Beta4.10.5-1.txt

                  Geant4 10.5-beta-01 Release Notes
                  ---------------------------------

                                                            29 June 2018

                 Migration Notes & Main New Features
                 -----------------------------------

  o Replaced POSIX threading with C++11 threading, allowing for use of
    multi-threading on Windows for use with either static or dynamic
    libraries.
  o Added track/step/volume information in G4ExceptionHandler when an
    exception happens while event is being processed.
  o Enabled VecGeom wrappers for G4ExtrudedSolid, G4TessellatedSolid and G4Hype.
    Updated to VecGeom version 1.0 for optional solids modeling.
  o Correction in G4MagneticField to no longer inherit from
    G4ElectroMagneticField. Introduced Bulirsch-Stoer method of integration,
    an alternative to Runge-Kutta based on the mid-point method.
  o Improved reporting of looping particle, adding info on density and
    fraction (%) of proposed step that was completed, and momentum.
    Changed default values for thresholds related to the killing of
    looping particles by transportation processes.
  o Use new parameter MscEnergyLimit() (default is 100 MeV) in EM physics
    constructors, allowing to customise transition from low-energy multiple
    scattering to high energy combination of multiple and single scattering.
  o New lateral displacement algorithm in G4UrbanMscModel; faster version
    used by default; more accurate distribution is used for Opt3 EM physics.
  o New gamma conversion model G4BetheHeitler5DModel.
  o New ion ionisation models G4LindhardSorensenIonModel, 
    G4AtimaEnergyLossModel, G4AtimaFluctuations.
  o Revisions of G4BetheHeitlerModel and G4PairProductionRelModel with
    improved screening function approximation, improved LPM function
    approximation, angular generation and efficiency of sampling 
    of the final state.
  o Extended strange pair production channels in Bertini Cascade.
  o Extended upper limit of applicability (from 1 to 100 TeV) for the
    diffuse-elastic processes.
  o New tuning of the FTF model parameters; introduced smearing of delta-isobar
    mass and improved di-quark fragmentation into baryons in Lund string
    fragmentation. First implementation of rotating strings in FTF with
    introduction of Mt distribution of hadrons.
    Improved process probability parameterizations for Pi-Nucleon interactions;
    corrected calculation of nuclear residual excitation energy; tuning of the
    Lund string fragmentation.
  o Major revision of the QGS final-state model (note: this was already
    included in 10.4.beta, but was not included in the final 10.4 release):
    implemented Reggeon cascading and "Fermi motion"; new algorithm for the
    determination of the kinematical properties of partons; improved formation
    of the residual nucleus. Tuning the parameters related to both string
    formation and string fragmentation to improve the description of
    thin-target data.
  o For both string models (FTF and QGS): first implementation of alpha
    cluster structure of carbon nuclei (affecting only hadron-Carbon
    interactions), according to P.Bozek et al., Phys.Rev.C90, 064902 (2014).
  o New class for diffuse-elastic process, where the underlying physics model
    code has been revised and significantly improved in performance.
  o G4LindhardSorensenIonModel, G4LindhardSorensenData: new relativistic ion 
    ionisation model.
  o Introduced switch in cross-sections to allow transuranic elements.
  o Removed unnecessary checks of process manager for generic ion and muonic
    atom in G4SteppingManager.
  o Implemented simple algorithm in visualization to allow colouring of
    detector elements based on materials density.
  o RayTracer now visualizes layered-mass-geometry volumes defined in 
    parallel worlds.
  o Porting of code on gcc-8.1 series.
  o Updated to CLHEP version 2.4.1.0.
  o New data sets G4NEUTRONXS-2.0 and G4EMLOW-7.4.

 ----------------------------------------------------------------------------

                             Technical Notes
                             --------------- 

  o Tested platforms:
    + Linux, gcc-4.8.4.
      Tested on 64 bit architectures (Intel or AMD) with CERN CentOS
      Linux 7 (CC7) (based on CentOS Linux 7).
      Versions of Geant4 have also been compiled successfully on other
      Linux distributions, Ubuntu, Debian, Suse or other RedHat systems.
    + MacOSX 10.13, clang-4.0 (Apple LLVM/Clang-9.1.0)
    + Windows/10 with Visual C++ 14.11 (Visual Studio 2017)

  o More verified configurations:
    + Linux, gcc-4.9.3/5.4.0/6.3.0/7.3.0/8.1.0, clang-3.9/5.0.
    + Linux, Intel-icc 18.0.
    + MacOSX 10.11/12 with clang-3.7/3.9

  o Geant4 10.5-beta-01 has been tested using CLHEP-2.4.1.0, required
    for external installation of the CLHEP library.

