CGAL users discussion list

[Mael <>]

Hello,

You may in theory have self-intersections in your input, but the result 
is not guaranteed as too many constraints might make it impossible to 
find a triangulation.

There seems to be some self-intersections (at vertices nonetheless) in 
your input; what result do you expect at those locations: should the 
tiny hole and the big hole be filled independently, or something else? 
For this reason, it is best to clean your input or just split your 
polyline into multiple polylines if you have these self-intersections.

Best,
Mael

On 12/06/2019 19:17, danielyan86129 wrote:
> Dear CGAL,
>
> I have a hole boundary represented as a polyline L = {v_1, ..., v_n}, v_i
> being index into some mesh structure. I've verified this is a simple polygon
> because every vertex is of degree 2. The 3d coordinates of the vertices are
> denoted as P = {p_1, ..., p_n}. I'm using triangulate_hole_polyline(P, F) to
> fill holes represented as polyline where F denotes the output patch faces
> (each being a triple of indices into P). However this polyline cannot be
> triangulated by the API, i.e. F is empty. Here is data V:
>
> https://drive.google.com/open?id=1zYT3wisdWXBatkARTJq1YzzG_eN_zaP5
>
> What could go wrong? Would the API fail if some line segments are
> geometrically intersecting?
>
>
>
> --
> Sent from: http://cgal-discuss.949826.n4.nabble.com/