On 2015-06-04 20:41:46 +0200, Andres Freund wrote:
> > 1. 15 years of the same argument (current source: pg_audit)
>
> I don't see getting rid of contrib helping with that. The only change
> will then be whether something should be in core.
>
> And there's stuff that's just very hard to develop out of core; but
> doesn't necessarily immediately belong into
> core. E.g. pg_stat_statements is relatively closely tied to things in
> core.
An even better example is pg_upgrade. It'd not have been possible to
bring it somewhere close to maturity without having it in contrib/. It
requires core cooperation, and we weren't immediately to declare it as
part of core.
