> writes:
> On Sat, Nov 9, 2019 at 3:58 PM Peter Eisentraut
> <peter.eisentraut@2ndquadrant.com> wrote:
>> On 2019-11-08 14:38, Amit Kapila wrote:
>>> I am planning to commit and backpatch this till PG10 where it was
>>> introduced on Monday morning (IST).  Pavel agreed that this is a good
>>> change in the other thread where we need it [1].  It is not an urgent
>>> thing, so I can wait if we think this is not a good time to commit
>>> this.  Let me know if anyone has objections?

>> I think the change makes sense for master, but I don't think it should
>> be backpatched.

> Fair enough.  Attached patch with a proposed commit message.

I don't have an opinion on whether it's appropriate to back-patch
this, but I do have an opinion that Monday morning is the worst
possible schedule for committing such a thing.  We are already
past the point where we can expect to get reports from the slowest
buildfarm critters (e.g. Valgrind builds) before Monday's
back-branch wraps.  Anything that is even slightly inessential
should be postponed until after those releases are tagged.

If it's HEAD-only, of course, it's business as usual.

            regards, tom lane



pgsql-hackers by date:

Previous
From: Hans Buschmann
Date:
Subject: AW: Missing constant propagation in planner on hash quals causes joinslowdown
Next
From: Tom Lane
Date:
Subject: int64-timestamp-dependent test vs. --disable-integer-timestamps