This reverts commit 30ea407b. Some
additional work is needed to size fonts appropriately with non-default
DPI.

This improves chart font size (and ideally other chart element size)
on HiDPI displays.

This reverts commit aa59cdbd.

If vtkRenderWindow was told to do renders on back buffer only
(i.e. offscreen renders), the connected QVTKOpenGLWidget could still end
up inadvertently display the rendered content onscreen if QOpenGLWidget
tried to compose the frame buffer together with the rest of the UI.

This was causing issues like paraview/paraview#17256.

Fixed by making QVTKOpenGLWidget flag such offscreen renders and ensure
that real render gets called before Qt decides to compose the frame
buffer.

There were two problems that needed to be fixed for the wheel events.
The first problem had to do with changing direction: since the
AccumulatedDelta kept the residual after the threshold was passed,
the first wheel click in the opposite direction was sometimes ignored.
Not keeping any residual fixes this, though it makes scrolling on a Mac
trackpad a bit less smooth.  The second problem is that the delta()
method is deprecated for Qt5, and angleDelta() should be used instead.

QtSQL headers are included in public headers of this module, so
dependent libraries need to link to Qt as well.

Utkarsh Ayachit authored 8 years ago and Ben Boeckel committed 8 years ago

Similar to what we're doing with Qt5, this change now adds called to
vtk_module_export_code_find_package() for Qt4. That way, when importing
a module using Qt4, we automatically find Qt4.

Also changed to use imported targets rather than library paths when
linking to Qt4 libraries.

QVTKOpenGLWidget would render see-through, when using Mesa
(paraview/paraview#17159). The symptom was seen on OsX when
the default format requests alpha buffers (which is now the default
for standard Render Window). While this may be exposing some odd
side effect in Qt 5, explicitly making the renderer use alpha=1.0
when filling the background overcomes this issue.

Anyways, there was no API to vtkViewport to explicitly set the
background alpha, only background color (RGB) was exposed. Exposed that
(with default unchanged). QVTKOpenGLWidget uses the new API to tell
all non-transparent renderer to render with alpha=1.0.

QVTKOpenGLWidget was not determining correctly if the FBO was using
multi-samples. Fixed that.

Special handling is needed since QOpenGLWidget changes QSurfaceFormat to
always have samples=0, since the OpenGL context is not a multisample
context. We have to query the render-buffer to know if its using
multisampling.

QVTKOpenGLWidget now ensures that vtkOpenGLRenderWindow is reinitialized
to obtain the current default framebuffer object id after resize. We
were not handling the case were the default frame buffer was recreated
but the active OpenGL context wasn't recreated (which happens in
resize).

Without this widgets in dialog were causing issues in the main window's
OpenGL widgets when they were closed. This resolved issues in my tests.

This enables retinal Mac to work, there is a little extra work to
account for the pixel ratio not triggering a full resize event as is
expected, and a question of whether we need to invoke a configure event.

This is continuation of 8e61675a. We missed the
`mouseDoubleClickEvent` event. We needed to handle that.

We were calling `QVTKInteractorAdapter::ProcessEvent` twice: once in
`QVTKOpenGLWidget::event` and another time in `QVTKOpenGLWidget::mouse..`
handlers. Fixed that,

I observed on left-click and drag that the entire window would get
dragged on Linux when using the KDE/Plasma desktop without this change.
Once applied then the left (and other mouse buttons) all behave as
expected. Simply returning early from the switch statement was not
effective in my testing - hence the additional overrides.

This may be necessary for the other events too, in order to ensure the
ordering is as expected in teh Qt signal-slot system, but I have not
done extensive testing with key presses etc.

vtkRenderWindow::MakeCurrent should not cause the active frame buffer
binding to change. Calling `QVTKOpenGLWidget::makeCurrent` did exactly
that. Updating QVTKOpenGLWidget code to avoid doing that.

QVTKOpenGLWidget's cache texture if left active could have side effects
in other code (e.g. `vtkDualDepthPeelingPass`). Overcome those by
unbinding the texture after we're done with it -- which is a good
practice anyways.

vtkGenericOpenGLRenderWindow now supports a notion that it's not ready
for rendering. This happens when the context is created by someone else
e.g. QVTKOpenGLWidget. In that case, the vtkGenericOpenGLRenderWindow
shouldn't start rendering before the parent has created the context etc.
This adds ability to do that.

QVTKOpenGLWidget was failing due to incorrect determination of
MaximumHardwareLineWidth. For the time being, fixed by adding ability to
override it in vtkGenericOpenGLRenderWindow.

Adding QVTKOpenGLWidget which enables using a vtkRenderWindow (rather a
vtkGenericOpenGLRenderWindow) with a QOpenGLWidget (added in Qt 5.4).

Kitware Robot authored 8 years ago and David Gobbi committed 8 years ago

This commit reindents the code with the following utility:
Utilities/Maintenance/vtk-reindent-code.py

This utility changes the positions of the braces so that they are no
longer indented relative to the code block they reside in.

