Tom Lane wrote:
Patch committed, with most of these fixed. On this item:
> AtSubCommit_Snapshot can leave us with multiple RegdSnapshotElt's for
> the same snap and s_level, which seems a bit bogus.
I agree it is bogus. I punted though, and left the code as-is, with a
comment stating the problem. We can revisit this problem later if it's
ever an issue.
Many thanks for the extensive help.
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support