vtkChartMatrix now features a SetSpecificResize allowing to move
the bottom left point of a single chart. This feature is used in
vtkScatterPlotMatrix to move the Big Chart plot a little on top right
in order to prevent overlap, using painter to compute axis bounding rect.
It also add a compare operator to vector in order to use it in the stl.

It doesn't seem to have got everything, but it's a start.

Some resulting whitespace is suboptimal, but one day
we can fix that with clang-format. :)

Just ran this command:
run-clang-tidy.py -checks=-*,modernize-use-override -fix

(I actually had to build my own clang to make clang-tidy
output "VTK_OVERRIDE" instead of "override".)

The OpenGL2 backend does not define a vtkOpenGL2ContextDevice2D class,
so this test will not run when built with OpenGL2. Change it to run if
OpenGL2 is enabled as the OpenGL2 backend requires higher than OpenGL
version 2 OR if the OpenGL1 backend is enabled but Open GL version 2
is available.

Added alternate baseline for TestScalarsToColors output to
accommodate minor differences in OpenGL2 backend.

Code extracted from:

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

at commit 41bdf05361007f8c5186f3df9944d86bb273ed13 (for/vtk).

Off by default, this new option adds labels at the ends of the axis
with the range values. A separate printf-style format string can be
specified for these labels. If a tick label overlaps one of the range
labels, it will not be displayed.

Modified TestAxes and its baseline to exercise this new option.

1. Fix the "mosaic" bug, due to transform be rewriten
2. Fix a bug concerning the mtime of the scene
3. Adaptating tick size to the scene scaling
4. Updating baseline

This fix is in reference to bug report 0015325. The formula for width now
avoids a conversion from int to double back to int, and bounds are correctly
computed. TestHistogram2D and its resultant image have also been updated to
more closely match the original test image.

...rather than anisotropic scaling, since we can't render fonts like that
anyway.

Now vtkAxis will show the same number of ticks and the
same relative text size when magnifying the image. Text
will still be a little funky if the scaling is anisotropic,
but at the moment our freetype engine doesn't support
anisotropic dpi (not sure if the MPL backend can or not).

New API: vtkContextScene::GetLogicalTileScale() returns
the renderwindow tilescale or (1, 1) depending on whether
ScaleTiles is enabled.

There are some random fixes where vtkAxis objects were being
used without a scene, which led to segfaults when trying to
look up the tile scale internally.

The numbers of digits in the exponent for a float printed
in scientific notation is undefined. We get 2 or 3 digits on
different platforms.

Adding a new plot to plot a region named, vtkPlotArea. This can be used
to plot a region enclosed between two series.

Also added a new test, TestAreaPlot, to test this new plot.

Plots may do extra work when axis' Mtime changes. Don't change its Mtime
unless absolutely needed.

OpenGL 3.2 does not support line width greater than 1 although
many drivers do support it. This adds valid images for cases
where line width is greater than 1.0

Also minor VisualStudio error check

Also a few tweaks to get closer to OpenGL 32 support
for mesa.

Change-Id: I2d81fe807e22d50913a57fa63c934deadc8e0787

TestChartXYZ test fails on various platforms.  It turns out it is
failing due to floating point issues. Basically

1)	By default the plot data points define the bounds of the axes
    for the chart
2)	the axes define the plot area (clip stuff outside that)
3)	which defines the clipping plane equations
4)	which can in turn clip the original data points subject to
    floating point precision

This is what is happening (FYI, the point that disappears in the middle
is actual on the boundary in the depth axis). This patch
“fixes” the test by defining axes that are slightly bigger than the
data and replacing the valid image.

Change-Id: I25e7f05736f274ffc16633d8110767b9a5a5907b

Change-Id: I6512c3c8cc441464ec5d093dca9852f1a30acb6b

Convenience methods are exposed for plots in vtkChartXY

Change-Id: Iccccc18011b74314af3fcd1474ac792a15627ceb

Change-Id: I403f8db9f0c1746b45d6df38b07e98b42e1f04d8

vtkChartXY did not use the space used by the title text and top-axes to
determine where to place the title text. That resulted in cases where
the title overlapped the top-axes labels or for larger font sizes,
simply didn't have enough room. Fixed that by ensuring that the title
placement/border determination code uses the title text's size.

Updated TestLinePlotAxisFonts test to test this new functionality.

Change-Id: Idac42b970eb27ebbaee95adf43748d79747cd1dc

This commit cleans up the bar chart update logic, adds support for
using a lookup table to map an array to colors for the bars. It also
adds a test verifying the points, and the bars, using C++, and the
same array/lookup table.

Change-Id: I11585739f5a0c5bca3240274aebafead302e89ec

Lots of places where we pass floating point values as integers; make the
loss of precision more explicit.

Change-Id: Icafde395f0463421b5f36d1e4e855fa25cab5ab4

The 'inc' variable will take care of the implicit float cast. Avoid the
implicit downcast when passing i and j as indicies though.

Change-Id: Ia84dc4b67c102c0da3c0bd5ef43bd89d0e861e0d

Change-Id: Id90e98ae40559ef6076e9b9323824f47a08395f1

Change-Id: I5f19b101063e1a25d4f5db5c4356ce5efffbd90d

Change-Id: I10a54acbd076e055d8071414f2fff1ebafe1653d

Fix the way how the bags are computed. They are now performed at
50 and 99% of the distribution.
Functional bag plots also create a median serie which is the
median line of all input series. Outliers series are also
tagged with an "_outlier" extension in their name.

Change-Id: I990428c9b79278be9be725491232b3e9d69190b0

Fix the problem with the position of Y axis: long labels were trunked.
This position was fixed, it is not flexible depending the size of the labels.

Add new functions to the chart API to allow manipulation of series
by id (vs by name).
Add support for manual plot labels setting.
Enhance the boxplot displacement, it now correctly follow the mouse
without ugly jumps.

Change-Id: I400cd064616f7b93415e17db9749d709ea2f7345