The bash command line used was the following:

for d in Charts Common Deprecated Domains Examples Filters GUISupport \
         Geovis IO Imaging Infovis Interaction Parallel Rendering \
         Testing Views Web Wrapping; do
  for e in cxx cxx.in txx txx.in hxx hxx.in h h.in c c.in; do
    find "${d}" -name "*.${e}" -exec \
      python Utilities/Maintenance/vtk_reindent_code.py {} +
  done
done

generated via:
cd $VTKSRC
perl Utilities/Doxygen/doc_header2doxygen.pl --to ~/tmp/vtkheaders .
cp -r ~/tmp/vtkheaders/* .

The current dependency relationship in vtk is unclear and misleading.
This MR tries to rewrite them based on header files inclusion of headers
and source files in each module. Corner cases are considered and
modules are sorted in alphabetical order to facilitate future reference.
See details in my gitlab python based script project. In future we can continue
using this script to clean the VTK Dependency easily from
time to time.

This is a fix for:
#14310

Windows applications that run for a long time report that rendered
objects do not change. This is because the modified time on a Windows
system is 32 bits. This causes overflows that defeat the modified time
mechanism. This patch defines a new type, vtkMTimeType that is 64
unsigned integer regardless of the architecture.

A mechanism to provide backward compatibility is introduced. The
preprocessor define "VTK_HAS_MTIME_TYPE" can be used in applications
that must build against VTK versions that use the "unsigned long" type
for MTime's.

Methodology used to find MTime occurences:
1) Identify files as follows:
   git grep "unsigned long" | grep ime | cut -d":" -f1,1 | sort | uniq
2) Hand edit each of those files replacing "unsigned long" with
   "vtkMTimeType" where appropriate.
3) Temporarily change typedef for vtkMTimeType to "double" to detect
   missing conversions

In module configuration files generated for Qt-based modules, we now
export find_package() calls that find specific components from Qt5 i.e.
`find_package(Qt5 COMPONENTS Widgets)` instead of
`find_package(Qt5Widgets)`.

Find/replace of:

;[ ]*//\s*Not implemented[\.]*
to
 VTK_DELETE_FUNCTION;

To catch a few remaining ones missed by previous greps.

Manually reverted changed files in VPIC and KWSys folders, and a couple other of places.

vtk(.*)\(const vtk\1&\);\s*//\s*Not implemented[\.]*
to
vtk\1(const vtk\1\&) VTK_DELETE_FUNCTION;

vtk(.*)\(const vtk\1 &\);\s*//\s*Not implemented[\.]*
to
vtk\1(const vtk\1 \&) VTK_DELETE_FUNCTION;

vtk(.*)\( const vtk\1 & \);\s*//\s*Not implemented[\.]*
to
vtk\1( const vtk\1 \& ) VTK_DELETE_FUNCTION;

vtk(.*)\( const vtk\1& \);\s*//\s*Not implemented[\.]*
to
vtk\1( const vtk\1\& ) VTK_DELETE_FUNCTION;

vtk(.*) \(const vtk\1&\);\s*//\s*Not implemented[\.]*
to
vtk\1 (const vtk\1\&) VTK_DELETE_FUNCTION;

(operator\s*=.*);\s*//\s*Not\s*implemented[\.]*
to
\1 VTK_DELETE_FUNCTION;

After that, this finds basically nothing:

operator.*implemented

then manually reverted changed files in VPIC and KWSys folders.

These files will eventually need vtkConfigure.h
for eventual VTK_DELETE_FUNCTION usage.

Control and shift key state was set in vtkRenderWindowInteractor
from QVTKInteractorAdapter, but the state of the alt key was not.
This patch addresses that oversight.

Hat tip to Mengda Wu <wumengda@gmail.com> for providing a patch
on the VTK users mailing list that describes the needed fix.

Calling iren->Render() in paint event is fraught with problems
especially for applications that want to show progress when rendering is
happening. Added a mode to QVTKWidget to allow deferring render requests
in paintEvent so they happen on idle. This avoids recursive paint call
errors due to updating progress bars to show progress of the render.

Qt5 doesn't define Q_WS_*, instead we need to use Q_OS_* flags. This
updates the code to check for Q_OS_LINUX in addition to Q_WS_X11 flags.

SafeDownCast performs a series of virtual calls and string comparisons,
which is quite slow, especially when used in worker functions.

vtkArrayDownCast will switch between SafeDownCast and the more
efficient FastDownCast (only available for common vtkAbstractArray
subclasses) depending on whether or not FastDownCast is defined for
the specific array type.

When Qt is used with a trackpad (rather than with a wheel mouse) the
MouseWheel events did not correspond to the actual amount by which the
finger had moved (or, at least, this was the case on OS X).

By accumulating the delta values and only invoking a ScrollWheel event
when the accumulated delta passes a certain threshold, the OS X trackpad
interaction can be fixed.

Kitware Robot authored 9 years ago and Dan Lipsa committed 9 years ago

Code extracted from:

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

at commit 41bdf05361007f8c5186f3df9944d86bb273ed13 (for/vtk).

vtkDataArray subclasses now use

ComponentValue --> TypedComponent
TupleValue --> TypedValue

for their type-specific methods.

# Conflicts:
#	Rendering/Annotation/vtkScalarBarActor.cxx

rendering to a memory HDC is not support beyond opengl 1.4
so change the print methods to use FBOs