On Wed, Oct 10, 2012 at 08:43:34PM -0400, Tom Lane wrote:
> > It seems to me that the
> > root of the issue here is that people is not that people expect two
> > snapshots -- indeed, a number of people strongly supported getting rid
> > of that behavior at the time -- but rather that they expect the
> > snapshot to be taken after locks are acquired.
>
> Sure. Maybe we can rejigger things in a way that does that, although
> I think the stumbling block is going to be parse-time calls to
> user-defined I/O functions for constants --- which might need a
> snapshot. It might be possible to redesign things so that all tables
> are locked before we do anything that requires a non-SnapshotNow
> snapshot, and then take a single "planning/execution" snapshot. But
> that is not this patch, and would be a lot more invasive than this
> patch, and would certainly not be back-patchable to 9.2.
>
> I think we have to revert and go back to the drawing board on this.
Is reverting going to adversely affect users who are already using the
9.2 behavior?
-- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB
http://enterprisedb.com
+ It's impossible for everything to be true. +