Hi,
I've been using CGAL to mesh (approximations to) fattened gyroid surfaces, confined to a box from an implicit function. Attached is a simple example code. Unfortunately, this results in a surface mesh containing a nonmanifold edge (with 4 faces attached to it). this problem seems to persist in the tet mesh that was generated (I wrote a little program to recalculate its surface mesh). Am I making some silly mistake here? Is there any way to fix it? Thanks, Patrick 3d_make_mesh.cpp (8K) Download Attachment 
Le Sunday, November 20, 2016 7:35:56 PM CET Patrick a écrit : > Hi, > > I've been using CGAL to mesh (approximations to) fattened gyroid surfaces, > confined to a box from an implicit function. Attached is a simple example > code. Unfortunately, this results in a surface mesh containing a > nonmanifold edge (with 4 faces attached to it). this problem seems to > persist in the tet mesh that was generated (I wrote a little program to > recalculate its surface mesh). Am I making some silly mistake here? Is > there any way to fix it?
There is an undocumented parameter that you may want to try... Add `facet_topology=CGAL::MANIFOLD,` in the parameters of `Mesh_criteria_3`, to enforce the manifoldness of the output. However, if mesh domain has not a manifold surface, the execution may loop and eventually crash. Use it carefully.
 Laurent Rineau, PhD R&D Engineer at GeometryFactory http://www.geometryfactory.com/ Release Manager of the CGAL Project http://www.cgal.org/

On Mon, Nov 21, 2016 at 11:57 AM, Laurent Rineau (CGAL/GeometryFactory) <[hidden email]> wrote:
Thanks for the advice, however, it seems that the nonmanifoldness of the surface mesh is introduced somewhere in the optimization steps. Adding the manifoldparameter alone didn't help, but adding it together with "no_lloyd(), no_odt(), no_perturb(), no_exude()" produced a proper manifold mesh. Just exude alone also seems to be fine, still. Best, Patrick 
Dear all,
I observed that: lloyd() and odt() introduces nonmanifold edges, however perturb() and exude() worked fine for me when CGAL::MANIFOLD is used. I notice also that the mesher tend to have issues with intel TBB when CGAL::MANIFOLD is used. Best, Adrien 
Le Wednesday, January 11, 2017 2:20:02 PM CET aleph a ?crit :
> Dear all, > > I observed that: lloyd() and odt() introduces nonmanifold edges, however > perturb() and exude() worked fine for me when CGAL::MANIFOLD is used. That's true. Both Lloyd and ODT optimiser are global optimizers, that do not check the topology of the resulting surface. > I notice also that the mesher tend to have issues with intel TBB when > CGAL::MANIFOLD is used. Indeed. That is a known issue, and the main reason we have not yet officialized the MANIFOLD option.  Laurent Rineau, PhD R&D Engineer at GeometryFactory http://www.geometryfactory.com/ Release Manager of the CGAL Project http://www.cgal.org/  You are currently subscribed to cgaldiscuss. To unsubscribe or access the archives, go to https://sympa.inria.fr/sympa/info/cgaldiscuss 
Free forum by Nabble  Edit this page 