 Please refer to the Geant4 User Documentation:
   http://cern.ch/geant4/support/userdocuments.shtml
 for further information about using Geant4.

 ----------------------------------------------------------------------------

List of features and fixes included in this Beta release since 10.4.p02:

  o Configuration:
    -------------
    + CMake:
      o Added "auto" dummy Thread Local Storage model; when "auto" is used,
        no explicit compiler flag is being set to select the TLS model;
        selection is left to compiler default.
      o Added configuration for optional profiling package TiMemory.
      o Skip validation checking for external zlib sources.
      o Added TESSELLATEDSOLID flag to enable optional wrapping of
        G4TessellatedSolid with VecGeom implementation.
      o Added configuration support for MT on Windows; removed restrictions
        and warnings on enabling MT on Windows.
      o Support use of WINDOWS_EXPORT_ALL_SYMBOLS for DLL builds with
        cmake >=3.4 retaining old capability for CMake 3.3.
      o Updated version of data sets: G4NEUTRONXS-2.0, G4EMLOW-7.4.
        Updated tags for 10.5-beta.
    + GNUMake:
      o In geomconf.gmk, added G4GEOM_USE_UTESSELLATEDSOLID flag to enable
        VecGeom wrapper for G4TessellatedSolid.

  o Analysis:
    --------
    + Updated to g4tools 4.2.3: fixed compilation warnings on gcc-8.1.0.
      See History_tools for the complete list of modifications.
    + Fixed a deadlock issue with G4RootPNtupleManager where the mutex
      was locked before being passed to Root.

  o Digits & Hits:
    -------------
    + Updated use of thread-local static variables to function returning
      a thread-local static reference. New model for allowing support of
      multi-threading on Windows DLLs.

  o Event:
    -----
    + Code cleanup in G4StackManager to remove branches for muonic atom
      settings.
    + Updated use of thread-local static variables to function returning
      a thread-local static reference. New model for allowing support of
      multi-threading on Windows DLLs.

  o Externals:
    ---------
    + Updated to Zlib version 1.2.11:
      o Use original files format and adopt C Compiler for installation.
        Addressing compilation warnings on gcc-8.1.0.
    + Updated CLHEP module to version 2.4.1.0:
      o Enabled thread-local storage and atomics on Windows and clang,
        as now all supported by c++11.
      o Ranlux64Engine: fixed compilation warning on Windows for shift count
        negative or too big.
      o Fixed compilation warnings from gcc-8.1.0 when using
        -Woverloaded-virtual compilation option.

