svn commit: r505245 - in head/databases: postgresql10-client postgresql10-contrib postgresql10-pgtcl postgresql10-server postgresql11-client postgresql11-server postgresql94-server postgresql95-cli...

Alexey Dokuchaev danfe at freebsd.org
Wed Jul 3 01:49:12 UTC 2019


On Tue, Jul 02, 2019 at 10:04:35PM +0200, Matthias Fechner wrote:
> Am 30.06.2019 um 11:13 schrieb Alexey Dokuchaev:
> > I tried to see what had changed but could only find version bump from 9.2
> > to 9.6 in commits like these:
> >
> > https://gitlab.com/gitlab-org/gitlab-ce/commit/b609c8f650ca9854d12edbf623f2419c216e73a7
> > https://gitlab.com/gitlab-org/gitlab-ce/commit/e1294022875138b40db7d34a9f33ac9f4a7dc655
> >
> > I'd try to lower those number to 9.5, I'd expect with high probability
> > things would work just as fine.
> 
> if I remember correctly it will break with postgresql < 9.6 if you use a
> more complex structure for groups.

That's why I don't like separate commits in code/docs area, at least when
one commit (documenting 9.2->9.6 requirement) does not reference the actual
code commit which makes use of new features: no one can/wish to pin-point
the exact issue later, just giving those IIRC's.

> You can check the manual on:
> https://gitlab.fechner.net/mfechner/Gitlab-docu
> 
> to see at which version I changed the version requirements.

If you mean this commit:

https://gitlab.fechner.net/mfechner/Gitlab-docu/commit/694fc5875d2d7ac9b285fdd706f318989e633dc6

then I don't see how it explains why 9.5 doesn't cut it, while 9.6 does.

Anyway, I didn't really want to argue about this, I don't use GitLab and
don't care about it (now even more, seeing how well it's being developed
and documented).  I jumped in mostly because bumping default PG version
to 11 in FreeBSD had raised my brows a bit. :-)  Peace,

./danfe


More information about the svn-ports-head mailing list