ComponentParameter move from const ptr* to unique_ptr

The main change is

typedef std::pair<std::unique_ptr<Trk::TrackParameters>, double> SimpleComponentParameters;
typedef std::vector<SimpleComponentParameters> SimpleMultiComponentState;
/**
 * Component parameter typedef
 */
typedef std::pair<const TrackParameters*, double> ComponentParameters;//Needed by MultiComponentState
} // end Trk namespace

to

typedef std::pair<std::unique_ptr<Trk::TrackParameters>,double> ComponentParameters;

The rest somehow follow from this ... Although there are naturally many places where this enters.

• @amorley with this change we do not need to have Simple versions for the unique_ptr semantics and this can helps us simplify the ownership chain quite a bit. We can know there is a "unique" owner of non-const things and ownership can be passed via from -> to functions accepting/returning these etc.

• Perhaps we can do one or two final rounds on reducing cloning (i.e we can modify the payload when ownership is passed) and moving a bit more things to unique_ptr or even by-value (e.g return statements)

• For the reviewer , some places are still far from ideal in how they manage things. Especially at the later stages where one communicates with EDM like objects that need / take ownership. I have been trying to use const ptr* to mark these cases. Also there is a bit of release -ing to them . This is seen more in the Smoother and fitter rather in the intermediate steps / calculation where ownership is clear .

On a side note

if (statesToMerge.size() <= m_maximumNumberOfComponents) {
    ATH_MSG_VERBOSE("State is already sufficiently small... no component reduction required");
    MultiComponentStateAssembler::addMultiState( cache, std::move(statesToMerge) );
    return MultiComponentStateAssembler::assembledState(cache);
  }


  if (statesToMerge.empty()) {
    ATH_MSG_ERROR("Attempting to merge multi-state with zero components");
    return nullptr;
  }

the 2nd check was practically doing nothing as if empty the size is already prb < a value greater than 0... Removed it for now ....