  o Geometry:
    --------
    + magneticfield:
      o Correction in G4MagneticField to no longer inherit from
        G4ElectroMagneticField; this allowed a user to provide a full EM field,
        but obtain the force only from the B-field when using the equation
        created in the G4ChordFinder constructor.
        It is no longer possible to use the equation of motion for a pure
        magnetic field for the case of a mixed electromagnetic field. Instead
        the full chain of classes for equation of full EM field, steppers and
        driver must be used for the case of the 'full' EM field.
      o Introduced Bulirsch-Stoer method of integration, an alternative to
        Runge-Kutta based on the mid-point method. Separate driver derived from
        G4VIntegrationDriver and implemented, using specialisation of
        G4IntegrationDriver template class.
      o Change in G4FieldManager::CreateChordFinder() to cope better with case
        where field argument is null.
      o Fix for missing interface method in G4VIntegrationDriver; added method
        RenewStepperAndAdjust() which used to exist in G4MagInt_Driver. This
        method must be defined in those derived driver classes which accept
        a Runge Kutta stepper.
      o Defined default implementation of the method in G4IntegrationDriver
        and G4FSALIntegrationDriver, and implementations in the templated
        classes G4IntegrationDriver and G4FSALIntegrationDriver.
      o Removed static data from G4FSALIntegrationDriver, G4IntegrationDriver
        and G4MagIntegratorDriver and moved to base class as normal data
        members. Fix required for allowing proper symbol exporting on
        Windows DLLs.
      o Use G4Exception warnings instead of streaming to cerr.
      o Added override final for virtual methods in G4UniformMagField.
    + management:
      o G4AffineTransform: added explicit copy constructor, assignment and
        destructor; InverseProduct() has been optimised, now it does 36
        multiplications instead of 45; optimised also for the case of pure
        translation; added methods for inverse transformation:
        InverseTransformPoint(), InverseTransformAxis(), InverseNetRotation()
        and InverseNetTranslation().
        Minor optimisation to multiplication of matrix by vector, to make use
        of temporaries in accessing vector components.
      o G4GeomSplitter: replaced use or realloc()/free() and memcpy() with
        normal allocation/deallocation through G4Allocator. Also addressing
        compilation warnings on gcc-8.1.0.
    + navigation:
      o G4Navigator::ComputeLocalAxis(): removed unnecessary check for
        IsRotated().
      o Replaced use of Inverse() from G4AffineTransform with new dedicated
        methods, to avoid creation of temporaries in G4RegularNavigation,
        G4Navigator and G4ReplicaNavigation.
      o G4PropagatorInField: fix for case of finishing integration in last
        iteration (so a looping particle will not be signalled incorrectly in
        that case). Improved reporting of looping particle, adding info on
        density and fraction (%) of proposed step that was completed, and
        momentum.
      o Use G4Exception instead of streaming to cerr in various classes.
      o Fixed Coverity defect in diagnostic report in G4MultiLevelLocator.
    + solids/Boolean:
      o G4MultiUnion: added protection for normal initialisation in
        DistanceToOutVoxels(). Minor optimisation in DistanceToOut(p,v,..).
    + solids/CSG:
      o Make G4UCons wrapper inheriting from vecgeom::GenericUnplacedCone,
        following the latest changes in VecGeom.
    + solids/specific:
      o Enabled VecGeom wrappers for G4ExtrudedSolid, G4TessellatedSolid
        and G4Hype.
      o Minor cleanup in G4TessellatedSolid::CreatePolyhedron(), no functional
        changes.
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.
    + volumes:
      o Replaced use of Inverse() from G4AffineTransform with new dedicated
        methods, to avoid creation of temporaries in G4TouchableHistory,
        G4PVPlacement and G4PVParameterised.
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.
      o G4PVPlacement: added copy number to G4Exception printing in
        CheckOverlaps().

  o Global:
    ------
    + G4Threading: replaced POSIX threading with C++11 threading
      (std::condition_variable, std::mutex + std::recursive_mutex, std::thread).
      Many of the former macros are now dummy macros (e.g. G4MUTEXINIT,
      G4MUTEX_INITIALIZER, etc.). Updated pin affinity
    + G4AutoLock is now a thin wrapper around std::unique_lock providing
      previous functionality + more features (e.g. try_lock, try_lock_until,
      etc.). Overloaded G4AutoLock member functions in sequential mode to not
      execute locking/unlocking.
    + Added G4RecursiveAutoLock for recursive mutexes.
    + Extended G4TemplateAutoLock to take advantage of the features
      std::unique_lock offers.
    + G4MTBarrier: removed WIN32 specifics, updated G4CONDITIONWAIT to use
      G4AutoLock (std::unique_lock) instead of G4Mutex, removed
      G4MUTEX_INITIALIZER from constructor.
    + Added G4TypeMutex<T> and G4TypeRecursiveMutex<T> function to threading
      for safely obtaining a static mutex that is unique to the template type T.
    + Replaced static G4Mutex in G4Cache<T> with calls to G4TypeMutex<G4Cache<T>>
      function calls to safely initialize mutexes on Windows.
    + Expanded DEFINED_PPC and DEFINED_INTEL in G4FPEDetection.
    + Migrated G4Timer real (wall) time to use std::chrono::high_precision_clock
      from C++11.
    + Removed obsolete MT wrappers for random numbers, no longer necessary.
      Simplified definition of G4Random types in Randomize.hh accordingly.
    + Full support for thread-local storage on clang (also MacOS clang) now
      enabled; updated tls.hh definitions accordingly; removed use of the
      G4USE_STD11 flag, now always enabled by default.
    + Use thread_local keyword on Windows, instead of __declspec(thread) in
      tls.hh. Avoid exporting symbols for Windows DLLs in multi-threaded builds.
    + Removed no longer necessary windefs.hh header.
    + Updated G4cout_p and G4cerr_p to point to std::cout and std::cerr before
      and after library load. G4cout before G4RunManager creation has been
      seen to fail on Windows.
    + Improvement to G4AutoLock in non-MT mode + bug fix (void try_lock() ->
      bool try_lock()) plus error handling within G4AutoLock if static mutex
      is attempted to be used at application cleanup (previously handled in
      G4Cache). Using atomics in G4Cache.
    + Define G4ThisThread, G4Promise, G4Future, and G4SharedFuture in
      G4Threading the promises and futures are not currently used but will be
      used in future.
    + G4TypeMutex and G4RecursiveTypeMutex now accept an index, e.g.
      G4AutoLock l(G4TypeMutex<G4SomeClass>(0)) and
      G4AutoLock l(G4TypeMutex<G4SomeClass>(1))
    + G4DummyThread typedef std::thread::id as id instead of id_type.
    + Updated use of thread-local static variables to function returning
      a thread-local static reference. New model for allowing support of
      multi-threading on Windows DLLs.
    + Added G4TiMemory header which defines an dummy initializer and empty
      macros when the profiling package TiMemory is disabled, while it includes
      the TiMemory headers and defines an initializer when built with TiMemory
      support enabled. Removed obsolete definition in G4Timer and G4SliceTimer,
      causing conflicts with TiMemory. Updated CMake scripts.
    + Increased accuracy of G4Pow::A13 in the most critical [1/4,4] interval.
    + Moved G4Exception functions to inline and extracted banner front/end
      strings for use in Qt parsing.
    + Minor code cleanup in G4UniformRandPool.
    + Changed date for release 10.5-beta.

