In THcScalerEvtHandler::AnalyzeBuffer there is a loop over
the scalers to keep track of scalers with a cut on the current.
There was a mistake in using scal_prev_read[nscal-1] instead of
dvars_prev_read[ivar] in the IF statement.

fGoodFlags holds event-by-event data, but was never cleared,
eventually growing to several hundred megabytes in a long analysis.
The effciency calculations in THcHodoEff that rely on the contents
of fGoodFlags have been most certainly garbage up to now.

I've replaced the event-by-event ClearEvent() method with the
standard virtual function Clear() and have added fGoodFlags.clear()
there.

Also, I've added assertions to the hodoscope code to check
the validity of some non-obvious array indices. Nothing triggers;
the index logic appears sound (despite its appearance ...)

This includes a particularly nasty leak in the shower code that caused
several hundred bytes to be lost _per event_.

Currently both CMake and SCons build not only libPodd, but also libHallA,
even though it is not required for hcana. hcana is linked only against
libPodd, however, so libHallA is simply a spurious byproduct.

Modify THcAerogel::CoarseProcess



For each PMT, fGoodPosAdcMult.at(npmt) was being set to the total number of positive PMTs with an
adc hit and fGoodNegAdcMult.at(npmt) was being set to the total number of negative PMTs with an
adc hit.

To be like the other dectectors make fGoodPosAdcMult.at(npmt)
and  GoodNegAdcMult.at(npmt) be the number of adc hits for that particular
PMT.

  For KaonLT, the kPulsePedestal values being returned for
  the helicity signal channels was 0 and 16380.  For SIDIS, the
  values for off and on are ~2000 and ~1400.  Made the default
  threshold 8000 instead of 1000.  Also added an optional parameter,
  "helicity_fadcthreshold" so that this threshold can be changed
  at run time.

Undoes delayed helicity reporting by learning seed of pseudo random
sequence used by injector helicity electronics.

This detector class should be added to the Trigger apparatus.
Delivers the global variables "hel" and "helrep".  "helrep" is the
reported delayed helicity and "hel" is the actual helicity.  + and -
helicity are reported as +1 and -1.  If the event is during a MPS
settling period, then the helicity variables are zero.  For the first
few seconds of a run, the "hel" variable will be zero as the pseudo
random sequence seed has not been determined.  (If the trigger class
is "T" and this detector is called "helicity", then these variables
will be T.helicity.hel and T.helicity.helrep.

By default, it is assumed that there is delayed reporting of 8 cycles
and that Quartets are used.  If delayed helicity reporting is not in
use, then set the parameter "helicity_delay" to zero.

  These signals are plus and minus signals, the MPS indicator, and
  the start of quartet indicator.  The time from the TI (with overflow
  detection) is also retrieved as this is needed when dealing with
  delayed helicity reporting.

  The helicity signals are in FADC250 channels.  The roc/slot/channel
  assignments for these signals are currently hardwired into the code.

Add the tree variable goodAdcMult which is the number of raw ADC pulses
  in an ADC channel per event for a given shower array block.

Fix problems in THcTrigDet.cxx

1) Read-in of the trigger reference time cuts was in ReadDatabase method
  when it should be in Init method. Moved to the Init method before
  call to InitHitList where the reference cuts are used.

2) Modified the selection of the good TDC hit for filling the TdcTimeRaw
and TdcTime. Previosly used the TdcTime but not picks the TDC hit with the TdcTimeRaw that is within
the time min and max windows that can be set by parameters. Default windows are min=0 and max=10000.
If two hits in the time window then selects last hit within window.
If not hits within time window then TdcTimeRaw=0 and TdcTime=0.
The TdcMultiplicity is always filled.

1) Read-in of the trigger reference time cuts was in ReadDatabase method
  when it should be in Init method. Moved to the Init method before
  call to InitHitList where the reference cuts are used.

2) Modified the selection of the good TDC hit for filling the TdcTimeRaw
and TdcTime. Previosly used the TdcTime but not picks the TDC hit with the TdcTimeRaw that is within
the time min and max windows that can be set by parameters. Default windows are min=0 and max=10000.
If two hits in the time window then selects last hit within window.
If not hits within time window then TdcTimeRaw=0 and TdcTime=0.
The TdcMultiplicity is always filled.

1) THcTrigDet
   a) Modify code to use string vector fTrigNames to identify which of the trigger
     trigger detector TDC signals to use for coincidence timing. This
      replaces the hard-coded names. The integer array fTrigId[4] keeps
     track of the array ID in the overall fTDcNames vector that goes
      with the fTrigNames[4]
   b) The fTrigNames must be in order of SHMS ROC1, HMS ROC1, SHMS ROC2
     and HMS ROC2.
   c) Default names are "pTRIG1_ROC1","pTRIG4_ROC1","pTRIG1_ROC2","pTRIG4_ROC2"
      The parameter is t_coin_trigNames
      For example:
      t_coin_trigNames="pTRIG1_ROC1 pTRIG4_ROC1 pTRIG1_ROC2 pTRIG4_ROC2"
      This parameter is read-in and the split into a string vector fTrigNames.
   d) In HcTrigDet::ReadDatabase , all members of fTrigID array are set
to -1 then a loop over all Tdc in the trigger detector looks for match between
fTdcNames amd fTrigNames to set fTrigId. If match not found the fTrigId = -1.
   e) In THcTrigDet.h made new method Get_CT_TrigTime(Int_t ii) to be used
in in THcCoinTime to get the 4 trigger times: SHMS ROC1 (ii=0), HMS ROC1 (ii=1),
SHMS ROC2 (ii=2) and HMS ROC2 (ii=3).  If fTrigId[ii]=0 then it returns 0
otherwise returns fTdcTime[fTrigID[ii]].

