Re: Process startup infrastructure is a mess

On 2017-09-15 01:06:54 +0100, Simon Riggs wrote:
> On 14 September 2017 at 22:44, Andres Freund <andres(at)anarazel(dot)de> wrote:
>
> > The way we currently start and initialize individual postgres (sub-)
> > processes is pretty complicated and duplicative. I've a couple
> > complaints:
> ...
> > I think we should seriously consider doing a larger refactoring of this
> > soon. I've some ideas about what to do, but I'd welcome some thoughts
> > on whether others consider this a serious problem or not, and what they
> > think we should do about this, first.
>
> Refactoring without a purpose is a negative for me. It takes time,
> introduces bugs and means the greater code churn over time introduces
> more bugs because fewer people have seen the code. That is arguable,
> but when we compare the priority of that against things people want
> and need there is little contest in my mind.

That's true for code that's not going to be touched anymore, largely bug
free, and can be understood with reasonable effort. Neither is the case
here.

> If we add something to an area then its a good time to refactor it
> since we were going to get bugs anyway.

We've added something to the area on a regular basis. As in last in
v10. The fundamental problem with your argument is that that makes it
impossible for most contributors to get feature in that touch these
parts of the code - many neither have the time nor the knowledge to do
such a refactoring. By your argument we should have rejected logical
replication for v10 because it complicated this further, without
cleaning it up.

Besides that, it also makes it harder to develop new features, not just
to get them in.

Greetings,

Andres Freund