On Fri, Apr 29, 2016 at 06:43:04AM -0400, Bruce Momjian wrote:
> On Fri, Apr 29, 2016 at 01:20:35PM +0300, Teodor Sigaev wrote:
> > >Doesn't this invalidate tsvector indexes upgraded by pg_upgrade? Should
> > >they be marked as invalid?
> > Directly, it affects on functional indexes i.e. over to_tsvector(). But it
> > affects tsvector column, it should be recreated if it was generated by
> > ts_vector() function.
>
> OK, so every tsvector column or expression index needs to be reported by
> pg_upgrade? Do we want to fix everything else in this same release?
I guess my point is that we should do all pg_upgrade-breaking tsvector
changes in a single release so we don't need to invalidate tsvector
columns and indexes in two releases. If we can't do them all in 9.6,
perhaps we should revert this change and do them all in 9.7.
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ As you are, so once was I. As I am, so you will be. +
+ Ancient Roman grave inscription +
