there are some paths in the vtkXOpenGLRenderWindow (CreateAWindow for instance) that would need better NULL pointer check to avoid SEGFAULT and allow a clean abort.
this fix should also be merged to the release branch

3e03d8c6 point pickig with msaa was using a msaa depth buffer
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2132

5df73b10 In vtkVolumeOfRevolutionFilter, fix assignments to output cell locations.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2133

0f750114 Remove ES2 support for VTK 8.0
Acked-by: Kitware Robot <kwrobot@kitware.com>
Reviewed-by: Berk Geveci <berk.geveci@kitware.com>
Merge-request: !2126

a4bc54ee use a better and smaller data file fix bugs this exposed
7fc4b92b add the regression test input dataset
b94d6a59 Fix crash related to deselected blocks.
2e87ecc6 strengthen the test
ab244c3e A new reader for GE/LANL's TRUCHAS simulation files
Acked-by: Kitware Robot <kwrobot@kitware.com>
Acked-by: John Tourtellott <john.tourtellott@kitware.com>
Acked-by: Bob Obara <bob.obara@kitware.com>
Merge-request: !2108

msaa was not being turned off prior to setting the depth
buffer on point picking. Mixing a msaa depth buffer with
a non msaa color buffer was causing some issues.

f06e3ccf Don't pass subject for Python DeleteEvent
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2131

19d616bc Fix an error with depth peeeling and glyph rep
Acked-by: Kitware Robot <kwrobot@kitware.com>
Reviewed-by: David Lonie <david.lonie@kitware.com>
Merge-request: !2123

When Python observed a VTK DeleteEvent, it called GetObjectFromPointer()
to try to get (or make) a Python object for the vtkObject that invoked
the event.  Obviously if a vtkObject is being deleted, we do not want to
create a new Python object that references it.

ES2 is causing headaches and code mess and is
getting old. Remove for 8.0 so we can focus on
more modern APIs. Most Apple and Adroid devices
from the past 4 years support ES3.

cd1c29e6 Clean up by destroying the vtkTkRenderWidget.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Reviewed-by: Brad King <brad.king@kitware.com>
Merge-request: !2125

Currently, the only way for Python/Tk to force Tcl to delete the
vtkTkRenderWidget is to destroy() the widget.  This has to be done with
care, since the widget must hold the only reference to its vtkRenderWindow
at the time of its destruction.  If the widget is not destroyed, then a
memory leak results.

d1cfc837 improve the speed of building the IBO for polys
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2118

177f37d8 Padding filenames for ExodusII writer based on number of processes
Acked-by: Kitware Robot <kwrobot@kitware.com>
Reviewed-by: Cory Quammen <cory.quammen@kitware.com>
Merge-request: !2119

bd53b8f6 Update vtkExternalOpenGLCamera to use new explicit proj property.
fc820483 Add option to set an explicit projection matrix.
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2089

When depth peeling is used the glyph was still trying
to set glyphColors even though they were not used
resulting in an error message.

In order to read in ExodusII files written out in parallel we
need to pad the process ids with zeros so that all of them are
associated together.

There was a spot where an ivar could be
set without checking if it already had a value that
needed to be released.

The BinningFilter ivar may also  be in a reference
loop as the input to this filter. Have to report it.

The context was getting cleaned too often maybe.
This fix reverts one line from commit
https://gitlab.kitware.com/vtk/vtk/commit/
3a8e4497

This commit fixes issue

vtk/vtk#16876

and adds a test for that condition as well as
a few other cases that should work but could
cause problems.

This change could introduce other issues limited
to windows as the code is fairly complex and
the use cases vary widely. At some point the whole
process of creating windows and handling offscreen
drawing needs to be reworked. It is a mishmash of 20
years of code and overdue for a real workover for all
platforms. But this topic is intended to fix this one
case until we have time to really look at it and rework
it.

Currently the new test is only active for OpenGL2 and
Windows. OpenGL1 and other platforms has issues with
it that need to be fixed in the future. But for now this
change fixes the reported Windows issues and adds a
test to maintain that fix.

There were no C++ tests that directly tested this class, and this test
also covers the recently-discovered bug for TranformInputSamplingOff().

A simple omission in the code caused TransformInputSamplingOff() to produce
incorrect results.  The default is TransformInputSamplingOn(), and it seems
that no tests exercised the class with it off.

Fix some long standing issues with multisamples
and stereo capable window requests on Win32.

This fixes an MPI deadlock issue when LIC is used in parallel.

30dd43d7 use vtkMath::IsNan
Acked-by: Kitware Robot <kwrobot@kitware.com>
Merge-request: !2120