GitLab

If TRANSVCUTOFF is negative or zero, these distribution work the same as before. (The default value for TRANSVCUTOFF is set to -3.0 and the other routine that reads it takes the absolute value.) If it is greater than zero, then sigx_m defines the Gaussian sigmas in x and y and particles greater than the radius defined by TRASNVCUTOFF are excluded. We made this change because SLAC is reporting lowered emittances using Gaussian cut transverse profiles comparted to the usual uniform profiles.

		MINR, MAXR, MINZ, MAXZ. 
- if a particle is out of cyclotron aperture or out of the field map, 
   it will be deleted after this step and its parameters are recorded in the *.loss file
- change data type of particle id from int to size_t
- many minor cosmetics of display massage

Change the arguments of CCOLLIMATOR element to XSTART, XEND, YSTART, YEND, WIDTH so as to model non-radial oriented collimator

Now opal-cycl can restart from hdf5 generated by opal-t, but the appended hdf5 file can not be open by H5root: Dude! Inconsistent number of scaler step attributes between first and last step 

Merge branch 'develop' into svn: using caches to speed up sc calculation (mainly relevant in gun calculations with many particles)

    - Prepare list of cavities to check against OUTSIDE particle loop
    - Integration timer reduced by 15 % wrt previous revision

    - Inlined ParallelCyclotronTracker::checkGapCross into ::push
    - Convert perpendistance from mm to m, do calculations in m
    - Only compute distOld when necessary, stop if is already clear that
      no cross occurs
    - Integration timer for firstTurn.in example is more than 15 % smaller
      (LF and MTS profit from that, not yet RK4 because RK4 uses old function)