Re: Refactoring speculative insertion with unique indexes a little

On Mon, Jan 18, 2016 at 8:36 AM, Alvaro Herrera
<alvherre(at)2ndquadrant(dot)com> wrote:
> If you refuse to post an updated version of the patch until Heikki
> weighs in some more, and given that Heikki has (for the purposes of this
> patch) completely vanished, I think we should mark this rejected.

I don't refuse. I just don't want to waste anyone's time. I will
follow all of Heikki's feedback immediately, except this:

"I think it'd be better to define it as "like CHECK_UNIQUE_YES, but
return FALSE instead of throwing an error on conflict". The difference
is that the aminsert would not be allowed to return FALSE when there
is no conflict".

That's because I believe this is quite broken, as already pointed out.

> If somebody else is open to reviewing the patch, I think that'd be
> another way to move forward, but presumably they would start from a
> version with the discussed changes already fixed. Otherwise it's a
> waste of time.

Your premise here is that what Heikki said in passing months ago is
incontrovertibly the right approach. That's ridiculous. I think Heikki
and I could work this out quite quickly, if he engaged, but for
whatever reason he appears unable to. I doubt that Heikki thinks that
about what he said, so why do you?

The point about CHECK_UNIQUE_YES I highlighted above felt like a
temporary misunderstanding to me, and not even what you might call a
real disagreement. It wasn't as if Heikki was insistent at the time. I
pointed out that what he said was broken according to an established
definition of broken (it would result in unprincipled deadlocks). He
didn't respond to that point. I think he didn't get back quickly in
part because I gave him something to think about.

If any other committer wants to engage with me on this, then I will of
course work with them. But that will not be predicated on my first
revising the patch in a way that this other committer does not
understand. That would be profoundly unfair.

--
Peter Geoghegan