Analysis.cxx and Analysis.h   Had to use a manual merge tool (meld)
to do this.

This brings in Shuya's version of Analysis.cxx
It also removes the Projects/T40 from the github repo.

Now there are two options:
    1. Ignore ring+sector matching entirely (set SIGMA of the window to zero).
If you do this you just get the ring assuming that you set TAKE_E_RING in the config file.
    2. Do the ring+sector matching as before (set SIGMA to some non-zero value).
If you do this, the code attempts to find a match just as before. However, if
it doesn't find one, it sets the sector to the fake "sector 9" and takes
phi at the center of the wedge. One can gate on matched vs. non matched events
in the resulting ROOT file simply by looking at sector number (sectors 1-8 mean
a match was found, sector 9 means it wasn't).

NB - changes are still untested; Eames will need to do this.

Fixed bug in the strip number calculation due to the 'cyclic' nature of the detector and leading to false phi strip numbers in some cases

What the code does is take the sector to be a "fake" sector at the center of
the wedge.  So this way all relevant things (phi angle, impact angle, etc)
are calculated off the center of the wedge later on in the code, which is the
best we can do w/o using the actual sectors.

Fixed error message w/ trying to calculate energy loss for negative values - round 2, also checking the second energy loss for hyball and barrel

Adding an argument for particle-3 phi in the lab to take calculate correctly recoil Phi angle, usefule for recoil spec and in some case to Doppler correction

sectors