[postgis-devel] a slight error in postgis TopologySummary

[Sandro Santilli]

On Thu, Sep 26, 2013 at 05:32:30PM +0200, Rémi Cura wrote:
> OK,
> thanks !
> Now I understand.
> I'm placing all my hope in this 2.1 version, as I can't work with the
> current version.

Let me know. 

Note that there's no way you can get code with no possibility of hitting
robustness issues. It's just not available with finite precision.
Not implemented in CGAL, not implemented in boost::geometry.
It currently doesn't exist, so don't look for it.

Instead, organize your workflow to expect exception and react to them.
The good thing about persistent topology is that you could
loose nothing and work on the specific problematic geometries.

--strk;