Re: partitioned indexes and tablespaces

On Wed, Nov 7, 2018 at 2:33 PM Tomas Vondra
<tomas(dot)vondra(at)2ndquadrant(dot)com> wrote:
> FWIW, it was mentioned that "your only concurring vote came from someone
> with whom you share an employer" which kind suggests opinions/votes from
> people working for the same company are somehow less honest/valuable. I
> find that annoying and even insulting, because it kinda hints the
> company (or companies) are pushing people to respond differently than
> they would otherwise. Which I find rather insulting.

It doesn't have to be companies exerting overt pressure on their
employees. It's natural that people are going to have a sympathetic
view of patches written by people with whom they work on a day-to-day
basis. And it's not even a bad thing; having friends and colleagues
whom we trust is good. Still, it's impossible to be be sure that
you're reacting in exactly the same way to a patch by someone you know
and trust as you would to a patch written by a stranger, which is why
when Amit and I recently posted some reviews of zheap-related patches,
we inserted disclaimers into the first paragraph that we are not
independent of those patches and that independent review is desirable.
We did that because *we know* that we may be biased and we want to be
fair about that and on guard against it. I don't think asking other
people to be similarly aware should be annoying or insulting. If your
Mom decided to take up PostgreSQL hacking and posted a patch here, can
you really say you'd evaluate that patch with the exact same
objectivity that you'd apply to a patch written by somebody you'd
never met before? Whether you have a good relationship with your Mom
or a terrible one, that seems like an impossible standard for any
normal human being to meet.

With regard to this patch, I think the new behavior is fine in and of
itself, but I *do not* think it should have been back-patched and I
*do not* think it should work one way for tables and another for
indexes. And, regardless of the technical merits, I strongly object
to things to which multiple people are objecting being jammed into a
back-branch just before a release wraps. That's just not cool.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company