dolfin team mailing list archive

["Garth N. Wells" <gnw20@xxxxxxxxx>]

Shawn Walker wrote:
> ok, I see some FIXME's.  I will try and do it better.  I think if I put 
> the initialization and creation of this higher order mesh_coordinates 
> function into the MeshEditor class, then that will be better.

Also, take a look at XMLFile. There was some dodgy code in there which 
caused a leak, but I've improved it now gotten around the leak by 
improving the Function interface. The improved interface should help 
with what you need.

Garth


> - Shawn
>
> On Thu, 11 Sep 2008, Garth N. Wells wrote:
>
> > Shawn Walker wrote:
> > >  ok, then I will wait until it is fixed, and then update my dolfin code
> > >  from the repository.
> >
> > Take a look now.
> >
> > Where possible use references rather than plain pointers.
> >
> > Garth
> >
> > >  - Shawn
> > >
> > >  On Wed, 10 Sep 2008, Garth N. Wells wrote:
> > >
> > > > > >  DOLFIN wrote:
> > > > >  One or more new changesets pushed to the primary dolfin 
> > > repository. A
> > > > >  short summary of the last three changesets is included below.
> > > > > > >  changeset:   4729:a84cf2b5cda6170367b30e1c6bcffa00bc4f36d1 tag:
> > > > >  tip parent:      4727:04152c9160962363af3a996773ced092caed339d
> > > > >  parent:      4728:a5ad46cb91f08dac998bd2a506f6c7e6aa4d0da1 user:
> > > > >  Anders Logg <logg@xxxxxxxxx> date:        Mon Sep 08 00:03:12 2008
> > > > >  +0200 files: description: Merge patch from Shawn Walker
> > > > > > > > >  changeset:   
> > > 4728:a5ad46cb91f08dac998bd2a506f6c7e6aa4d0da1 parent:
> > > > >  4703:0e3198c7a287f5d7e8604f0f26fa8cc958fbc5b4 user:        Shawn
> > > > >  Walker <walker@xxxxxxxxxxxx> date:        Wed Sep 03 10:56:23 2008
> > > > >  -0400 files:       dolfin/function/DiscreteFunction.h
> > > > >  dolfin/function/Function.cpp dolfin/function/Function.h
> > > > >  dolfin/io/XMLMesh.cpp dolfin/io/XMLMesh.h dolfin/io/XMLObject.cpp
> > > > >  dolfin/io/XMLObject.h dolfin/io/XMLVector.h
> > > > >  dolfin/mesh/MeshEditor.cpp dolfin/mesh/MeshEditor.h
> > > > >  dolfin/mesh/MeshGeometry.cpp dolfin/mesh/MeshGeometry.h 
> > > description:
> > > > >  This patch adds in support for reading in an additional 
> > > 'coordinates'
> > > > >  tag in the mesh.xml file, and uses this information to create a
> > > > >  Function* called mesh_coordinates that is contained in the
> > > > >  MeshGeometry class.  This function stores higher order mesh
> > > > >  coordinate data that comes from assuming the mesh elements are 
> > > mapped
> > > > >  using a finite element function higher order than linear.  Another
> > > > >  boolean array is created inside MeshGeometry that indicates which
> > > > >  elements are ACTUALLY curved, and which are just straight 
> > > (standard).
> > > > >  This is completely backwards compatible with previous mesh formats.
> > > > > > >  I've commented out some of this, so it's not functional at 
> > > the moment.
> > > >  The problem was that it introduced a number of memory leaks. Once 
> > > we get
> > > >  the Function ownership sorted (hopefully I'll get this done 
> > > tomorrow) it
> > > >  should be easy to turn it back on leak-free.
> > > > >  Garth
> > > > > > > >  changeset:   4727:04152c9160962363af3a996773ced092caed339d 
> > > user:
> > > > >  Anders Logg <logg@xxxxxxxxx> date:        Sun Sep 07 23:51:54 2008
> > > > >  +0200 files:       bench/fem/jit/bench.py
> > > > >  dolfin/la/SLEPcEigenSolver.cpp description: Fix bug in
> > > > >  SLEPcEigenSolver (bug reported by Evan Lezar)
> > > > > > >  
> > > ----------------------------------------------------------------------
> > > > >   For more details, visit http://www.fenics.org/hg/dolfin
> > > > > _______________________________________________ DOLFIN-dev mailing
> > > > >  list DOLFIN-dev@xxxxxxxxxx
> > > > >  http://www.fenics.org/mailman/listinfo/dolfin-dev
> > > > >
> > > > _______________________________________________
> > > >  DOLFIN-dev mailing list
> > > >  DOLFIN-dev@xxxxxxxxxx
> > > >  http://www.fenics.org/mailman/listinfo/dolfin-dev
> > > >