[postgis-devel] Can we put back GEOS 3.5 support in 3.0?

[Sandro Santilli]

On Wed, Jan 23, 2019 at 04:44:47PM +0300, Darafei "Komяpa" Praliaskouski wrote:

> If we, PostGIS developers, are affected by the same plague, we basically
> can't ever close a ticket in PostGIS GEOS milestone

My policy has always been to _close_ a "PostGIS GEOS" ticket as soon
as an official GEOS release was out fixing the GEOS problem. Leaving
tests failing when the GEOS version is too old would be fine with me.
It tells builders that they need to upgrade to a newer GEOS if they
want a knowing functional PostGIS.

> A recent case is someone showing on PostGIS IRC asking why
> ST_Subdivide won't work as expected in their 2.4.latest and the issue was
> they used unpatched GEOS.

Did they build PostGIS themselves ? And GEOS ?
Did they (or their packagers) run the testsuite ?

> We didn't have a release of GEOS for a lot of time for the same reason:
> nobody cares about GEOS. Either we start caring, or we should drop
> supporting it as mandatory thing and rewrite its algorithms into liblwgeom,
> as it's at least supported.

I'm happy with making GEOS optional, but of course even happier if
more people will raise the love for GEOS. I see lots of activity
from Paul Ramsey, Daniel Baston, Regina Obe and some other occasional
contributor recently in there, so I suspect "nobody cares" is not
actually true.

--strk;