Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> writes:
> ! 	if (!(tuple->t_infomask & HEAP_XMAX_INVALID))
>   	{
> ! 		if (!(tuple->t_infomask & HEAP_XMAX_IS_MULTI))

How about just one test,

	if (!(tuple->t_infomask & (HEAP_XMAX_INVALID | HEAP_XMAX_IS_MULTI)))

But other than that quibble, yeah, it's a bug.  XMAX_INVALID means just
that: the xmax is not to be thought valid.

			regards, tom lane

In response to

• heap_tuple_needs_freeze false positive at 2012-02-02 01:01:27 from Alvaro Herrera

pgsql-hackers by date

Next:	From: Alvaro Herrera	Date: 2012-02-02 04:33:13
	Subject: freezing multixacts
Previous:	From: Tom Lane	Date: 2012-02-02 03:57:48
	Subject: Re: spgist text_ops and LIKE