Re: Inaccuracy in VACUUM's tuple count estimates

Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> The reason this stuff is not too carefully spec'd is that when
> HTSV was written, there was no expectation that there was any
> correctness issue around which of these cases was returned.  I
> wonder whether SSI should be using HTSV at all.

That's certainly a reasonable question.  When I was writing the SSI
code, I was blissfully unaware of HTSV and had coded up a way to
check this which seemed to work for all tests we had.  Jeff Davis,
reviewing the code, was concerned that such separate code was more
likely to miss something or to break as visibility handling
changed.  He argued that HTSV was basically checking for the same
things I was, and a redundant and version which did the check
differently was a bad idea.  Here is where it was discussed during
development:

http://www.postgresql.org/message-id/flat/1296499247(dot)11513(dot)777(dot)camel(at)jdavis#1296499247(dot)11513(dot)777(dot)camel@jdavis

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company