History.h

Go to the documentation of this file.

#ifndef canvas_Persistency_Provenance_History_h
#define canvas_Persistency_Provenance_History_h

//----------------------------------------------------------------------
//
// Class History represents the processing history of a single Event.
// It includes ordered sequences of elements, each of which contains
// information about a specific 'process' through which the Event has
// passed, with earlier processes at the beginning of the sequence.
//
//
//----------------------------------------------------------------------

#include "canvas/Persistency/Provenance/BranchListIndex.h"
#include "canvas/Persistency/Provenance/EventSelectionID.h"
#include "canvas/Persistency/Provenance/ProcessHistoryID.h"

namespace art {
  class History {
  public:
    typedef std::size_t size_type;

    // Compiler-generated default c'tor, copy c'tor, assignment and
    // d'tor are all correct.

    // Return the number of 'processing steps' recorded in this
    // History.
    size_type size() const;

    // Add the given entry to this History. When a new data member is
    // added to the History class, this function should be modified to
    // take an instance of the type of the new data member.
    void addEventSelectionEntry(EventSelectionID const& eventSelection);

    void addBranchListIndexEntry(BranchListIndex const& branchListIndex);

    EventSelectionID const& getEventSelectionID(size_type i) const;

    EventSelectionIDVector const& eventSelectionIDs() const;

    ProcessHistoryID const&
    processHistoryID() const
    {
      return processHistoryID_;
    }

    void
    setProcessHistoryID(ProcessHistoryID const& phid) const
    {
      processHistoryID_ = phid;
    }

    BranchListIndexes const&
    branchListIndexes() const
    {
      return branchListIndexes_;
    }

    BranchListIndexes&
    branchListIndexes()
    {
      return branchListIndexes_;
    }

  private:
    // Note: We could, instead, define a struct that contains the
    // appropriate information for each history entry, and then contain
    // only one data member: a vector of this struct. This might make
    // iteration more convenient. But it would seem to complicate
    // persistence. The current plan is to have parallel vectors, one
    // for each type of item stored as data.
    EventSelectionIDVector eventSelections_;

    BranchListIndexes branchListIndexes_;

    mutable ProcessHistoryID processHistoryID_;
  };
}

#endif /* canvas_Persistency_Provenance_History_h */

// Local Variables:
// mode: c++
// End: