thanks Andrew for finding these

these tests fail with it the other way.
C_CTestXdmfWriter (Failed)
C_CTestXdmfAggregate (Failed)
Cxx_TestXdmfAggregate (Failed)
Cxx_TestXdmfFunction (Failed) new fail
Cxx_TestXdmfSubset

xml file for aggregate is no different, but reading
balks at lack of "Content" the other way around

vtk side changed CMAKE_BINARY_DIR to CMAKE_CURRENT_BINARY_DIR
so that xdmf would build cleanly in its own place within projects
that use it. Doing so moved xdmfcore.py and xdmfcoredsm.py
but then ctest wouldn't find them. This change puts the python
submodules into the top level.

rather than force -fPIC into CMakeLists.txt we now explicitly turn
on position independent compilation for each of the three internal
object libs. Do it this way because fpic is not valid on all
compilers.

Changes from ARL's Repo (C Wrappers, DSM rework, Custom Types, Templates, GridControllers, TIFF integration, XdmfAggregate)

allow wrapping without DSM



See merge request !6

Add support for Polyhedron cells

Elements of type Polyhedron (i.e. N face cells, where each face is a M edge polygon) are in the following format: 
[nCellFaces, nFace0Pts, id0_0, id0_1, ..., nFace1Pts, id1_0, id1_1, ..., ...]

This has first been proposed in a MR to VTK (vtk/vtk!1255), which contains the necessary changes to VTK. The VTK MR also has example/test files attached that demonstrated the functionality.

Ping: @demarle 


See merge request !5

Added a top-level README.md file in Markdown format.

Added a `README.md` file in Markdown format. This file contains brief instructions to build XDMF from source code. The contents of this file are displayed on the project page at Gitlab. Additional documentation for new users may be added to this file in the future. 

@demarle

See merge request !4

Elements of type Polyhedron (i.e. N face cells, where each face is a M edge
polygon) are in the following format:
[nCellFaces, nFace0Pts, id0_0, id0_1, ..., nFace1Pts, id1_0, id1_1, ..., ...]

Elements of type Polyhedron (i.e. N face cells, where each face is a M edge
polygon) are in the following format:
[nCellFaces, nFace0Pts, id0_0, id0_1, ..., nFace1Pts, id1_0, id1_1, ..., ...]

One of the commands was not in a code block. Fixed it.

Added a top-level README file in Markdown format. Contents of this file are displayed on the project page on Gitlab.

Address a regression introduced when backporting vtk's changes



See merge request !2

make submitted test results go to public cdash

now run
  ctest -D Experimental
and see results at
  https://open.cdash.org/index.php?project=XDMF

See merge request !1

now run
  ctest -D Experimental
and see results at
  https://open.cdash.org/index.php?project=XDMF

vtk's regression tests are symbolic links to other files that
do not end with .xmf or .h5. When we try to open those we
segfault.

Some style. Some fixes, especially for windows with the exports and template changes.

More so adapt this to work well as a subdirectory in a larger project. Don't force settings
that the parent project owns, allow parent project to turn on tests without xdmf's tests and
make the name of the library an option so that parent can have two xdmfs.

Merging the Xdmf"2" repository history into the Xdmf"3" repository
to keep both histories together in one unified repository.

Merging the Xdmf"2" repository history into the Xdmf"3" repository
to keep both histories together in one unified repository.

Checking for shared faces is not enough when inserting vertices to
generate higher order elements. Matching faces depends on the
order of traversal of the elements. Now we check for shared edges in addition
to faces - this solves the traversal ordering problem as we will pick up cases
where two elements share an edge before the faces can be matched. Before, this
caused duplicate points to be inserted on the shared edge.