Robert Haas <robertmhaas@gmail.com> writes:
> On Wed, Apr 25, 2012 at 2:48 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Would it be reasonable to shoot for wrapping 9.2beta1 next week?
> I suspect that's a bit aggressive. We have a couple of issues that I
> think we should fix before beta1:
> - the pg_stat_statements cleanups we were discussing on the other
> thread, since that's going to involve changing column names/types
> - the Hot Standby vs. index-only scans stuff, which needs a WAL format
> change, and is my top priority as soon as I get unburied
> - SP-GiST is unsafe for use by hot standby slaves, since I think
> queries returning wrong answers = bad
> - http://archives.postgresql.org/message-id/1333124720-sup-6193@alvh.no-ip.org
I agree that fixing the first two before beta is a necessary thing.
I'm not convinced the other two are beta blockers. We should not have
a mindset of "there must be no known bugs in beta1"; we want to
parallelize not serialize this process, so that testing has to be able
to happen concurrently with fixing non-blocker bugs. I'd say that
anything that isn't going to require an initdb to fix, and that most
testers are unlikely to hit, ought not be a blocker.
> I'm not sure we can commit to a timeline for beta until we know who is
> taking care of each of those things.
I already promised to deal with the first one, and you are indicating
willingness to deal with the second, so I think we have the blockers
covered.
regards, tom lane