Re: vacuum_truncate configuration parameter and isset_offset

On Wed, Mar 26, 2025 at 11:10 AM David G. Johnston
<david(dot)g(dot)johnston(at)gmail(dot)com> wrote:
> I'm willing to say "I don't know why this is so very important to Nikolay, but I trust him that it is, and since my opinion isn't that strong and this isn't a big deal, so I will accommodate the person screaming that adding this will make their life miserable in a real way." Maybe others need more evidence of what that misery looks like?

What I'm upset about is that it feels to me like Nikolay is trying to
win the argument by yelling. I don't want that to be the way we do
things around here. I admit that sometimes it is, and I think that is
bad, no matter who the yeller is and who is getting yelled at. People
get upset, including me, and that is life, but whether people are
upset should never be the determinant of what goes into the tree. I
fear that this argument will make Nathan - or other committers - more
cautious in committing routine patches for no real reason, and I think
that would be bad for the project. Committing patches frequently
results in a bunch of people getting annoyed at you because of some
oversight, and it's one of the reasons why it takes so long to get a
patch committed around here -- people are risk-averse and afraid of
doing something wrong, so they do nothing. I don't really know how we
can solve that problem, but I definitely do not want to see us start
getting upset with people when they didn't even do anything wrong. And
that's how I see this case. Somebody could have chosen to solve this
problem in a few ways, Nathan picked one of the reasonable ones, and
now we're spending dozens of email messages arguing about whether
Nathan is a terrible person for doing something that looks totally
normal.

In my mind, that is absolutely clear. Nathan is not a terrible person
for that, and I feel rather strongly that saying otherwise is harmful
to the development community and to the project as a whole. I have no
problem with a rational discussion of what the best option is here,
but I am absolutely not OK with vitriolic rhetoric about how things
are awful when, AFAICS, nothing has happened here beyond the totally
routine. In a certain sense, the damage here has already been done.
Nathan has already had to spend a significant amount of time engaging
with this thread over what I think should be a complete non-event, and
will probably have to spend more, and all that takes away from time
that could, for example, be spent reviewing and committing other
patches. And for what?

If it ever happens that the design decision that this patch made
caused a real problem, some future patch could have reverted it and
substituted something else and probably nobody would have cared. Even
now, if some committer other than Nathan cares enough to change
something here, I doubt that Nathan will really care. But I cannot see
any world in which pinning Nathan beyond a barrel and demanding action
is a win for the project overall. If we argued this much about every
design detail of my patches, I would have quit working on this project
long ago.

--
Robert Haas
EDB: http://www.enterprisedb.com