GitLab

Found with -Winconsistent-missing-destructor-override

The destructor in vtkAOSDataArrayTemplate would automatically
free memory in its Buffer data member instead of allowing the Buffer
to free memory when its reference count went to 0. This caused
issues for shallow copying data arrays.

Found and auto-fixed by clang-tidy’s readability-redundant-control-flow check.

Though I did have to manualy fix whitespace after it.

This commit will ensure that the "base" of stream used to output
vtkVariant storing VTK object is restored to "dec". Otherwise, any
integer streamed afterward would be displayed in hexadecimal base.

MSVC 2010 appears to assert that arguments to `std::copy` are non-NULL
and will crash even when both input pointers are NULL and the copy is
really a no-op.

Add a runtime check to prevent the `std::copy` call from being made when
there is no data.

With fewer and fewer sprintf remaining, many of these needed new tempories to be added.

Types of warnings:
- mismatch between parameters names in declaration vs definition
- make certain parameters references, or const references
- checking array index after its use
- prefer preincrement
- change return type of operator=

There are some very minor API chages here (const/reference).

These are already informally commented as legacy, this just
allows them to be truly compiled-away.  Will help totally removing them one day,
which is desirable because they violate C++ naming rules by having
2 or more adjacent underscores.

snprintf is now part of C++11 and is already simulated for old versions of VS.

This commit fox a warning in vtkDataWriter
for a incorrect usage of sprintf with vtkIdType

This commit replace a bunch of stat() usage by Stat
which is more robust especially on windows.

This will re-build the BelowRange, AboveRange, and Nan color entries
in the lookup table if they are out of date.  Without this change, if
a user calls SetBelowRangeColor() etc. after the table has been set or
built, then the table will remain unchanged.

Accidentally merged topic instead of just running tests...

This reverts commit 19d37821, reversing
changes made to a07ee519.

Kunda authored Mar 11, 2017 and Cory Quammen committed Mar 14, 2017

doxy and misc. typos pt2

This improvement has been mainly done to custom the OpenVR home button
The same logic has been conserved for this home button in VR (Exit App when
triggered)
but it is now customisable by the user by either
  - adding an Observer on vtkCommand::FourthButtonReleaseEvent
                                    (see below other vtkCommand added)
                                    ::FourthButtonPressEvent
                                    ::FifthButtonReleaseEvent
                                    ::FifthButtonPressEvent

  - overriding vtkInteractorStyle::OnFourthButtonUp()
                                 (see below other methods added)
                                 ::OnFourthButtonDown()
                                 ::OnFifthButtonUp()
                                 ::OnFifthButtonDown()

This is needed to resolve a cycle causing a leak as a result of
add a numpy array to a dataset where the array was derived from
an array associated with the dataset.

This had me confused as I recently learned that VTK
renders transparent actors in an order.

Also removed unneeded cast, and updated other comments.

Out-of-range colors were not being set up correctly when
vtkLookupTable::SetTable() was called. Instead of just resizing the
table for the special colors, actually build them.

Added test for this use of vtkLookupTable.

As suggested in vtk/vtk#16916

In vtkAbstractArray::SetNumberOfValues(), we would get less values
when the passed in value was 189527780 due to round-off issues
for single component arrays. Making this function more robust
to avoid memory errors.

Fix compiler error

In the general case this avoids a branch in FunctionValue.
It also gives the option to specialize EvaluateFunction so
that it can be further optimized and/or run in parallel.

Add floor option to show a floor

completely rework the matrix and coordinate code
to be cleaner and use standard VTK conventions (right handed)

add support for a dashboard overlay with options to subclass
and support for saving/loading camera poses and adjusting
motion factors and scale factors.

Improve the resetcameraclippingrange to work better for VR

change how eye position is handled so that VolumeRendering
will work properly

Handle the case where poeple make the cubes invisible which
results in a 308 error on loading the model.

Add support for the user to specify the mapping of physical
coordinates into world coordinate. Allows for different
view up vectors other than the old code which was hard
coded to Y.

Add middle mouse support mapped to the grip button.

Add some support for a hardware clipping plane.

This adds the member functions vtkDataArray::ComputeFiniteRange
to retrieve the range of data array value, ignoring NaN and infinity.

This is the ground work to allow the user to pass a function for a
buffer that should be called to free it.

When VTK is passed an aligned buffer on windows, it was unable to properly
free the memory as it wasn't aware of _aligned_free. This issue was detected
when passing VTK-m allocated buffers to VTK.

At camera positions close to the bounding box (outside), certain parts of the
proxy geometry would fall behind the near plane getting those fragments
(and the volume image chunk sampled by those rays) clipped off. To avoid this, the
IsCameraInside criteria is changed to check if the camera's near plane is
intersecting the bounding box (by performing Plane-AABB intersection). With this
approach, the proxy geometry is clipped by vtkClipConvexPolyData accordingly
(see OGLGPUVolumeRayCastMapper::RenderVolumeGeometry), ensuring all of the
geometry falls within the camera's range.

- Made a few APIs' params const
- Changed many usages of the evil GetBounds() to expect a const
  return value, even though it does not (yet?) return const
- Slightly changed return values of vtkTextActor3D::GetBounds()
  to return its own ivar instead of its ImageActor's

These tests were taking way too long on some or all
test platforms. In some cases more than two minutes
to run.

Test just needed some tightening up.