ufl team mailing list archive

[ufl@xxxxxxxxxx (UFL)]

One or more new changesets pushed to the primary ufl repository.
A short summary of the last three changesets is included below.

changeset:   323:f59642d7192a0c94d8c63dd6dbd132a76018801f
tag:         tip
user:        "Martin Sandve Alnæs <martinal@xxxxxxxxx>"
date:        Thu Oct 16 20:48:30 2008 +0200
files:       ufl/algorithms/__init__.py ufl/algorithms/ad.py
description:
Minor fix.


changeset:   322:b6c21508cd983ce2653e50150ffd5e2dd3330581
user:        "Martin Sandve Alnæs <martinal@xxxxxxxxx>"
date:        Thu Oct 16 20:28:17 2008 +0200
files:       ufl/algebra.py ufl/algorithms/indexalgorithms.py ufl/base.py ufl/differentiation.py ufl/indexing.py ufl/tensors.py
description:
Added function repeated_indices() to UFLObject, with default
implementation "return ()". Implemented by Product,
SpatialDerivative and Indexed.

(I'm trying to avoid undocumented access to
protected "._foo"-variables)


changeset:   321:d9a1d79b5606ad124b0e8a86e5efe4f12f208e35
user:        "Martin Sandve Alnæs <martinal@xxxxxxxxx>"
date:        Thu Oct 16 20:14:46 2008 +0200
files:       ufl/base.py ufl/tensors.py
description:
Added free_index_dimensions() to UFLObject with a default
implementation that should work fine for most cases.
The exception is when creating tensors with dimensions
different from the spatial dimension. We'll probably
need to implement this function some more places,
but may be able to get away with a general default
implementation for most classes.

We need this for two reasons. One is for ComponentTensor
to compute the correct shape, and the other is for
Product.repeated_index_dimensions() to get the correct
dimensions from its operands.

----------------------------------------------------------------------
For more details, visit http://www.fenics.org/hg/ufl