f7e651fa When building the multi_backend code suppress warnings from thrust.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Kenneth Moreland <kmorel@sandia.gov>
Merge-request: !672
634a6136 Add license exception to version.txt
a6719638 Add Version.h header file
a575ec1d Add warning if git version disagrees with version.txt
3ffd16a8 Add ability to get VTK-m version from git
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !671
This allows source files to use the VTK-m version.
Note that Configure.h does not include the version. This is intentional.
First, I don't want the entirety of VTK-m to recompile every time the
version is updated (which is every time a commit happens in git).
Second, it is important that external project source files be able to
get the VTK-m version without actually loading any VTK-m code.
c6fdceaf Fix an error in the clip tables
Acked-by: Kitware Robot <kwrobot@kitware.com>
Reviewed-by: Sujin Philip <sujin.philip@kitware.com>
Merge-request: !669
There was an error in the clip tables for clipping lines. It was
referring to edge 1 for one of the case, but of course a line has only 1
edge at index 0. Fix the error in the clipping table and also add a
check for if we get an index to an invalid line.
There was an error in the CleanGrid filter where if the
CompactPointFields flag was off it would still try to compact the point
coordinates. Add an extra check so that the coordinate systems are just
passed when this flag is off.
There was a static assert in a templated function that was never
supposed to be used. Some compilers were fine as long as there was no
use of the templated function, but others errored out anyway. Change the
implementation to use the =delete keyword instead.
This won't have as nice of an error message, but you would have to jump
through some hoops to get to this point anyway. If this still does not
work, we could just probably remove the invalid templated function. The
compiler error would get even more obfuscated, but again I don't expect
anyone to actually run into it.
The implementation was calling PrepareForOutput on the delegate arrays
rather than PrepareForInPlace, do when used with CUDA you did not get
the data on the device.
Also added a regression test to check this.
This reduces the number of weak vtables vtkm generates, resulting in
a reduction of binary sizes for projects that include vtkm classes in
multiple translation units.
Class that need to be passed across dynamic library boundaries such as
DynamicArrayHandle need to be properly export. One of 'tricks' of this
is that templated classes such as PolymorphicArrayHandleContainer need
the Type and Storage types to have public visibility.
This makes sure that all vtkm storage tags have public visibility so
in the future we can transfer dynamic array handles across libraries.
