GitLab

30ddd95b vtkTimerLog: use Schwarz counter
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Cory Quammen <cory.quammen@kitware.com>
Acked-by: Ben Boeckel <ben.boeckel@kitware.com>
Merge-request: !7444

30ddd95b vtkTimerLog: use Schwarz counter
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Cory Quammen <cory.quammen@kitware.com>
Acked-by: Ben Boeckel <ben.boeckel@kitware.com>
Merge-request: !7444

bb1b4ee4 Merge branch 'upstream-netcdf' into update_netcdf_2020_11_24
16bbddf8 netcdf 2020-11-16 (3fd31153)
1fc608c2 Updated netcdf tag to for/vtk-20201116-4.7.4
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Ben Boeckel <ben.boeckel@kitware.com>
Merge-request: !7441

7a6312d2 Fix issues with vtkArrayListTemplate use.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Will Schroeder <will.schroeder@kitware.com>
Merge-request: !7443

7a6312d2 Fix issues with vtkArrayListTemplate use.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Will Schroeder <will.schroeder@kitware.com>
Merge-request: !7443

16530928 vtkCompositeDataSet: copy-structure fixes
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Cory Quammen <cory.quammen@kitware.com>
Merge-request: !7442

16530928 vtkCompositeDataSet: copy-structure fixes
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Cory Quammen <cory.quammen@kitware.com>
Merge-request: !7442

vtkTimerLog used a static `std::vector` that caused a issues on certain
platforms (see discussion on paraview/paraview!4565).

Updating vtkTimerLog to use a Schwarz counter idiom to alloc/dealloc the
vector to avoid double-free.

320464e6 made `vtkArrayListTemplate` depend on
`vtkDataSetAttributes::RequiredArrays` which means that removing arrays
from output DSA after calls to `InterpolateAllocate` etc. is no longer
valid. When array are excluded, these filters were doing exactly that
causing issues e.g. paraview/paraview#20344.

Fixed by ensuring that before we call InterpolateAllocate, we mark the
excluded arrays as `CopyOff` so they are not allocated in the output DSA
in call to `InterpolateAllocate`.

vtkCompositeDataSet::CopyStructure would not copy data-information and
other meta-data. Since CopyStructure is often used in executive and
other filters in place of shallow-copy where leaf datasets are to be
shallow-copied, this causes meta-data like time to be dropped. This was
manifesting as paraview/paraview#20052.

Fixed that.

3983d154 FindNetCDF: fix regex and mangle function.
35d6579b Merge branch 'upstream-ioss' into mpi-netcdf-ioss-fixes
4e10851f ioss 2020-11-24 (e6b06185)
1f7f35cf tpl: update ioss for MPI-awareness fixes
f0db1727 netcdf: detect parallel-support
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Ben Boeckel <ben.boeckel@kitware.com>
Merge-request: !7440

3983d154 FindNetCDF: fix regex and mangle function.
35d6579b Merge branch 'upstream-ioss' into mpi-netcdf-ioss-fixes
4e10851f ioss 2020-11-24 (e6b06185)
1f7f35cf tpl: update ioss for MPI-awareness fixes
f0db1727 netcdf: detect parallel-support
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Ben Boeckel <ben.boeckel@kitware.com>
Merge-request: !7440

81c02c87 fix progressive refinement start with ospray ray cast render
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !7438

81c02c87 fix progressive refinement start with ospray ray cast render
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !7438

special case mode handling for raycaster needed to happen after
the comparison with previous mode

# By netcdf Upstream
* upstream-netcdf:
  netcdf 2020-11-16 (3fd31153)

Code extracted from:

    https://gitlab.kitware.com/third-party/netcdf.git

at commit 3fd311538003e313cfebcf59eab0dc0611713a1c (for/vtk-20201116-4.7.4).

c30d90ab Fixed out-of-bounds NSArray access by checking index first
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Cory Quammen <cory.quammen@kitware.com>
Merge-request: !7437

50be6c42 fix crash when ray trace molecule data that has no bonds
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !7439

50be6c42 fix crash when ray trace molecule data that has no bonds
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !7439

# By Seacas Upstream
* upstream-ioss:
  ioss 2020-11-24 (e6b06185)

Code extracted from:

    https://gitlab.kitware.com/third-party/seacas.git

at commit e6b0618587774d36fd6b7d64a75b3f764a6e3354 (ioss/for/vtk-20201124-v2020-04-27).

fixes paraview/paraview#20345

NetCDF now detects (and reports) parallel support to CMake.

156fb524 vtkmClip: reduce compilation memory overhead
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: Robert Maynard <robert.maynard@kitware.com>
Merge-request: !7397

Bug introduced in 080ba5d7.
Unlike other languages, in Obj-C it’s invalid to access past the end of an array. Similarly, exception handling isn’t mean for normal control flow.

Benchmark
---------

Version | Size sum of object files | Minimum memory to compile
--------|--------------------------|--------------------------
Old     | 74M                      | 4GiB
New     | 107MiB                   | 3GiB

Changes
-------

- Splits `vtkmClip` TU into two TUs, one of each solely including
  its corresponding vtkm filter Clip{WithField|WithImplicitFunction}.

- Created an internals class to further encapsulate several
  types and methods needed for the previous point.

- It removes unneeded includes to the vtkm filter .hxx files.
Signed-off-by: Vicente Adolfo Bolea Sanchez <vicente.bolea@kitware.com>