Re: Sigh, we need an initdb

On 2014-06-04 17:03:52 -0400, Tom Lane wrote:
> Actually, that statement makes me realize that if we fix
> PG_CONTROL_VERSION then it's a good idea to *also* do some regular catalog
> changes, or at least bump catversion. Otherwise pg_upgrade won't be able to
> cope with upgrading non-default tablespaces in beta1 installations.

Heh. That's not a particularly nice property, although I can see where
it's coming from. Probably not really problematic because catversion
updates are so much more frequent.

> For the moment I'll just go bump PG_CONTROL_VERSION, assuming that we have
> enough other things on the table that at least one of them will result in
> a catversion bump before beta2.

The slot_name vs slotname thing seems uncontroversial enough since
slot_name is the thing that already appears everywhere in the docs and
it's what we'd agreed upon onlist. It's just that not everything got the
message.

> I have no objection to these as long as we can get some consensus on the
> new names (and personally I don't much care what those are, but I agree
> "xmin" for a user column is a bad idea).

I won't do anything about this one though until we've agreed upon
something.

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services