I apparently forgot to update the version before the last patch release.

Utkarsh Ayachit authored Nov 10, 2011 and Kenneth Moreland committed Nov 10, 2011

Setting C_TEST_PATH to include the configuration type only for generators
supporting multiple configurations.

C_TEST_PATH is now complete and correct. So all tests can simply use it without
worry about platforms or generators.

Utkarsh Ayachit authored Nov 09, 2011 and Kenneth Moreland committed Nov 09, 2011

MPI_C_LIBRARY and MPI_C_EXTRA_LIBRARY flags are no longer supported.

Utkarsh Ayachit authored Nov 09, 2011 and Kenneth Moreland committed Nov 09, 2011

On several machines, FindMPI picks mpicxx by default. Linking with this C++
library causes issues with the C-based test executables.

FindMPI.cmake has gone revisions in recent versions of cmake (in 2.8.6 for
example). FindMPI now provides language specific MPI flags. Fixed the code to
use these MPI_<lang>_<variable> flags to avoid the afforementioned issue.

Apparently MSVC does not support the same environment setting commands
that POSIX uses.

The MPI_IN_PLACE feature, which allows you to do collective operations
such as gather with the local data already in the destination array, is
a feature that is not available in MPI 1.  To continue to support MPI
version 1, implement an alternative copying version that is only used
for that version.

The way the translation transform was picked, I was getting z shifts
that were so close that after OpenGL transforms you ended up with the
same depth value, which of course threw off the compositing.  This
change limits the precision used for the Z shift which seems to avoid
really close values but does not seem to encourage the same value
(although that is still a possibility).

This behavior makes includes in modules shipped with CMake still look
in the module directories defined by the project.  The change does not
effect building IceT on its own.  (At least, there are currently no
modules included that are also included with CMake.)  However, this is
currently necessary to prevent warnings in ParaView.

Support latest means of loading packages.  Add check for version.
Export targets correctly.

This completes the changes in the documentation.

- icetUnsafeStateGet now returns const.
- new ICET_COLLECT_TIME timing metric.
- ICET_COLLECT_IMAGES option.
- ICET_VALID_PIXELS_* state variables.
- Matrix helper functions.
- Input array for icetSendRecvLargeMessages became const.
- Background correction in strategies.

The latest MSVC debugger apparently checks to make sure that any
variable used as an argument to a function has been set.  There was
an instance in the reduce strategy that violated it.  I don't think
it could ever cause a problem (the arguments are ignored in this
particular state), but it's still bad programming style.  Thanks
to John Biddiscombe for bringing this up.

- ICET_MAGIC_K and ICET_MAX_IMAGE_SPLIT CMake, env, & IceT variables.
- icetCompressedCompressedComposite.
- Sparse image copy and split.
- Image interlacing.
- Image equal functions.
- icetDecompressSubImage.
- icetRenderTransferSparseImages.
- Sparse image null functions.

Now accepts and returns sparse images.  Also no longer collects image
pieces.

- Radix-k and its selection.
- icetAddTiles raises ICET_INVALID_VALUE for 0-pixel tile.
- icetFindRankInGroup and icetFindMyRankInGroup.
- New communication functions.
- Sequential strategy skips some collective operations.
- icetImageAssignBuffer and icetSparseImageAssignBuffer.
- ICET_SIZE_TYPE.
- ICET_IN_PLACE_COLLECT.