Create a simple 1:1 SpectraDetectorMap implementation

[Martyn Gigg]

This implementation will not need to store the whole map, it will just need to know the start and end spectra/IDs.

[Martyn Gigg]

(In [11990]) Create a simple 1:1 implementation of a spectra-detector map. Also the specid_t and detid_t types have been moved to their own header so that unnecessary dependencies can be removed, i.e. changing ISpectraDetectorMap now only recompiles 3 files in Geometry not nearly all of them. Re #3067

[Martyn Gigg]

(In [11991]) Remove some test code that shouldn't have gone in. Re #3067

[Martyn Gigg]

(In [12087]) Replace uses of the SpectraDetectorMap with the new interface. Note that mutableSpectraMap has now gone in favour of using MatrixWorkspace::replaceSpectraMap with a brand new implentation. Re #3067

[Martyn Gigg]

(In [12257]) Have a default 1:1, starting at 1, correspondence created by the spectra axis. Update the default 1:1 that is created with the MatrixWorkspace::init method so that it also starts at 1. This removes the need for most tests that care about it to bother creating their own spectra map and axis. Refs #3067

[Martyn Gigg]

(In [12258]) Fix the test. I accidentally reverted this without realising before I checked in. Re #3067.

[Martyn Gigg]

(In [12312]) SpectraAxis now understands how to build itself from a SpectraDetectorMap. Also, split the AxisTest into NumericAxis and SpectraAxis tests as they should be. Re #3067

[Martyn Gigg]

(In [12443]) Fix for an issue that Spencer reported around CreateWorkspace. The default SpectraAxis should now be sufficent and should not need updating byt the algorithm. Re #3067

[Stuart Campbell]

This ticket has been transferred to github issue ​3914