[postgis-devel] PostGIS 2.5 what should be minimum requirements?

[Regina Obe]

I like that idea.  Try to convince strk of it.

 

He's always the hardest to budge and it's always pulling teeth to get him to let go of supporting older versions of things.

 

For PostgreSQL I really only wanted to ever support 3 or 4 versions of PostgreSQL on any release of PostGIS and now we are using more PostgreSQL machinery in our code, each version always breaks something.  So I'm tempted to say we shouldn't support any more than 3 versions of PostgreSQL in any release.

 

From: postgis-devel [mailto:postgis-devel-bounces at lists.osgeo.org] On Behalf Of Daniel Baston
Sent: Friday, September 29, 2017 10:55 AM
To: PostGIS Development Discussion <postgis-devel at lists.osgeo.org>
Subject: Re: [postgis-devel] PostGIS 2.5 what should be minimum requirements?

 

This is tangential to this specific proposal, but I am wondering whether we should establish some general rules for how long to support old versions of dependencies.

 

For example, we could decide to support <dependency n> in new versions of PostGIS until <timespan> after <dependency n+1> is released.  As an example, we could decide to support GEOS 3.4 in new versions of PostGIS until 2 years after GEOS 3.5 is released.

 

Or alternatively, we could say that we support the latest released version of <dependency>, plus any versions released within the past <timespan>.  So we could decide to support GEOS 3.6, plus any versions of GEOS released in the past three years (3.4 dates back to 2013, so this would limit it to 3.5 and 3.6).

 

I'm just throwing these out for the sake of examples, but I think some guidelines along these lines could help make these decisions simpler and more predictable for users/packagers.

 

Dan

 

On Fri, Sep 29, 2017 at 10:40 AM, Regina Obe <lr at pcorp.us <mailto:lr at pcorp.us> > wrote:

While we are still in infancy of PostGIS 2.5 (and probably at least a year
away from release), I'd like to get requirements out of the way and propose
the following:

1) Drop support for PostgreSQL 9.4 and 9.5  (so PostgreSQL 9.6 - PostgreSQL
11 will be supported)
9.4 has the main annoyance of not supporting true KNN and I'm tired of
explaining this to folks and having people with pg_upgrading from these
lowers being screwed when upgrading because they can no longer do

ALTER EXTENSION ...  cause they are already at 2.5.0

9.5 doesn't support Parallelism and I suspect we may need to do some
restructuring for some aggs in 2.5, I'd just assume not have to make special
concessions for folks trying to pg_upgrade from a PostgreSQL 9.5 2.5
extension to 9.6+

2) Make GEOS 3.5+ the minimum (for 2.4  GEOS 3.4 was the minimum)  - we've
got too much stuff that requires 3.5 already that is turned off for lower
and newer GEOS has some robustness improvements.

3) Make Proj 4.9+ the minimum currently our minimum is proj 4.7 I think.  As
discussed it's on the table to redo some of the geography stuff like
ST_Segmentize using proj 4.9+ features.


Anyone have issues with the above?

PSC folks, if you are okay with all the above please give a +1

If you are okay with some and not others, I can break apart so we can at
least decide on some pieces.


Thanks,
Regina

_______________________________________________
postgis-devel mailing list
postgis-devel at lists.osgeo.org <mailto:postgis-devel at lists.osgeo.org> 
https://lists.osgeo.org/mailman/listinfo/postgis-devel

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-devel/attachments/20170929/c1fa3402/attachment.html>