Issue with CPU version

[RaulPPelaez]

Calling SetPositions twice makes the solver segfault.

After creating a valid solver using the cpu as device, the second call to SetPositions crashes the code:
For instance:

import numpy as np
from common_interface_wrapper import FCMJoint
device = 'cpu'
domType = 'DPSC'
nP = 2 #Number of particles
has_torque = True # Set to True if angular displacements are needed
viscosity = 0.957e-3
L=128
xmin = -L/2
ymin = -L/2
H = 2
zmin = 0
zmax = H
solver = FCMJoint(device)
solver.Initialize(numberParticles=nP, hydrodynamicRadius=1.0155, kernType=0,
                  domType=domType, has_torque=has_torque,
                  xmin=xmin, xmax=xmax, ymin=ymin, ymax=ymax, zmin=zmin, zmax=zmax,
                  viscosity=viscosity, optInd=0, ref=False)
np.random.seed(1234)
xP = np.zeros(3 * nP, dtype = np.double) # Arbitrary xP inside the Z domain
xP[0::3] = np.random.uniform(xmin, xmax, nP)*0 #X
xP[1::3] = np.random.uniform(ymin, ymax, nP)*0 #Y
xP[2::3] = np.random.uniform(zmin, zmax, nP)*0 #Z
xP[3] += 1.000000;
torques = np.random.normal(0, 1, 3 * nP)*0
forces = np.random.normal(0, 1, 3 * nP)*0
solver.SetPositions(xP)
V, T = solver.Mdot(forces, torques)
solver.SetPositions(xP)
print("This is never printed")

[snatesh]

I'm unable to reproduce this error. The script is executed completely on several machines I've tried. One thing to try is make sure zmin/zmax are doubles and not integers. I will look more into this and the other issue over the weekend.