Re: MultiXact truncation, startup et al.

Andres Freund escribió:
> Hi,
>
> Thanks, looks saner than my version ;)

Here are my versions of these patches.  One thing not yet addressed in
these patches is the change to HeapTupleGetUpdateXid() that has been
mentioned repeatedly in these threads (we'll work on that next).  Note I
still have the FIXME comments in the freeze patch previously mentioned.

I added some comments about the output parameters of
vacuum_set_xid_limits, per Robert's complaint.  Please give that a look.

--
Álvaro Herrera                http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services