Re: Fixing a couple of buglets in how VACUUM sets visibility map bits

On Sat, Dec 31, 2022 at 4:53 PM Peter Geoghegan <pg(at)bowt(dot)ie> wrote:
> The first patch makes sure that the snapshotConflictHorizon cutoff
> (XID cutoff for recovery conflicts) is never a special XID, unless
> that XID is InvalidTransactionId, which is interpreted as a
> snapshotConflictHorizon value that will never need a recovery conflict
> (per the general convention for snapshotConflictHorizon values
> explained above ResolveRecoveryConflictWithSnapshot).

Pushed this just now.

Attached is another very simple refactoring patch for vacuumlazy.c. It
makes vacuumlazy.c save the result of get_database_name() in vacrel,
which matches what we already do with things like
get_namespace_name().

Would be helpful if I could get a +1 on
v1-0002-Never-just-set-the-all-frozen-bit-in-VM.patch, which is
somewhat more substantial than the others.

--
Peter Geoghegan