- Jun 12, 2023
-
-
-
2184a4cd Add support for rotations in VRML Transform nodes d60adbd4 Detect and handle incompatible depth attachment formats c02c0da6 vtkOpenGLProjectedTetrahedraMapper: fix depth buffer blit 897a852b vtkOpenGLProjectedTetrahedraMapper: move shader binding 269b7b03 openfoam: add a test for decomposed GGI / global face zone 2864e102 openfoam: ignore extended number of faces (fixes paraview#14235) 55558016 Replace utf8 right quote with ascii quote 82837c8c Remove bad byte 0x0f from a source file ... Acked-by:
Kitware Robot <kwrobot@kitware.com> Acked-by:
buildbot <buildbot@kitware.com> Reviewed-by:
Julien Fausty <julien.fausty@kitware.com> Merge-request: !10268
-
2184a4cd Add support for rotations in VRML Transform nodes d60adbd4 Detect and handle incompatible depth attachment formats c02c0da6 vtkOpenGLProjectedTetrahedraMapper: fix depth buffer blit 897a852b vtkOpenGLProjectedTetrahedraMapper: move shader binding 269b7b03 openfoam: add a test for decomposed GGI / global face zone 2864e102 openfoam: ignore extended number of faces (fixes paraview#14235) 55558016 Replace utf8 right quote with ascii quote 82837c8c Remove bad byte 0x0f from a source file ... Acked-by:
-
Also fixes crash if tabs are used as values separators
-
- When the depth formats are incompatible, sample the depth with shaders. Otherwise, blit color and depth attachments together.
-
Fix depth buffer copy when a floating point framebuffer is available. When the bit depth of VTK's render FBO and the one used by this mapper are not the same, direct blitting from one FBO to the other does not work. Fix that by separating the color buffer blit from the depth buffer blit.
-
Move binding of the shader program needed for the projected tetrahedra rendering algorithm closer to where it is needed. This makes it less likely to interfere with other shader programs that may be used in the algorithm.
-
-
- occurs with foam-extend globalFaceZone, in which the faces list is longer than owner/neighbour lists (with the extra faces belonging to the globalFaceZone). Remedy by simply resetting the known NumFaces to be consistent with owner/neighbour information. The excess faces will simply be dragged along without being addressed. This is a minor memory overhead (versus truncating the face list), but only affects cases with globalFaceZone and still allows some possibility to handling in the future if desired. Adjustment consistent with CreateCellFaces() private method.
-
-
The file vtkImageData.cxx had control byte 0x0f (^O) at the end of one of the comment lines. It has been removed.
-
- Jun 08, 2023
-
-
Upsized to to int64_t before multiplying. Fixes some unit test failures under UBSan.
-
The vtkQuadricClustering algorithm would overflow when compiled with VTK_USE_64BIT_IDS=OFF, because it would store huge integer values in a std::set<vtkIdType>. This was noted in particular with vtkQuadricLODActor. A 64-bit integer is needed in order to avoid the overflow.
-
- errors on the root processes would broadcast and return, except there was no corresponding broadcast on the other processes (ie, it would block). - reworked logic to simplify the handling. - misc code cleanup for broadcasts (code style only). - replace memmove with memcpy for broadcasting vtkStringArray contents (guaranteed to be non-overlapping) - remove BroadcastStatus private method (now unneeded)
-
-
Previously, theses legacy reader/writer didn't support field data for some composite data structure like vtkPartitionedDataSet. As it is use by extractor, field data was lost during a catalyst simulation for example.
-
Wanted to make clear I took that test from TestMappedGridDeepCopy and didn't entirely write it myself.
-
Tests vtkUnstructuredGrid::ShallowCopy with a vtkMappedUnstructuredGrid source and simulates pipeline execution where executives calls vtkUnstructuredGrid::Initialize before copy.
-
Made vtkUnstructuredGrid::ShallowCopy check and create internal arrays as necessary if input is a vtkUnstructuredGridBase
-
Improve documentation about `dataset_adaptor.append(array, name)` method. Specially, specify two cases of `array`: * shape[0] is matching number of elements so `array` is used "as is" * otherwise `array` is **copied** at each index of a `newarray`, where `newarray.shape[0]` is the number of element In the second case, numpy may throw a `MemoryException` because of a too big memory requirement. Handle this to add explanation about the `shape` expectation.
-
The Java wrappers were not blocking templated parameters, even though they are unable to wrap them. This resulted in build errors for methods like this one: void AddArray(vtkAOSDataArrayTemplate<int>* array); The fix for these build errors is to reject all methods with parameter types that include "<" as part of their name.
-
-
821fe510 vtkXRenderWindowInteractor: Fix a logic error with repeating timers Acked-by:
David Gobbi <david.gobbi@gmail.com> Reviewed-by:
Jaswant Panchumarti <jaswant.panchumarti@kitware.com> Merge-request: !10264
-
- Jun 07, 2023
-
-
Mathieu Westphal (Kitware) authored
-
- May 31, 2023
-
-
-
d9553d30 Merge branch 'upstream-cgns' into cgns-missed-hdf5-compat c1883071 cgns 2023-05-31 (24ce45b6) 96eea346 cgns: fix a missing version check Acked-by:
Mickael PHILIT <mickey.phy@gmail.com> Merge-request: !10244
-
Ben Boeckel authored
# By cgns Upstream * upstream-cgns: cgns 2023-05-31 (24ce45b6)
-
Code extracted from: https://gitlab.kitware.com/third-party/cgns.git at commit 24ce45b6ebb640cf0e8509249489acb45bcdfcae (for/vtk-20230531-4.2.0). -
Ben Boeckel authored
Also backports https://github.com/CGNS/CGNS/pull/712 to support 1.8-era HDF5 files.
-
-
1903dc0d Fixed bug #18995: corrected order of deallocation Acked-by:
-
- May 30, 2023
-
-
Sean McBride authored
Call super last, as per Objective-C requirements.
-
- May 24, 2023
-
-
-
e5c87764 Add release note for wireframe shading improvement 81783448 Make wireframe normal computation more robust Acked-by:
-
e5c87764 Add release note for wireframe shading improvement 81783448 Make wireframe normal computation more robust Acked-by:
-
David Gobbi authored
-
David Gobbi authored
This allows wireframes to render correctly on linux with amdgpu drivers. The previous code always used dFdx as the line vector, unless dFdx.x was equal to zero. The new code adds dFdx and dFdy to get the line vector, after using dot product to get them both pointing in the same direction (since they might be pointing in opposite directions along the line).
-
- May 16, 2023
-
-
-
b7aee2da Merge branch 'upstream-netcdf' into netcdf4.9.2-update b6f42059 netcdf 2023-05-05 (12d264fe) f480035c Updated netcdf update.sh to tag `for/vtk-20230505-4.9.2` Acked-by:
Ben Boeckel <ben.boeckel@kitware.com> Merge-request: !10199
-
- May 12, 2023
-
-
Sean McBride authored
# By netcdf Upstream * upstream-netcdf: netcdf 2023-05-05 (12d264fe)
-
