* unrecognize -e option on various linux platform

  with OS X 10.9
* Remove self assignment warning in TMust2Physics (issue warning, not
 affecting logic)
* Remove self assignment warning in TCatsPhysics, actually correct a
 potentiatlly nasty bug

* linking fail with gcc

  - Getter was returning the whole vector instead of the single element
  - Not take a unsigned int as argument to specify the element to return

        - Renaming the file TTiaraData into TTiaraBarrel
        - Adding variable Name, Getter and Setter

 - The generated and copied file are now ignored

     and 'ebad93e4331f7e4117babdd1024d81d11cc38944'

 - Some old machine requiered the -e flag in front of echo to interpret correctly the \t command as a tab
- some new machine ignore -e and the make fail later
- The solution to both is to suppress the -e flag (done before) and replace \t by the ascii code \011

     and '7c82e3f88c0f723c336731836b94dcefdab2cbe4'

- The Beam info were filled but cleared too soon
- A flag telling whenever the info need to be cleared has been added

* Remove double declaration of header in Simulation.cc file

  in NPSimulation/Tigress

* Change CLHEP unit system for NPSystemUnit in NPAnalysis
  for W1 detector

  header and src files of each defined detector are now
  symlinked in the include and src directory instead of
  being hard copied

* Remove compilation warnings in Tigree.cc

     and 'ebd0455fede9857160b27010db6b644a08ff88f6'

   + geant4 files are kept in the include and src directories
   + one directory is created for each detector. This directory
     contains the associated headers and sources files
   + a script directory is created with some scripts allowing the
     conditional compilation

 * Add #ifdef statements in DetectorConstruction.cc file
   + this applies to each detector header

 * Move G4MultiFunctionnalDetector.hh header from Sharc.cc to 
   Sharc.hh

  this branch is dedicated to the conditional compilation of NPS

- Sin theta was not applied to the cross section

 - In the case of multiinteraction in the same volume, the scorer was keeping only the last energy deposit, instead of summing them
 - The threshold has been move in the Read sensitive and is no longer in the scorer, as the energy deposit can be down in multiple step each of them below threshold and be ignored if the threshold is applyed individually.

* Fixing GeometricalEfficiency.C macro to fit new TInitialCondition
* Adding support for TInitialCondition in Sharc

 - The code was returning the atomic mass instead of the nuclear one

 - The DetectorThickness token has no effect