MuonChamberHoleRecovery - Ensure that drift circles are unique

For one reason or another, the can be 2 hits for the same Mdt tube in the PrepData collection.

MuPatTrackBuilder.Mu...MuonChamberHoleRecoveryTool    22    21 SUCCESS Two hits recorded for the same tube
*** current: MDT BIS eta -3 phi  6 ml 2 lay 3 tube 14  r  4617 z -2443 phi -1.907 theta 2.057 lPos 14.89 error  0.07  r_drift 14.89
*** previous: MDT BIS eta -3 phi  6 ml 2 lay 3 tube 14  r  4624 z -2430 phi -1.907 theta 2.055 lPos  0.22 error  0.23  r_drift  0.22
MuPatTrackBuilder.Mu...MuonChamberHoleRecoveryTool    30    23 SUCCESS Two hits recorded for the same tube
*** current: MDT BOS eta  1 phi  2 ml 1 lay 1 tube 16  r 10386 z   496 phi 1.267 theta 1.523 lPos 14.89 error  0.07  r_drift 14.89
*** previous: MDT BOS eta  1 phi  2 ml 1 lay 1 tube 16  r 10386 z   483 phi 1.267 theta 1.524 lPos  1.69 error  0.16  r_drift  1.69
MuPatTrackBuilder.Mu...MuonChamberHoleRecoveryTool    30    23 SUCCESS Two hits recorded for the same tube
*** current: MDT BOS eta  1 phi  2 ml 1 lay 3 tube 16  r 10438 z   496 phi 1.267 theta 1.523 lPos 14.89 error  0.07  r_drift 14.89
*** previous: MDT BOS eta  1 phi  2 ml 1 lay 3 tube 16  r 10438 z   487 phi 1.267 theta 1.524 lPos  5.72 error  0.08  r_drift  5.72

If none of the hits is picked-up by the segment maker, the chamber hole recovery puts both back on track, triggering a duplicate hit warning from the GlobalChiFitter. This MR introduces a proper treatment by choosing the hit giving the better residual. I let the decision to @kluit, @stavrop to investigate the underlying reason at the bytestream converter level.