On Tue, 22 Feb 2011 20:02:09 +0100, Xavier Puig <xpuigr(a)gmail.com> wrote:
> No, the installation is new
> but I do it "rm -Rf build" too, and get the same problem.
> Emmmm, when I execute configure I put the absolute path in
> boost-python-libname=/usr/lib/libboost_python-mt-py26.so (/usr/lib/)
> If I don't do this I can't compile:
> /usr/bin/ld: cannot find -llibboost_python-mt-py26.so
> Can it be the cause of the problem?
> I create simbolic links
> in /usr/local/include, /usr/include /usr/local/lib of
> libboost_python-mt-py26.so and don't found it!
-l implies the "lib" bit. Try
i compiled boost and mespy, but when I run:
Traceback (most recent call last):
File "test.py", line 2, in <module>
from meshpy.tet import MeshInfo, build
"/usr/local/lib/python2.6/dist-packages/MeshPy-0.91.2-py2.6-linux-i686.egg/meshpy/tet.py", line 2, in <module>
import meshpy._tetgen as internals
ImportError: /usr/local/lib/python2.6/dist-packages/MeshPy-0.91.2-py2.6-linux-i686.egg/meshpy/_tetgen.so: undefined symbol: _ZNK5boost6python7objects21py_function_impl_base9max_arityEv
What I'm doing wrong?
Thanks in advance
On Mon, 21 Feb 2011 15:33:03 +0500, Andrey Bezrukov <bezrukov.andrey(a)gmail.com> wrote:
> I try to use meshpy in my work to make 2D and 3D interplolations.
> Unfortunately, there is not enough documentation on the web, and mailing
> list is not accessible. Is it possible to hear from you ore some one
> else some explanations?
> this is my conceptual code:
> mesh_info = MeshInfo()
> print 'Building mesh'
> mesh = build(mesh_info, max_volume = 10000, allow_boundary_steiner = True) #
> points_out = 
> for p in mesh.points:
> My problem is, that I want to know exactly how initial set of points
> corresponds to output points_out. In most cases there is exact connection:
> points_out[0:N] == points_xy[0:N],
> where N = len(points_xy).
> But sometimes an extra point appears, for example K, such that
> points_out[0:K] == points_xy[0:K] and points_out[K+1:N+1] == points_xy[K:N].
> What is the right way to identify points?
point_markers is the right tool. They're vertex numbers that persist
through the meshing process.
> In documentation it is not
> described, what are point_markers should be used for. Attempt to use
> point_markers to identify points fails, the program tels me that
> point_markers array is not allocated.
Call mesh_obj.point_markers.setup(). (Please let me know if that
helped--if so, this needs to be added to set_points() in meshpy.common.)
Hope this helps,