Noah Misch <noah@leadboat.com> wrote:
> Incidentally, I contend that we should write frozen tuples to
> new/truncated tables unconditionally.
+1
> The current behavior of making old snapshots see the table as
> empty violates atomicity at least as badly as letting those
> snapshots see the future-snapshot contents.
Right, there was no point where the table existed as empty at the
end of a transaction, so it is quite broken as things stand now.
> But Marti has a sound proposal that would interact with your
> efforts here to avoid violating atomicity at all
Well, getting it right is certainly better than moving from a slow
non-conforming behavior to a fast non-conforming behavior. ;-)
-Kevin