Re: BUG #14057: vacuum setting reltuples=0 for tables with >0 tuples

On 2017-03-16 17:28:24 -0400, Tom Lane wrote:
> Andrew Gierth <andrew(at)tao11(dot)riddles(dot)org(dot)uk> writes:
> > "Andres" == Andres Freund <andres(at)anarazel(dot)de> writes:
> > Andres> Seems like a good thing to include in the tree. I'd be ok with
> > Andres> just including the simpler version in the relevant branches.
>
> > Ok.
>
> I dunno ... there doesn't seem to be any meaningful portability risk
> here, and it's not clear to me what class of future bug this test might
> hope to catch. Do we really need to spend test cycles forevermore on
> this?

We had previous bugs around this, so it doesn't seem like a bad idea to
test it. Also it should be so short in comparison to the rest of the
isolationtests that it won't matter wrt total runtime?

- Andres