Andres Freund writes:
> On 2018-01-10 13:06:50 -0500, Tom Lane wrote:
>> So I think we should get rid of that micro-optimization, which is
>> probably useless anyway from a performance standpoint, and do the
>> comparison honestly. Any objections?
> No, absolutely none. You're going to change it
Hi,
On 2018-01-10 13:06:50 -0500, Tom Lane wrote:
> So I think we should get rid of that micro-optimization, which is
> probably useless anyway from a performance standpoint, and do the
> comparison honestly. Any objections?
No, absolutely none. You're going to change it?
Greetings,
Andres Fre
While analyzing a recent crash report[1], I noticed that bufmgr.c's
ckpt_buforder_comparator is coded to assume that no two CkptSortItems
could have equal page IDs; it therefore skips the final comparison
and will never return 0 (equal). I do not think that assumption is
correct. I do not see any
