More unit tests for MATLAB replacer classes

[willGraham01]

Vertices class | Closes #170

• initialise() successfully assigns memory and decrements all stored values by 1, to be consistent with C++, rather than MATLAB indexing.

FieldComponentsVector class | Closes #172

• Instances can be initialised from structs that contain the appropriate fields, and have the expected sizes.
• Can handle both vertical and horizontal arrays being passed in.

ComplexAmplitudeSample class | Closes #173

• Construction via the empty struct does not cause an error, but exits construction appropriately.
• Otherwise, a struct with two fields provided will correctly initialise the attributes components and vertices.

GratingStructure class | Closes #174

• Fixes a double-deallocation bug when assigning from MATLAB arrays
• Checks appropriate construction/destruction/assignment methods

Pupil class | Closes #175

• Checks initialise() method performs correctly, given the MATLAB input
• Fixes a double-deallocation bug

EHVec class | Closes #176

• Can correctly assign fftw_complex to the entries of the object
• fftw_free used in destructor

FrequencyExtractVector class | Closes #177

• Constructor can handle both vertical and horizontal arrays
• max() function correctly identifies maximal element
• Data can be read in from MATLAB datastructs

CCollection class | Closes #178

• Error thrown when struct passed is of the wrong shape/size
• Correctly extracts components when fields are labelled correctly
• is_disp_ml and is_multilayer are "correctly" identified, although has flagged a logical flaw in one of these

DCollection class | Closes #179

• Similar functionality to CCollection, minus the possibility of being provided with 9 fields over 6.

CMaterial class | Closes #184

• Construction with incorrect fields throws an error
• Updates common method in matlabio.cpp to avoid seg-faults if the input has an incorrect fieldname

DMaterial class | Closes #185

• Literally identical to CMaterial barring a 6 instead of a 9.

[samcunliffe]

LGTM! Despite my picky comments about using full words for variables these are all nice tests.

Just to be clear: both me and Google prefer full names wherever possible...

Do not worry about saving horizontal space as it is far more important to make your code immediately understandable by a new reader.

Perhaps the only exception for us is if it's actually the name of a maths symbol from Peter's thesis/convention e.g. $C$, $D$. (Although even then we should probably try to understand the object and name it.)