Add comment to the prior import of MPI in xcorr.py & setupclass.py:
on EULER, this leads to a fork() warning:
http://gfs.sourceforge.net/wiki/index.php/System(),_popen(),_fork()_and_OpenMPI

+ minor changes

+ no "Found XXX detections" for normal run (only for "--verbose")

It used to be a debug message

Don't trust the chunk number written to file, but instead check each
processed chunk time range. This avoid bugs when resuming work for new
days (in a cotinuous scan that doesn't have a end date)

- append chunk overlap also to last chunk
- always save chunks.dat when preparing a scan
- move logging of "New template found" from TemplateManager to TemplateMatching
  (when its indeed new)
- save more often (yes, Luca, I like to check progress more often :-)))
   (hope it's okay for you)
   ... it's not much what is saved anyway...

Once a chunk has been processed by a node, keep assigning the same
chunk (with new templates) to the same node. This improve chunk
cache usage ad also avoid that multiple nodes loads too many chunks.
This is relevants when the nodes reside  in different machines and
loading a chunk require network I/O

Allow tolerance of 0.1 seconds