  o Graphics Representations:
    ------------------------
    + Replaced DBL_MAX by G4VisAttributes::veryLargeTime = 1.e100*ns,
      (i.e. about 1.e75 billion years), to avoid for consistency DBL_MAX
      which is dimensionless.
    + Added G4GREPS_ALLOC_EXPORT flag for DLL symbols export and apply it to
      data in G4VisAttributes.
    + Disabled deprecated static data member G4VisAttributes::Invisible for
      Windows systems.

  o Intercoms
    ---------
    + Updated use of thread-local static variables to function returning
      a thread-local static reference. New model for allowing support of
      multi-threading on Windows DLLs.
    + Fixed compilation warnings for not used return parameters.

  o Interfaces:
    ----------
    + G4Xt: fixed compilation warnings for casts on gcc-8.1.0.
    + Fixed compilation warnings for not used return parameters.

  o Particles:
    ---------
    + Removed obsolete MaxNumberOfProducts enumerator in G4DecayProducts.
    + Removed G4MUTEX_INITIALIZER from constructor in G4VDecayChannel.
    + Updated use of thread-local static variables to function returning
      a thread-local static reference. New model for allowing support of
      multi-threading on Windows DLLs.

  o Persistency:
    -----------
    + GDML:
      o Clear auxiliary map information in G4GDMLReadStructure::Clear().
        Addressing problem report #2064.

  o Physics Lists:
    -------------
    + Constructors:
      o electromagnetic:
        - G4EmStandardPhysics, G4EmStandardPhysics_option1/2/4, G4EmLowEPPhysics,
          G4EmPenelopePhysics, G4EmLivermorePhysics, G4EmStandardPhysicsGS,
          G4EmLivermorePolarizedPhysics: use new parameter MscEnergyLimit()
          (default is 100 MeV) allowing to customise transition from low-energy
          multiple scattering to high energy combination of multiple and single
          scattering.
        - G4EmStandardPhysics_option1/2: use Livermore photo-effect model.
        - G4EmStandardPhysics_option4, G4EmLivermorePhysics,
          G4EmPenelopePhysics: set upper energy limit to 100 TeV.
        - G4EmLowEPPhysics: use new 5D gamma conversion model below 80 GeV,
          use new Lindhard-Sorensen ion ionisation model above 20 MeV/u.
      o gamma_lepto_nuclear:
        - G4EmExtraPhysics, G4EmMessenger: added extra UI commands and modifiers
          to define flags for neutrinos and for neutrino-electron interactions.
        - Removed G4MUTEX_INITIALIZER from constructor in G4EmExtraPhysics.
      o hadron_inelastic:
        - G4HadronInelasticQBBC: use G4ParticleInelasticXS both for neutrons
          and protons.
      o factory:
        - Moved G4_REFERENCE_PHYSCONSTR_FACTORY macro calls out of
          G4PhysicsConstructorRegistry., where they cause an implicit circular
          dependency when building granular libraries; moved into inline source
          where they can be included as necessary in a higher level compilation
          unit. REFERENCE macro calls can be made unique.
        - Allow override of exclusion of G4RegisterPhysicsConstructors
          definitions when using WIN32 by setting USING_STATIC_LIBS.
        - Code cleanup.
    + Lists:
      o Use new G4RegisterPhysicsConstructors inline definitions to ensure that
        for static builds G4PhysicsConstructorRegistry is populated.
      o No longer register ShieldingLEND as a wrapper, but use explicit one.

