Re: Should contrib modules install .h files?

On 2018-07-31 17:53:19 -0400, Tom Lane wrote:
> Peter Geoghegan <pg(at)bowt(dot)ie> writes:
> > On Tue, Jul 31, 2018 at 1:17 PM, Andres Freund <andres(at)anarazel(dot)de> wrote:
> >> I'm a bit surprised that you decided to push to the 11 branch - the
> >> consensus in this thread seem to have gone the other way by my read?
> >> Given that that's the rule, and pushing non-fixes is the exception, I'm
> >> not particularly ok with just ignoring that?
>
> > +1
>
> By my count of people expressing opinions, we had Michael -1, Stephen -1,
> me -0.1 or so, Alvaro +1, Peter -1, presumably +1 from Andrew; and Andres
> made a comment about not waiting, which perhaps Andrew read as a +1 for
> backpatching. So it's not unreasonable for Andrew to have decided that
> it was about tied. Nonetheless, it does seem like a feature and we're
> past feature freeze, so the default assumption ought to be "no backpatch"
> IMO.

Yea, I don't think it's an entirely unreasonable to decide to backpatch
based on these votes, but I think if the stated opinions are like you
count, it's pretty reasonable to at least announce that the more
controversial choice is the plan and give a chance to more vigorously
object.

Greetings,

Andres Freund