Re: pgsql default version

[Greg Troxel <gdt%lexort.com@localhost>]

nia <nia%NetBSD.org@localhost> writes:

> ping?

(I don't really understand how much we care about having a default that
some packages don't build with.  This is easy to adjust for building
from source, and running postgresql seriously but not wanting to build
packages to me doesn't really go together.  But I think the default
shouldn'tq exclude lots of things.)


Last I heard, where we were was:

  Taylor proposed moving to a pretty recent version.

  I pointed out that lots of things don't build with that.

  I pointed out that we were too close to freeze.

So now we are comfortably away from freeze and some things are updated.

I reran my script (after updating to 95 as base) on the 9.0 aarch64
2020Q2 packages, and:

  MISSING 10 postgresql10-pgroonga

  MISSING 11 collectd-postgresql11
  MISSING 11 postgresql11
  MISSING 11 postgresql11-client
  MISSING 11 postgresql11-contrib
  MISSING 11 postgresql11-pg_repack
  MISSING 11 postgresql11-pgpool2
  MISSING 11 postgresql11-pgroonga
  MISSING 11 postgresql11-plperl
  MISSING 11 postgresql11-plpython
  MISSING 11 postgresql11-pltcl
  MISSING 11 postgresql11-redislog
  MISSING 11 postgresql11-repmgr
  MISSING 11 postgresql11-server
  MISSING 11 proftpd-postgresql11-1.3.6c.tgz

  MISSING 12 postgresql12-pg_repack
  MISSING 12 postgresql12-pgpool2
  MISSING 12 postgresql12-pgroonga

and it seems that postgresql11 did not even build.



So I am in favor of:

  flip to 10 immediately (I can, or you are welcome to)

  hope somebody updates pgroonga

  maybe fix 11 for aarch64 or see if that's real

  somebody take a look at pg_repacka nd pgpool2, and update if we're
  overdue, poke upstream, and basically give up if this doesn't work

  after seeing how the updates in previous item play out, flip to 11 or
  12 before 8/15 to be comfortabley before Q3


What do you think we should do?