1.  If there are multiple ROCs with scalers (hardware, TI, FADC), then
there will be multiple event 129s.  In the case where analysis of these
events is defered to the end of the analysis, all the event 129s will
be analyzed as if a single event, resulting in one entry in the tree.
  2.  Scalers are read either when a given time interval has passed since
the last scaler read (usually 2 seconds), or every time an event is tagged
as a sync event.  In the case of the "timed" reading of events, different ROCs
with scalers are not guaranteed to read their scalers at the same event.  But
or sync event scaler reads, all ROCs read their scalers.  This commit adds an
option SetOnlyUseSyncEvents(Bool_t).  If set, only scaler events triggered
by the sync event will be analyzed.

Fix bug if more BCMs defined in db_PScalevt.dat than in gscalers.param

Fix bug that confused names "BCM1" and "BCM17" by adding ".scal" to the BCM name

1) Add X,Y offsets parameter for each plane, instead of X,Y offset per chamber
2) ModifyTHcDriftChamber::CorrectHitTimes to work with SHMS chambers
3) Add GetReadoutSide method to THcDCHit class.

  Eliminate duplicated/redundant configuration code in SCons system
  Let SCons also clean ROOT 6's *_rdict.pcm class layout cache files
  Move helper python script to site_scons
  Remove EVIO from scons as podd build will take care of it.
  Remove define of HALLC_MODS in SConstruct
    (There is no longer any Hall C conditional code in Podd)

Modify SConstruct to prepend $ROOTSYS/bin to PATH, rather and append, so that SCons will use the most recently defined user-installed version of ROOT, or the system version, if no user version exists.

Also, make ROOT environment definition consistent with Podd.

  Fixes broken MacOSX building

  Cache the Cherenkov detector and parent pointers in Init

  Fixes building when using scons 3.0 which apparenlyt uses python3.
  Fixes broken building on travis

  For both the layers and the array part
  Correct calculation of coordinates of the Array modules in the
    calorimeter effciency calculations.

Previous was trying to return the reconstructed target
quantities as values which was totaly wrong.

Changed so that the references are returned. Now get the
reconstructed target quantities properly.

  fpTimes, fScinTime, fScinSigma and fScinZpos are not used.
  These arrays were given fixed size of 53.
  This commit finishes issue #53

  Add #include <set>
  Don't use new method of looping over vector

- Apparently  $?loginsh is not a good indicator of whether tcsh is
  interactive (bad Google...) so we'll look for 'setup.csh' in the
  current directory and if it is present assume it is the hcana/
  directory.
- The if($?prompt) clause should print some info text if you're
  interactive.  (Although it's not definitive either -- dammit csh.)

-- Brad

- The underlying problem here is that $_ is assumed to contain the path
  to the sourced file, but that isn't true for any version of tcsh I've
  tried... (I guess it must have worked for some version of tcsh?)
  Anyway, that fails and $1 is also empty in the else clause so the
  script fails.

- Reliably finding the path to a sourced file is next to impossible in
  tcsh, so just encourage/force people to pass a path to the script
  directory as an argument.

-- Brad

  When the db file is read, a list of crate #s and a list of
  module types is made.  The tag of each bank searched for scalers
  must have a tag that is in one of these lists.
  If all scalers are in banks tagged with the module type, then,
  SetOnlyBanks(kTRUE) should be called in which case bank tags will
  be compared only against this list of module types.

  In the end method of the Hall C rocs, the scalers are read for a final
  time and a user event 129 is created with the scaler counts.
  These events do not always appear at the end of the coda file, leading to
  the risk that triggered scaler reads will appear after the event 129s.

  If the event type of these final scaler reads is passed to the scaler
  object with the SetDelayedType(int evtype) method, then the scaler analyis
  will cache these events and analyze them at the end of the analysis.

Update podd version