This test compares harmonic distance restraints from colvars to
using fix spring in LAMMPS. Apart from numerical differences the
trajectories need to be the same. The run is broken down in chunks

1) run 100 steps with inactive restraints
2) run 100 steps with active restraints and write restart
3) continue for 100 more steps with active restraints
4) reset completely and restart from previously written restart file
   this should result in repeating the previous 100 steps.
5) reset again and restart from a previously written data file
   this should result in a slightly different trajectory
   due to re-initialization of the thermostat, but it should
   be different in the same way between fix spring and fix colvars.
   
 
