dolfin team mailing list archive

[Ola Skavhaug <skavhaug@xxxxxxxxx>]

Hi,

I'm trying to assemble a pde using pyDOLFIN, but I get some template related
errors.

The code:


from dolfin import *

P1 = FiniteElement("Lagrange", "triangle", 1)
element = P1 + P1

(v,q) = TestFunctions(element)
(s1,h1) = TrialFunctions(element)
(s0,h0) = Functions(element)

class K(Function):
    def __init__(self, element, mesh, value):
        Function.__init__(self, element, mesh)
        self.value = value
    def eval(self, values, x):
        values[0] = self.value

class Alpha(Function):
    def __init__(self, element, mesh, value):
        Function.__init__(self, element, mesh)
        self.value = value
    def eval(self, values, x):
        values[0] = self.value

class Beta(Function):
    def __init__(self, element, mesh, value):
        Function.__init__(self, element, mesh)
        self.value = value
    def eval(self, values, x):
        values[0] = self.value


mesh  = UnitSquare(10,10)
k     = K(P1, mesh, 1.0)
alpha = Alpha(P1, mesh, 1.0)
beta  = Beta(P1, mesh, 1.0)

a = v*s1*dx + q*h1*dx
L = v*s0*dx + q*h0*dx + k*( \
        alpha*s0*dot(grad(h0),grad((1-s0)*v)) \
        -beta*(1-s0)*dot(grad(h0),(grad(s0*v))) \
        + alpha*s0*dot(grad(h0),grad(q)) \
        + beta*(1-s0)*dot(grad(h0),grad(q)) \
        )*dx

A   = assemble(a, mesh)
rhs = assemble(L, mesh)


Crashes:

Traceback (most recent call last):
  File "sim_error.py", line 57, in <module>
    rhs = assemble(L, mesh)
  File
"/home/skavhaug/extsoft/lib/python2.5/site-packages/dolfin/assemble.py", line
49, in assemble
    coefficients_.push_back(c.f)
  File "/home/skavhaug/extsoft/lib/python2.5/site-packages/dolfin/dolfin.py",
line 8456, in push_back
    return _dolfin.STLVectorFunctionPtr_push_back(*args)
TypeError: in method 'STLVectorFunctionPtr_push_back', argument 2 of type
'std::vector<dolfin::Function * >::value_type'


Does anyone know what the problem is here? The c.f variable is a ffc_Function,
so perhaps the assembler doesn't know what to do. Is there a way of sending
the correct functions (concrete cpp_Functions, for instance) to the assembler?

Ola