CGAL users discussion list

[Giles Puckett <>]

Thank you for the insight. It was a loss of precision problem this end - 
coordinate transformations have pulled the two boxes apart slightly, 
most likely - if I take more care with precision they stay together and 
the problem goes away. Longer term solution for robustness may be to 
make sure they penetrate each other by a small epsilon.

Merci,

G.

On 8/06/2020 11:51 pm, "Sebastien Loriot (GeometryFactory)" 
( via cgal-discuss Mailing List) wrote:
> If the output of the union is non-manifold and the function is returning
> false to indicate it.
> The two boxes are sharing only an edge so the union will be non-manifold
> and cannot be represented in a halfedge data structure (without 
> creating self-intersections).
>
> There is some on-going work to handle those operations with 
> self-intersections but nothing is publicly available for now.
>
> Best regards,
>
> Sebastien.
>
>
>
>
> On 6/8/20 11:40 AM, Giles Puckett wrote:
> > Hello all,
> >
> > I have caught PMP (CGAL 5.0) failing to do a corefinement and union, 
> > after creating meshes out of simple shapes (cylinders and rect 
> > prisms). No self-intersection exception is thrown.
> >
> > I have reproduced the problem by writing OFF files in double 
> > precision. A test program and files are here:
> >
> > https://www.dropbox.com/sh/7sv3p0yxomsjrxn/AAC_xP_HH9EoMFajZSpK4m71a?dl=0 
> >
> >
> > I built on Windows 10 using Visual Studio 2019 (tool set 1.42). The 
> > source is a modification of one of the PMP examples and it builds 
> > under that environment. CGAL is at 5.0. Let me know if you want me to 
> > raise an issue.
> >
> > G.