Fixed now

re #13938

re #13938

In addition to this, identified subtle bug in copyToMD

As part of this issue, I found that calculating the inverse of a 1x1
matrix results in a segmentation fault

With this change, the previously failing test passes.

The stream operators still use commas as before but the free functions
now allows users to specify the delimiter. Refs #7199

The Vates quantification requires using the Sobol sequence for producing
a point in ND space. The interface as it was made it difficult to
swap out other random number generators. There is now a generic ND
interface that returns a vector of numbers.

can not compare results as the results widely differ on test data provided.

refs #4813 further step in separating transformation matrix from ConvertToMDEvents

and made it more general.

The wrong coordinate transformation was being calculated. Tests happened to pass because the transformation in the previous tests were too simple

This required a const to be added to a member of Matrix and
similarly to OrientedLattice.

Refs #455, Fixes #3195: ISAW peaks file's HKL's are flipped so as to be consistent with internal Mantid representation of Q = ki-kf. Fixed the LoadIsawUB so that the U matrix is a pure rotation (no -1 factor).

Refs #3175. Sorry for the huge commit but certain things require Matrix to be in Kernel. Given how Quat, V3D and Matrix inter cooperate it seemed sensible for them to be in the same place and arguably this is where they belong anyway.

Refs #3175. Add a MatrixProperty that will allow Geometry::Matrix objects to be passed into algorithms. Also renamed the MantidMat typedef to DblMatrix so that it's clearer what the thing actually is.

OrientedLattice exposed to python. Check if U is proper rotation. Partial fix for LoadIsawUB. Refs #455

refs #2745 changed the location of bounding box wrt to the  beam line and detector to calculate angular detector's sizes more accurately in Cartesian topology case;

Refs #3157 Sort our System.h. It was becoming a dumping ground. It should only be touched when absolutely necessary as it affects every file in Mantid. Also combined this with Refs #2345 to sort out DLLExport/DLLImport usage. It's done for Kernel, Geometry and API but stuff in the plugins could do with updating as well.

Refs #3137: Used coord_t and signal_t in appropriate spots in MD* classes. Right now both are typedef'd to double but it builds if typedef'd so something else e.g float.