Heikki Linnakangas <heikki@enterprisedb.com> writes:
> Heikki Linnakangas wrote:
>> Tom Lane wrote:
>>> Anything that involves having VACUUM re-evaluate index expressions is a
>>> nonstarter ... or have you already forgotten the optimizations we put
>>> into 8.2 that assume, eg, no sub-transactions within a VACUUM?
> I think I found it. Is this what you're talking about (in
> commands/vacuum.c):
That's part of it, but I seem to recall other things too --- in
particular, the point about subtransactions troubles me. Whatever you
think about an index function looking at other tables, it is perfectly
legitimate to have an exception block in an index function, and that
requires subtransactions (at least as plpgsql is now implemented).
regards, tom lane