2) THcCoinTime
  a) Previously used the raw trigger times and had a hard-coded 0.1 ns/channel
conversion. Now use the trigger times which already are in ns using the
ns/channel conversion set in the TrigDet.
  b) Now use the new TrgiDet method Get_CT_TrigTime(ii) to get the
trigger times for SHMS ROC1, HMS ROC1, SHMS ROC2 and HMS ROC2.

Added TClonesArray* fRawHits which an array of THcDCHit object for all
hits in the drift chamber. This is filled during Process method.

Added tree variable raw.wirenum whihc is array of wire numbers for all hits in the drift chamber.

* updated mispointing calculations

* added TMath::Abs() for HMS angles, since it has a '-' sign in the kin. file

Don't confuse index of scalers with slot number

See https://github.com/JeffersonLab/analyzer/commit/cd7c26b5b1161115a0e35a4824148b78a189e986

NOTE:
This changes the format of the "variable" lines in the db map file that
defines the scaler layout.

The number after the variable directive is now the slot number instead of
the index of which scaler module (starting from zero) is being used.
Updated DBASE/db_PScalevt.dat and DBASE/db_HScalevt.dat files which have
been generated by an updated MAPS/SCALERS/make_scaler_db.py.  This change is
made so that the "variable" line format is the same as used in
THaScalerEvtHandler.cxx.

  Determine readout side in THcDriftChamberPlane at initialization time.

instead of using GetParent() every hit to get DC version.

Main change was to allow the beam X and Y positions and angle
at the target to be read in through parameters instead of
always using the EPICS data. This gives added flexibility.

If either of the beam position parameters (gbeam_xoff or
gbeam_yoff) are read-in as parameters, then the code uses
the read-in parameters as the BPM position at target. If
there are other beam position or angle parameter that are
not read in then they are set to zero.


1) Modify THcRaster.h
   a) Added variables for X and Y beam angles at target.
   b) Added variable fFlag_use_EPICS_bpm which is a flag
     that is determines whether the code uses the parameters
     or EPICS reads.

2) Modified THcRaster.cxx
   a) set default BPM z-positions to Fall 2018 survey.
   b) Modified Process method so that the BPM position/angle variables
     that are set by parameters are no longer recalculated. Just
     use the existing variables fXbpm_tar, fYbpm_tar, fXpbpm_tar
      and fYPbpm_tar throughout the method.

Add the variable ebeam_epics to the tree. This is the EPICS
value of HALLC:p  . For now this is good for monitoring changes in the
beam energy but needs to be check how it compares to the Hall C
beam energy measurement.

1) THcDCWire.h
  a) added sigma to THcDCWire constructor which fills fSigmaWire
  b) added GetSimga method
  c) added SetSigma method

2) THcDCHit.h
 a) Added GetWireSigma method

3) THcDriftChamberPlane.h
 a) Added flag fUsingSigmaPerWire
 b) Added array fSigmaWire

4) THcDriftChamberPlane.cxx
 a) In ReadDatabase, set fUsingSigmaPerWire=0 by default
    with reading of parameter h_using_sigma_per_wire optional
    for setting fUsingSigmaPerWire.
 b) If fUsingSigmaPerWire=0 then sets all wires in array fSigmaWire to
   fSigma ( which is the simga per plane).
 c) If fUsingSigmaPerWire=1 then reads in the sigma per wire from
   a parameter file.
 d) Used new THcDCWire constructor to set the sigma for wire to fSigmaWire[nwire]
 e) Also change code so that the tzero offset is set in the  THcDCWire constructor
 f) In ProcessHits , use wire->GetTOffSet instead of the fTzeroWire array when setting
    the time.

5) THcDC.cxx
 a) In TrackFit, replace fSigma with hit->GetWireSigma()

1) add variable fHodoEff_CalEnergy_Cut

2) THcHodoEff::ReadDatabase
   Optionally read in parameter hHodoEff_CalEnergy_Cut (HMS) or
        pHodoEff_CalEnergy_Cut (SHMS) and fill fHodoEff_CalEnergy_Cut
   If no parameter then defaults to fHodoEff_CalEnergy_Cut = 0.050

3) fHodoEff_CalEnergy_Cut
    In "if" statements that select events for hodoscope efficiency
    replay hard-coded "0.05" with fHodoEff_CalEnergy_Cut

  Previously, if load was called with a run number, array continuation
  lines in included files were being interpreted as lists of run numbers.
  Still we have the limitation that in the top file, if we define an array
  with multiple lines, the continuation lines will be interpreted as
  run number lists and mess things up.  But it is unlikely that we would
  need to define arrays in the top file when we are in run number
  mode.

Added setting of spectrometer mispointing according the the input
spectrometer angle if the mispointing is not set by a parameter.

The formulas for deteriming the mispointing comes from fits to sureys.

If one wants to use different spectrometer mispointings then
it is best to set hmispointing_x,hmispointing_y for HMS
and pmispointing_x,pmispointing_y for SHMS in the kinematics setting file
such as standard.kinematics .