- Sep 05, 2024
-
-
Ben Boeckel authored* static-crash: token 2024-08-26 (a6a7aedd) Revert token 3rd-party… Add a release note. Force order of unregistration of cell-grid responders. Clean up registration of cell-grid responders before exit. Allow applications to clean up before leak reporting. Acked-by:
Ben Boeckel <ben.boeckel@kitware.com> Acked-by:
buildbot <buildbot@kitware.com> Merge-request: !11392
-
- Aug 26, 2024
-
-
David Thompson authored# By token Upstream * upstream-token: token 2024-08-26 (a6a7aedd)
-
Code extracted from: https://gitlab.kitware.com/utils/token.git at commit a6a7aeddcf0148d07fe63da43f8074b04371c6ca (for/vtk). -
David Thompson authored
… so we can start updating it properly.
-
David Thompson authored
-
- Aug 16, 2024
-
-
1dcf73e3 vtk-m: vtkmAccelerator use new vtk-m API 8374972b vtk-m: update to v2.2.0 Acked-by:
Kitware Robot <kwrobot@kitware.com> Tested-by:
-
- Aug 14, 2024
-
-
David Thompson authored
Cell-grid responder classes often hold other VTK objects. If the registrar holding responders happens to be destroyed after the vtkDebugLeaks or the vtkFilteringInformationKeyManager classes perform their finalization, this can result in either false-positive leaks or crashes (when information-keys that have been deleted are dereferenced) or both. Add vtkCellMetadata::ClearResponders() to force all cell-grid responders to be unregistered and ensure it is called before vtkDebugLeaks::ClassFinalize or vtkFilteringInformationKeyManager::ClassFinalize are run.
-
David Thompson authored
This uses the new `vtkDebugLeaks::AddFinalizer()` method to unregister cell-grid responders before leaks are reported.
-
David Thompson authored
-
- Aug 09, 2024
-
-
14305452 Merge branch 'upstream-ioss' into update-ioss-exonull 863a314d ioss 2024-08-08 (0dda1fa8) b81ec74a ioss: update.sh Acked-by:
Cory Quammen <cory.quammen@kitware.com>
Merge-request: !11378 -
Spiros Tsalikis authored
# By Seacas Upstream * upstream-ioss: ioss 2024-08-08 (0dda1fa8)
-
Code extracted from: https://gitlab.kitware.com/third-party/seacas.git at commit 0dda1fa8d652ec7cfd9949802e147a788463989b (ioss/for/vtk-20240808-v2024-06-27). -
Spiros Tsalikis authored
-
- Aug 06, 2024
-
-
Vicente Bolea authored
(cherry picked from commit 6bd65269)
-
Vicente Bolea authored
(cherry picked from commit 49afe9b2)
-
- Jul 31, 2024
-
-
777a7465 vtkCompositeDataProbeFilter: Test point validity for composite HTG 43cf6487 vtkCompositeDataProbeFilter: Fix composite HTG point validity c5267791 vtkWrapPythonMethodDef: fix `bugprone-casting-through-void` lint c20549db vtkWrapPythonMethod: avoid unnecessary casts when wrapping `vtkObjectBase` 793a84a7 ImageMap: Fix an issue with indexing a multi-component image 1766bd31 Cast `char` fields durint the PrintSelf for correct display 9aea378e Allow for nullptr entries in the ConcatenateDataArrays method 5ef7fad9 Release note for vtkSkybox changes ... Acked-by:
Louis Gombert <louis.gombert@kitware.com> Merge-request: !11328
-
- Jul 29, 2024
-
-
(cherry picked from commit 3a0201d3)
-
When used with a composite dataset composed of HTG, probe points were not marked as valid for the global composite probe. (cherry picked from commit 73d7c734)
-
The generated code triggers a warning from `clang-tidy`. The `void*` cast is wrong because function pointers should not be cast through data pointer types. (cherry picked from commit 65fe2f7a)
-
New `clang-tidy` warns when casting a value of type `T` to `T` itself. Avoid generating code that trips this warning. (cherry picked from commit 219a7d6e)
-
Indexing into a multi-component image, the pointer was walking an additional step using the scalar size. Since, the pointer is templated over the underlying type, indexing directly into the active component yields the right result. (cherry picked from commit e070775b)
-
(cherry picked from commit 9eb02a52)
-
Those are simply skipped, and does not lead to crash anymore. (cherry picked from commit 1ef0bcfb)
-
(cherry picked from commit b4723506)
-
(cherry picked from commit 3217a646)
-
(cherry picked from commit fb71c796)
-
- closes vtk/vtk#19395 - without it, the batched polydata mappers would not update the buffers even when the composite data display attributes may have been modified due to a scalar visibility setting change. - the desktop mapper worked because it's `GetNeedToRebuildBufferObjects` indirectly compared the number of polydata in the current batch and previously rendered batch. This happened in `vtkStateStorage::operator !=()`. When the last block's scalar visibility was toggled from off to on, it recognized that the batch size has changed and uploaded the new batch of polydata. - the wasm mapper did not because it's `IsUptoDate` did not know about batch size and returned true even when the composite data display attributes were modified. - the fix could record the number of polydata in a batch and compare the number of polydata in `IsDataObjectUptoDate` with the last render. However, a better fix is to consider the MTime of the parent `vtkCompositePolyDataMapper` (that also considers the MTime of `vtkCompositeDataDisplayAttributes`) as the MTime concept was meant to fulfill this kind of state tracking need. (cherry picked from commit 4081a153)
-
- address vtk/vtk#19395 (cherry picked from commit b87561c3)
-
(cherry picked from commit 85fd10a6)
-
(cherry picked from commit 50f1238e)
-
* the `Scalars` array was still used to get Increments (to navigate in the data array raw pointer) leading to some inconsitencies when Scalars has a different memory layout than the effective InputArrayToProcess see issue #19386 Test was disabled due to ImageComparison (see #19221) Reimplement it in C++ to use `CompareDataObjects` instead of rendering. (cherry picked from commit 636b907b)
-
(cherry picked from commit 32049f80)
-
(cherry picked from commit 21ec9a0c)
-
- Jul 25, 2024
-
-
bc12f91a Fix order of sides created by vtkCellGridComputeSides. e121cc43 Test that vtkCellGrid sides are properly round-tripped. d130a1cd Properly serialize vtkDGCell::Source. 13f68aa7 Add "legacy" IO support for cell grids. 76582f81 Fix templated cell-grid methods. Acked-by:
-
- Jul 24, 2024
-
-
David Thompson authored
This makes testing and debugging easier (as well as user inspection).
-
David Thompson authored
-
David Thompson authored
The `SideType` ivar was not being recorded, which resulted in rendering errors upon a round trip.
-
David Thompson authored
This simply packs cell-grid JSON data into a "legacy"-like file so that ParaView can transfer geometry from server to client (when dataset size is small).
-
David Thompson authored
These methods did not properly deal with either iterator types or type-casting their return values from a vtkSmartPointer.
-
- Jul 22, 2024
-
-
42aa6902 Merge branch 'upstream-glew' into glew-osmesa-macros bd80a1df glew 2024-07-22 (fa840e66) 7231836e glew: update to work with newer OSMesa Acked-by:
Ryan Krattiger <ryan.krattiger@kitware.com>
Merge-request: !11341
-