  o Electromagnetic Processes:
    -------------------------
    + DNA:
      o Updated G4DNAOneStepThermalizationModel and other classes for chemistry.
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.
      o Fixed compilation warnings for not used return parameters.
    + High Energy:
      o G4AnnihiToMuPair: added Coulomb correction to cross-section at threshold.
      o Minor code cleanup in G4eeToHadronsMultiModel, G4eeToHadronsModel and
        G4mplIonisationWithDeltaModel.
    + Muons:
      o G4MuMultipleScattering, G4MuPairProduction, G4MuIonisation,
        G4ePairProduction, G4MuBremsstrahlungModel, G4MuBremsstrahlung,
        G4MuBetheBlochModel: use default destructor.
      o Added minor protection in G4MuPairProductionModel.
      o G4ePairProduction: fixed model name.
    + Polarisation:
      o G4PolarizedAnnihilationModel: implemented sampling of final state
        at-rest.
      o Code cleanup; removed unused headers and methods.
    + Standard:
      o G4GoudsmitSaundersonMscModel: implemented interface method for
        macroscopic first transport cross-section computation per volume
        (used only for testing).
      o G4GSMottCorrection, G4GSPWACorrections: fixed initialization of first
        moment correction factor.
      o G4WentzelOKandVIxSection: fixed form-factor for proton.
        Added Mott corrections to e+-.
      o G4BetheBlochModel: use form-factor for projectile ion as in formulation
        of NIM A 488 (2002) 282.
      o G4LindhardSorensenIonModel, G4LindhardSorensenData: new ion ionisation 
        model for moderate and high energies.
      o G4UrbanMscModel: new lateral displacement algorithm, faster version
        is used as default, more accurate version is used as a second type of 
        parameterisation in all EM physics.
      o G4BetheHeitler5DModel: new 5D model.
      o Added new models, G4eplusTo2GammaOKVIModel and G4eplusTo3GammaOKVIModel.
      o G4AtimaEnergyLossModel, G4AtimaFluctuations: new ion ionisation model.
      o G4ScreeningMottCrossSection: added extra method for more CPU effective 
        computation of the Mott correction.
      o G4BetheHeitlerModel, G4PairProductionRelModel: new versions with
        improved screening function approximation, improved LPM function
        approximation, efficiency, documentation and cleanup.
        Corrected call to selecting target atom in final state sampling.
      o G4PairProductionRelModel, G4BetheHeitlerModel: use G4ModifiedTsai
        angular generator for the final state sampling.
      o G4PairProductionRelModel: fixed computation of cross-section in the
        new version of the high energy model.
      o G4DipBustGenerator, G4ModifiedTsai: implemented new interface for
        sampling of directions of e+e- pair.
      o G4GoudsmitSaundersonTable: fixed gcc-8.1.0 compilation warnings.
      o G4eplusAnnihilation: moved sampling of final state from AtRestDoIt()
        method to SamplingSecondaries()
      o Code cleanup in G4eeToTwoGammaModel, G4BetheHeitlerModel and
        G4NISTStoppingData.
      o G4eBremsstrahlungRelModel, G4PairProductionRelModel, G4PAIySection,
        G4PAIxSection, G4ICRU73QOModel: moved inline method which use static
        constants to source. Improved code formatting.
   + Utils:
      o G4EmCorrections, G4VEmModel, G4ionEffectiveCharge, G4VAtomDeexcitation:
        avoid use of static class members in inline methods; moved long methods
        to source; added thread safe definition of model index for atomic 
        de-excitation.
      o G4EmCalculator: fixed GetCrossSectionPerVolume() method in case of
        multiple scattering process; the proper value of the first transport
        cross-section is retrieved now.
      o G4EmParameters, G4EmParametersMessenger: added 5D gamma conversion
        model parameters; added MscEnergyLimit() and corresponding UI command;
        addressing problem report #2045; added extra DNA UI directory
        and command; improved dump of EM parameters.
      o G4VEmAngularDistribution: added SamplePairDirections() method for
        sampling of e+e- pair production.
      o Added new enumerator G4DNAModelSubType.
      o G4LossTableManager: moved inline run time method to source.
      o Minor code cleanup in G4VMscModel and G4VEmModel.
      o G4VEmProcess: moved few private members to protected.
    + Xrays:
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.

