Re: Documentation improvements for partitioning

On Fri, Feb 24, 2017 at 8:06 AM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> Simon, this is ridiculous. We're 4 or 5 days away from the start of
> the last CommitFest. We have time to fix bugs and improve
> documentation and maybe tweak things here and there, but 3 and 4 are
> significant development projects. There isn't time to do that stuff
> right now and get it right. You don't get to show up more than two
> months after the feature is committed and start complaining that it
> doesn't include all the things you want. Which things ought to be
> included in the initial patch was under active discussion about a year
> ago this time.

I take that back. You can complain as much as you like; everybody has
a right to complain. But it's not reasonable to expect Amit (or
anyone else) to go fix the things you're complaining about in time for
v10, or really ever. He doesn't have to write any more partitioning
patches ever, and if he does decide to do so, he doesn't have to write
the ones you or I or anyone other than his employer wants him to write
(and he only has to listen to his employer if he doesn't want to get
fired).

Also, I'm very much against against any major tinkering with this
feature in this release. We've got enough work to do stabilizing
what's already been committed in this area, and the last things we
need is a bunch of patches that significant change it showing up at
the eleventh hour without time for adequate reflection and discussion.
Most if not all significant patches for this release should already
have been submitted; again, the last CommitFest will be starting
shortly, and we should have seen those patches in the previous
CommitFest. We should be focusing on getting all the good patches
that have already been written committed, not creating new ones at the
last minute.

Contrary to what you may think, neither changing the way partition
pruning works nor inventing a system for indexes to roll down to
partition children is a minor fix. Even if you restrict the scope to
simple cases, there's still got to be a level of design agreement so
that we know we're not boxing ourselves into a corner for the future,
and the patch quality still has to be good. That's not going to
happen in the next couple of days, barring a dramatic reversal of how
the development process in this community has always worked before.

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