segfault again, meshpy.triangle before and after numpy.

Hi, in the following script, if I put `import meshpy.triangle as triangle` before `import numpy as np`, everything goes fine. However, if I put it after numpy, there will be a segfault.

the versions are

```
ipython-2.3.1
python-2.7.8
numpy-1.9.1
```

and using the meshpy-git version.

PS, thanks for the help in the mailing list in generating the facet_markers.

``` python

from __future__ import division

import meshpy.triangle as triangle
import numpy as np

def round_trip_connect(start, end):
return [(i, i+1) for i in range(start, end)] + [(end, start)]

def refinement_func(tri_points, area):
max_area=0.1
return bool(area>max_area);

def main():
points = [(1, 0), (1, 1), (-1, 1), (-1, -1), (1, -1), (1, 0)]
facets = round_trip_connect(0, len(points)-1)
markers = [2,2,2,2,2,2]

outter_start = len(points)
points.extend([(2, 0), (2, 2), (-2, 2), (-2, -2), (2, -2), (2, 0)])
facets.extend(round_trip_connect(outter_start, len(points) - 1))
markers.extend([3,3,3,3,3,3])

# build
info = triangle.MeshInfo()
info.set_points(points)
info.set_holes([(0, 0)])
info.set_facets(facets, facet_markers=markers)

#
mesh = triangle.build(info, refinement_func=refinement_func)

#
mesh_points = np.array(mesh.points)
mesh_tris = np.array(mesh.elements)
mesh_attr = np.array(mesh.point_markers)

print mesh_attr

import matplotlib.pyplot as plt
plt.triplot(mesh_points[:, 0], mesh_points[:, 1], mesh_tris)
plt.xlabel('x')
plt.ylabel('y')
#
n = np.size(mesh_attr);
inner_nodes = [i for i in range(n) if mesh_attr[i]==2]
outer_nodes = [i for i in range(n) if mesh_attr[i]==3]
plt.plot(mesh_points[inner_nodes, 0], mesh_points[inner_nodes, 1], 'ro')
plt.plot(mesh_points[outer_nodes, 0], mesh_points[outer_nodes, 1], 'go')
plt.axis([-2.5, 2.5, -2.5, 2.5])
#plt.show()
#
fig = plt.gcf()
fig.set_size_inches(4.2, 4.2)
plt.savefig('../../figs/sec5-meshpy-triangle-ex5.pdf')

if __name__ == "__main__":
main()
```