Transversely shifted distribution in temporal monitors

In temporal monitors the distribution is shifted in both transverse directions by half a centimeter.

There are more problems with monitors than initially thought.

• The reference of a (infinitesimal thin) monitor is its center. This is handled correctly for monitors that are placed with ELEMEDGE but not with absolute coordinates.
• the path length stored in the h5 file isn't correct in case of "temporal" monitors
• the stored position of the bunch relative to the reference particle in case of "temporal" monitors is shifted

Some more information on this. If I set a monitor at a position where Opal also stores the phase space then the phase space of the temporal monitor should be the same. Furthermore the stored s-pos and the position of the reference particle in lab coordinates should be the same as the position of the monitor (in input file and data/_ElementPosition.txt).

For an example from bERLinPro I got
Position of monitor:
s-pos: 0.690843
3D: (3.036208, 0, 9.772690)

Position of reference particle in temporal monitor:
s-pos: 0.694659
3D: (3.0345, 1.16469e-14, 9.76799)

The phase space: