* Test that BinMD does copy basis vectors so that query can later be made
* Change the rebin mode to use the MDGeometry allBasisNormalized property
* Change the line mode to also use the allBasisNormalized property

Introduce test and then functionality to MDGeometry.

Now WorkspaceSingleValue has dimensions = 0, so this would throw.
However there does not appear to be a particular reason for throwing, so
we remove it.

Refs #11512

The main change here is to what CutMD can do. It Clears BOTH the transforms and the original workspaces off the output workspace.

1) This fixes the issue with the labeling when plotMD is run, because there is no original workspace to try to take dimension information from.

2) However, the plots generated from CutMD from the SliceViewer are now garbage. This could because the line cuts (BinMD) are operating over a Histogram workspace already (CutMD created with NoPix=True), but is something I will need to get to the root of.

Not strictly related to this ticket, but I was surprised when I made a
change in ImplicitFunctionParameter.h and everything recompiled.
This turned out the be the reason.

This has been prevented thus far, because use of coord transforms has been based on the presence of original workspaces. Original workspaces will not be present upon a reload.

I've also exposed the new getters on MDGeometry to python.

Also added tests for estimateResolution

Mainly stub which should be coordinated with other parts of MD Geometry.

Also, as agreed during the discussion, renamed QhQkQl to Q3D and made changes which would make similar modifications easier in a future.

The framework should be mostly clear of cppcheck errors now.

Also make doxygen more explicit.

all things prepared to implementation of the ticket but nothing substantial have been done yet.

To reduce warnings when coord_t = float. Also, fix a warning #4473.

which should fix CloneMDWorkspace. Also, fix to BinMD to strip spaces from dimension names.

Like BinMD and SliceMD.

To fix segfaults on destruction. Multiple virtual inheritance is hard ;)

Works on MDHisto and MDEvent.

Most observers that watch for workspace deletion do not need a copy of the
shared pointer. The PostDeleteNotification only has the name of what
was deleted meaning the object should only be kept alive by a notification
if absolutely necessary. The PreDeleteNotification does what the original
DeleteNotification did.

when the original workspace is removed from AnalysisDataService. This should allow the destructor to be called properly, freeing memory and closing the file. Should let windows system test pass.

Refs #3907: Made MDEventWorkspace inherit from IMDWorkspace. Added tests to MDGeometryTest. Clarified the getDimension() calls, which were ambiguous.

Refs #3672: Moved CoordTransform up to API namespace. Added basis vectors to the output MDHistoWorkspace when calling BinToMDHistoWorkspace.