Re: Collation-aware comparisons in GIN opclasses

On Mon, Sep 15, 2014 at 11:45 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Peter Geoghegan <pg(at)heroku(dot)com> writes:
> > On Mon, Sep 15, 2014 at 8:28 AM, Alexander Korotkov
> > <aekorotkov(at)gmail(dot)com> wrote:
> >> Rename such opclasses and make them not default.
> >> Create new default opclasses with bitwise comparison functions.
> >> Write recommendation to re-create indexes with default opclasses into
> >> documentation.
>
> > I certainly think this should be fixed if at all possible, but I'm not
> > sure about this plan. Can we really rename an opclass without
> > consequence, including having that respected across pg_upgrade?
>
> No. And we don't know how to change the default opclass without
> breaking things, either. See previous discussions about how we
> might fix the totally-broken default gist opclass that btree_gist
> creates for the inet type [1]. The motivation for getting rid of that
> is *way* stronger than "it might be slow", but there's no apparent
> way to make something else be the default without creating havoc.
>

I've read thread about gist opclass for inet type. But that case is more
difficult because conflict is between builtin opclass and contrib opclass.
This case seems to be much simpler: we need to change builtin opclass to
builtin opclass and contrib opclass to contrib opclass. I realized that
it's problematic to rename builtin opclass due to pg_upgrade. However, it
seems still possible to create new opclass and make it default.

------
With best regards,
Alexander Korotkov.