Ready for PostGIS 3.5.0 - beta1 in about a week?

[Regina Obe]

> "Regina Obe" <lr at pcorp.us> writes:
> 
> > PostgreSQL 17 RC1 will be released soon.  If we want a chance to make
> > the PostgreSQL 17 cut we should release soonish.  By soonish I mean a
> > month or less.
> >
> > PostgreSQL last RC generally stick around for 3-5 weeks before the
> > official is cut, and they usually release one or 2 RCs, usually one
these days.
> >
> > As supported by these stats
> >
> > https://git.postgresql.org/gitweb/?p=postgresql.git;a=tags
> >
> > Is everyone okay with a beta1 coming out say around 9/14 and then
> > followed a week after with an RC1 and then shortly after by release?
> >
> > Or do people want to punt PostgreSQL 17 release and take our sweet
> > time polishing 3.5.0?
> >
> > I created a PostGIS 3.6.0 milestone and have started pushing some
> > things to it.
> 
> You say 'make the PostgreSQL 17 cut' but I don't understand what that even
> means.
> 
> I don't understand why 3.5.0 is linked to 17.  We are about to have a
> 3.4.3 that works with 17, so anyone on a system where they want to update
> pgsql to 17 can also update postgis.

Because many packagers tie each release of PostgreSQL to exactly one version
of PostGIS.
If we don't get out before PG17, then we will be stuck with everyone running
3.4.3 for at least a whole year extra.
Which means we have to support both 3.4.3 for one extra year than we care
to.

> I think it's far more important that when 3.5.0 comes out it is not
troubled,
> that rushing leads to trouble, and that if things don't need to be coupled
it is
> better for them not to be.
> 
It won't be troubled - we've been stewing this thing for more than a year
and honestly
nothing earth shattering has happened.  We've got a lot of cleanup in here,
a lot of performance improvements.
The biggest addition is the sfcgal list of functions and renames.

If we never have a sense of urgency we will never release anything.

> Personally I try not to be bleeding edge on anything, mostly.  I did just
change
> my system from pgsql 12 to 16, since I realized I was on 12 which is now
old,
> and given 17 talk 16 seems like issues would be worked out.  I wouldn't
want
> to update a production system to 17 until 17 had been through several
micros,
> and I can't imagine anyone running in production would want to do that
either.
> 
> So I'm -1 on linkage.

That's you.  I like to have the latest stable version of everything.
As a member of the team, I like more people on my latest stable, cause it
means I don't have to have lots of Cis testing old systems and old versions
of PostgreSQL that are dead.