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

1) Add to tree variable arrays stub_x,stub_xp,stub_y and stub_yp
    and ncombos. These are arrays of the number of spacepoints in
    the chamber.

2) In ReadDatabase
   a) new variable  fRatio_xpfp_to_xfp which is set differently
       for HMS and SHMS. Used in method LeftRight
   b) Set default value of optional parameter fStubMaxXPDiff = 999.

3)  Modified method LeftRight
   a) Previously only for the old fHMSStyleChambers would
      the code only select LR combinations where
      the difference between stub_xp to space_point_X*ratio is
        with in the fStubMaxXPDiff .
      The ratio came from the HMS optics.
   b) Changed LeftRight so that if fStubMaxXPDiff < 100
         then the code will only select LR combinations where
      the difference between stub_xp to space_point_X*ratio is
        with in the fStubMaxXPDiff . The ratio is
      set in ReadDataBase according to the spectrometer.


5) In FindSpacePoints add some comments.

First, in h_pattern_recognition.f we have:

https://github.com/zahmed9/engine/blob/mkj/HTRACKING/h_pattern_recognition.f#L139
(hdc_wire_center(yy)-hdc_wire_center(yyprime))**2 .lt. (hspace_point_criterion(ich))

So I changed
https://github.com/zahmed9/hcana/blob/develop/src/THcDriftChamber.cxx#L292
TMath::Abs(fHits[plane_hitind]->GetPos() - fHits[planep_hitind]->GetPos()) < fSpacePointCriterion

to

pow((fHits[plane_hitind]->GetPos() - fHits[planep_hitind]->GetPos()),2) < fSpacePointCriterion

Second, the value of MAX_SPACE_POINTS in ENGINE is 100 so in hcana I changed it from 50 to 100.

In tree they are called tnhit and trawhit.
Created two new variables for THcDriftChamber, fNthits and fN_True_RawHits.
In tree they are called tnhit and trawhit.
THcDC defined subdetector THcDriftChamber to have name Ch1 and Ch2 to do used in the tree.
THcDriftChamberPlane defined new method GetNRawhits to return fNRawhits for each plane.
Added member fNRawhits to THcDriftChamberPlane

  Fix from Mark Jones with additional cleanup for
  saw's misunderstanding of space point criterion

  THcDriftChamber::AddPlane didn't know if style was HMS, because
  style was only determined in Init.  Now fHMSStyleFlag is set in
  only one place.  Right now determined from spectrometer name, but
  could be a parameter in the future.

Modified ReadDatabase : Added fdebugstubchisq and Corrected fhdebugflagpr to be Int
Added PrintDecode( void ) to print the drift chamber info in fHits
Modified FindSpacePoints to use the proper PlaneInd and PlanePInd
    for HMS and SOS style chambers
Changed THcDriftChamber::FindEasySpacePoint(Int_t yplane_hitind,Int_t yplanep_hitind) to
THcDriftChamber::FindEasySpacePoint_HMS(Int_t yplane_hitind,Int_t yplanep_hitind)
Added  THcDriftChamber::FindEasySpacePoint_SOS(Int_t xplane_hitind,Int_t xplanep_hitind)
commented out many fhdebugflagpr  output statements

  THcDriftChamber
    New flag fHMSStyleChambers to determine if HMS or SOS style
    tracking should be done.  Flag is set on if the spectrometer
    name begins with 'h'.
    THcDriftChamber::LeftRight fSmallAngleApprox must be set and
    fHMSStyleChambers must be off to do SOS style.
    If not HMS style, ignore all the Yplane optimizations
  THcDC
    SOS code does "project to chamber" in link stubs.  Need to evaluate
    if this is necessary, or if HMS style should also do it.

   {prefix}dc_fix_lr
        Historically, in the ENGINE, if a hit is used in multiple
        space points/stubs, the left/right assignment for that hit, which
        is later used in track fitting, is that assigned for the last stub
        encountered with that hit.  Properly, the left right assignment
        should be allowed to be different in different space points.  If
        this parameter is set to zero (e.g. in hcana.param), then the
        ENGINE behavior is used.  For new analyses, it should be set to one.
        (Which is the default)
   {prefix}dc_fix_propcorr
        When a hit gets put into a stub, the distance of the hit from the
        discriminator can then be estimated.  In the engine, a correction to
        the drift time (and thus drift distance) is applied.  However, if
        that hit ends up in another stub, the correction will get applied again,
        resulting in a over correction.  Setting this flag to 1 will give that
        hit a different corection for each stub that it is in.

These flags will default to the new "correct" way of handling hits if
the above parameters are not set in a parameter file.  Currently, both
flags are set to zero in hcana.param to replicate the ENGINE behavior.

To implement these changes, the propagation correction and L/R information
for each hit is saved in space point and track classes.  This information
is still saved in the hit class, but only used if in ENGINE compatibility
mode.

The THcDCTrack class now saves a list of space point pointers instead
of space point indices.

The AddSpacePoint method now also copies all the hit information into
the track object so that THcDC doesn't need to explicitely copy all
the hits.

The FindStub method, which fits a stub track to a space point is passed
the space point rather than a list of hits

Set fDebugDriftCh in DefineVariables
Put if fDebugDriftCh in front of all cout statements
In THcDriftChamber::DestroyPoorSpacePoints replace line (*fSpacePoints)[isp] = (*fSpacePoints)[osp];
In THcDriftChamber::SelectSpacePoints() replace line (*fSpacePoints)[sp_count] = (*fSpacePoints)[isp];

  Compiles and doesn't crash.

  THcDriftChamber::SpacePoint -> THcSpacePoint
  That way space points can be put in a TClonesArray and be made known
  to THcDC when stub information is added.

  New class THcDC is whole set of wire chambers.
  THcDriftChamber is one chamber (6 planes)
  Implement space point finding from h_pattern_recognition up to point
     before where space point lists from the two chambers are combined.
  Includes
      FindEasySpacePoint   (h_find_easy_space_point)
      FindHardSpacePoints  (find_space_points)
      DestroyPoorSpacePoints (h_sp_destroy)
      SpacePointMultiwire  (h_sp_multiwire)
      ChooseSingleHit 	   (h_choose_single_hit)
      SelectSpacePoint	   (select_space_points)

Space points are arrays of structures in the THcDriftChamber class.  Will
probably need to take space points out class so that they can be looked at
at the THcDC level.

New code is only tested to see that it compiles and that space points
are sometimes found.

      Start adding to the drift chamber code.   Setup a class structure
      similar to what podd uses for the VDCs.  Raw hit class renamed
      to THcRawDCHit.  Signal hits are how THcDCHit, modeled after
      podd hits.  Add THcDCWire, and classes for time to distance conversion.
      Time to distnace conversion doesn't do anything yet.

      Find some improved DC parameter files so that min and max TDC limits
      are good for the run we are using.

      THcDriftChamberPlane asks ThcDriftChamber for many parameters.

be made either in output.def or by analyzing root tree.

TClonesArray ConstructedAt method.

Raw hit class for drift chambers and skeleton drift chamber that prints
out drift chamber package hits.

Move Compare method from THcHodoscopeHit to ThcRawHit since it should
work for any derived class.

Get detector name -> ID mapping from the map file comments

THcHodoscope, a multiplane scintillator array.  Make a Hall
C style spectrometer (so that default Hall A detectors are not
included.)  Can read a Hall C data file and print out hodoscope hits
where the left/right TDC/ADC are matched up per counter.

get the hits.

Going to back out of this and go to multiple inheritance.

Copy the intermediate classes needed for THcHodoscope.  Many
changes needed.