Hi Barend,

Thank you very much for your detailed reply.

 

> In the output duplicate points are skipped too. Even if the input does not contain duplicate points, the raw  output might in some configurations contain duplicate points, and they are explicitly skipped (however, this is not checked). This is done in a later phase of the library, at first they were just included.

 

Not sure what exactly that means. So, when I call, e.g., boost::geometry::difference(poly1, poly2, polyOut) – can any assumptions be made about the presence of duplicate points in polyOut?

 

> You can use boost::geometry::unique(polygon). But this is not required, the algorithms we provide should handle them correctly. However it is always possible that you at one time need to get rid of them yourself.

Follow-up question: In case of an open (implicitly closed) representation, does boost::geometry::unique(polygon) remove the last point if it is equal to the first? Why (not)?

 

Regards

   Volker