Re: [Libmesh-users] <kein Betreff>

[Roy S. <roy...@ic...>]

On Thu, 14 Feb 2008, Roy Stogner wrote:

> Thank you for the report.  On first glance, it looks like the problem
> here is that FE<>::init_shape_functions is what resizes the storage
> space for second derivatives, but that function gets overridden in
> FEXYZ, which was never updated with second derivative support.  I'll
> fix it in SVN this weekend.

A beginning of a fix (it should work in 1D, or in 2D/3D if you make
specific requests of the FEXYZ object which don't include calculating
d2phi) is now in the SVN head.  I'll get the rest of the fix in later,
but anyone whose code uses XYZ elements might want to try out the
latest subversion libMesh.  I don't think we use discontinuous bases
in examples, and my code is mostly C1 stuff not discontinuous, so we
don't have any FEXYZ test coverage yet.
---
Roy