[Gmsh] two bugs
cgeuzaine at ulg.ac.be
Tue Oct 7 21:06:25 CEST 2008
John Doe wrote:
> Hi Christophe,
> Thanks for your response. I am rebuilding the stable source with
> Tetgen right now.
> Just for my understanding, both Netgen and Tetgen produce a Delaunay
> tetrahedralization? In my finite volume code, do I need to look out
> for elements where the Delaunay condition is not met?
You'll probably mostly want to avoid bad quality elements. Always apply
Optimize and/or Optimize (Netgen) in that case.
> On Fri, Oct 3, 2008 at 1:19 AM, Christophe Geuzaine <cgeuzaine at ulg.ac.be> wrote:
>> John Doe wrote:
>>> According to the web page, bugs should be filed on the mailing list.
>>> I have two.
>>> Bug #1:
>>> If I create a 3d box with corners 1e-5 apart with 5e-7 characteristic
>>> length at each corner, the program crashes. If I do the same with 1
>>> and 0.05. This is with the stable source code version
>> Hi John - it's a known problem with the 3D Netgen algorithm. Using
>> Tetgen+Delaunay should solve this.
>> (Try first with a binary from the website. To build from sources, you need
>> to download tetgen from the web and copy the .cxx and .h files into
>>> Bug #2:
>>> Compilation of the stable source code fails if GL/gl.h is missing.
>>> The configure script misses this on linux.
>> We rely on fltk's "fltk-config" script to detect OpenGL on Linux. We should
>> maybe include additional tests...
>>> gmsh mailing list
>>> gmsh at geuz.org
>> Prof. Christophe Geuzaine
>> University of Liege, Electrical Engineering and Computer Science
Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science