CGAL users discussion list

[Bryant <>]

When taking the difference between a /simple/ Polygon_set_2 and a /simple/
Polygon_2, the result may be non-simple. This appears to occur when one
would expect the difference to result in two polygons with holes that share
a point.

I use the following definitions:
- A Polygon_set_2 is simple if the Polygon_with_holes_2s composing it are
simple.
- A Polygon_with_holes_2 is simple if its outer boundary and holes are
simple.

Here's a minimum working example:
        CGAL::Polygon_2<K> poly1;
        poly1.push_back(CGAL::Point_2<K>(0, 0));
        poly1.push_back(CGAL::Point_2<K>(2, 0));
        poly1.push_back(CGAL::Point_2<K>(2, 1));
        poly1.push_back(CGAL::Point_2<K>(1, 1));
        poly1.push_back(CGAL::Point_2<K>(1, 2));
        poly1.push_back(CGAL::Point_2<K>(0, 2));
        CGAL::Polygon_set_2<K> poly_set(poly1);

        CGAL::Polygon_2<K> hole;
        hole.push_back(CGAL::Point_2<K>(0, 0));
        hole.push_back(CGAL::Point_2<K>(1, 0));
        hole.push_back(CGAL::Point_2<K>(1, 1));
        hole.push_back(CGAL::Point_2<K>(0, 1));

        poly_set.difference(hole);

Note: I have only tested with the EPEC Kernel.

The above code results in poly_set containing the following polygon:
(0, 1), (1, 1), (1, 0), (2, 0), (2, 1), (1, 1), (1, 2), (0, 2)

I would expect it to contain the following polygons:
(1, 0), (2, 0), (2, 1), (1, 1) 
        and
(0, 1), (1, 1), (1, 2), (0, 2)




--
Sent from: http://cgal-discuss.949826.n4.nabble.com/