ffc team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

This is a bug in FFC, thanks for noticing. Pull the latest hg version
or make the following fix in the function build_rhs in the file
src/ffc/compiler/optimization.py.

Where it says

  # Special case, no terms
  if len(terms) == 0:
      return "0.0"

change to

  # Special case, no terms
  if len(terms) == 0:
      return ("0.0", 0)

/Anders


On Wed, Sep 06, 2006 at 07:48:02PM -0400, osman b wrote:
> Hi,
> Just installed 0.3.3 on my SuSE10.0 AMD 64, tried 
> ffc  NavierStokes.form
> it worked.
> But -O gives an error:
> osman@linux:~/SRC/ffc-0.3.3/src/demo> ffc -O -d1 NavierStokes.form
> This is FFC, the FEniCS Form Compiler, version 0.3.3.
> For further information, go to http://www/fenics.org/ffc/.
> Parsing NavierStokes.form
> Output written to NavierStokes.py
> 
> Compiling form: w0_a1(dXa2/dxa0) | vi0[b0]*va1[a0]*((d/dXa2)vi1[b0])*dX
> Number of terms in form: 1
> Number of functions (coefficients): 1
> Number of projections (coefficients): 1
> Number of constants: 0
> Finite element of test space:  Vector Lagrange finite element of degree 1 on a 
> tetrahedron with 3 components
> Finite element of trial space: Vector Lagrange finite element of degree 1 on a 
> tetrahedron with 3 components
> Finite elements for functions: [Vector Lagrange finite element of degree 1 on 
> a tetrahedron with 3 components]
> Compiling tensor representation for interior
> Compiling term 0
> Created geometry tensor: a[3, 12, 3], b[]
> Pretabulating basis functions at quadrature points
> Initializing quadrature.
> Total degree is 2, using 2 quadrature point(s) in each dimension
> Precomputing table of basis functions at quadrature points.
> Computing table for factor v = vi0[b0]
> Computing table for factor v = va1[a0]
> Computing table for factor v = ((d/dXa2)vi1[b0])
> Computing product of tables
> Computing the reference tensor (15552 entries), this may take some time...
> ................................................................................
> Reference tensor computed in 0.0685 seconds
> Created reference tensor: i[12, 12], a[3, 12, 3], b[3]
> All terms compiled
> Generating code for reference tensor
> Generating optimized code for element tensor
> Computing optimization, this may take some time...
> Traceback (most recent call last):
>   File "/usr/bin/ffc", line 128, in ?
>     main(sys.argv[1:])
>   File "/usr/bin/ffc", line 82, in main
>     execfile(outname, ns)
>   File "NavierStokes.py", line 32, in ?
>     compile([a, L, M], name, "dolfin", {'blas': False, 'no-gpl': False, 
> 'optimize': True, 'debug-no-geometry-tensor': False, 
> 'debug-no-element-tensor': False})
>   File "/usr/lib/python2.4/site-packages/ffc/compiler/compiler.py", line 55, 
> in compile
>     forms = build(sums, name, language, options)
>   File "/usr/lib/python2.4/site-packages/ffc/compiler/compiler.py", line 124, 
> in build
>     form.AKi = ElementTensor(form.sum, "interior", format, cKi_used, gKi_used, 
> options, None)
>   File "/usr/lib/python2.4/site-packages/ffc/compiler/elementtensor.py", line 
> 70, in __init__
>     self.aK = self.__compute_element_tensor(format, options)
>   File "/usr/lib/python2.4/site-packages/ffc/compiler/elementtensor.py", line 
> 124, in __compute_element_tensor
>     return self.__compute_element_tensor_optimized(format)
>   File "/usr/lib/python2.4/site-packages/ffc/compiler/elementtensor.py", line 
> 184, in __compute_element_tensor_optimized
>     return optimize(self.terms, format)
>   File "/usr/lib/python2.4/site-packages/ffc/compiler/optimization.py", line 
> 54, in optimize
>     (value, num_mult) = build_rhs(rhs, j, iindices, aindices, num_terms, 
> format, num_mult)
> ValueError: too many values to unpack
> 
> Also get the same error with:
> NeumannProblem.form, Mass.form, Elasticity.form, PoissonSystem.form, 
> Optimization.form, and Stokes.form
> 
> Do I need new Ferrari? (Mine is 0.1.0)