  o Generic Processes:
    -----------------
    - Biasing:
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.
    + Optical:
      o G4OpMieHG: changed GetProperty() to use enum indeces instead of strings.
    + Scoring:
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.
    + Transportation:
      o G4Transportation, G4CoupledTransportation: changed default values of
        two thresholds related to the killing of looping particles: 
        Important_Energy: particles with MORE energy are killed ONLY after
                    multiple steps (fThresholdTrials) marked as 'looping'.
        Warning_Energy: particles with LESS energy are killed SILENTLY.
        Intermediate value: particles between these energies are killed after
                    one 'looping step' (with warning.)
        New values:  Warning_Energy = 1.0 KeV
                     Important_Energy = 1.0 MeV
        NOTES:
        This can happen if a track's propagation in field requires a large
        number of integration steps (default number is 1000, and it is
        controlled by G4PropagatorInField's Get/SetMaxLoopCount() methods).
        Usually the material is vacuum or very low density material and
        particles have either low momentum or momentum with a very small
        component along the direction of the magnetic field (zero or nearly
        zero).
        Improved reporting of the issue of dropped particles (above 'warning
        threshold'): now providing the type of particle, position, momentum,
        material density and step length in addition to the energy; added
        recommendations for users along with the warning (printed only for
        the first 5 warnings).
        Addressing problem report #2063.
      o New class G4TransportationLogger to consolidate reports from the
        Transportation processes.
      o Fixed Coverity defects.

  o Hadronic Processes:
    ------------------
    + cross_sections
      o G4NeutrinoElectronTotXsc: new class for total (i.e. neutral current
        plus charged current) cross-section for neutrino-electron interactions.
      o Introduced switch to allow transuranic elements.
      o G4NeutrinoElectronCcXsc, G4NeutrinoElectronNcXsc: added method.
      o G4PiNuclearCrossSection: for transuranic elements, scale the
        cross-sections from those of Uranium (using the ratio of atomic
        weights), instead of using directly the latter.
      o G4NeutronElasticXS, G4NeutronInelasticXS, G4NeutronCaptureXS:
        adopting new data structure from G4NEUTRONXS-2.0 data set,
        incompatible with previous.
      o G4ParticleInelasticXS: new inelastic cross sections for neutrons, 
        protons and light ions based on G4NEUTRONXS-2.0. 
      o G4HadronXSDataTable: new class for cross-section handling.
      o G4CrossSectionDataStore: added new method ComputeCrossSection(); reduced
        number of type conversions, number of computation of cross-sections and
        code branches; use const pointers to G4Material, G4Element and 
        G4Isotope; fixed selection of isotope, for energy below 20 MeV use
        isotope cross-sections, above 20 MeV - isotope abundaces.
      o G4VCrossSectionDataSet: added new flag 'isForAllAtomsAndEnergies' and 
        get/set methods allowing to reduce the list of cross-sections per
        particle.
      o Commented out debug printouts in G4ProjectileFragmentCrossSection
        header.
      o Replaced file-scoped mutexes in G4CrossSectionFactoryRegistry to use
        G4TypeMutex<G4CrossSectionFactoryRegistry> which addresses a static
        mutex initialization bug on Windows.
    + management
      o G4HadronicProcess: fixed compilation warnings on gcc-8.1.0 regarding
        catching polymorphic type by value, instead of by reference.
        Use const pointers and "nullptr" for G4Material and G4Element; use
        ComputeCrossSection() method; removed creation of new materials at run-
        time inside GetElementCrossSection(), all materials should be created
        in the user code.
        Use correct return type when calling CheckResult() from PostStepDoIt():
        fixes case when the value of pointer 'result' does not get properly
        reset to NULL in case a non-conservation is detected, thus "bad"
        interaction does not get re-sampled.
      o Code cleanup in G4HadronicProcess, G4VLeadingParticleBiasing and
        G4HadLeadBias. G4HadronicProcess, G4HadronInelasticProcess: added
        keyword  "override"; cleanup of cross-section scale factor and weights;
        G4HadronicProcessStore: use "nullptr" and added flag 'buildXSTable'
        with get/set methods.
    + models/abla
      o Updated some parameters; code cleanup.
    + models/cascade
      o Added strange pair production channels to list of 6, 7, 8 and 9-body 
        final states in classes G4CascadeT31piNChannel and
        G4CascadeT33piNChannel. Modified template values accordingly.
      o G4CascadeT11pizNChannel: extended strange pair production channels to
        6,7,8 and 9 body final states.
      o Fixes in G4ElementaryParticleCollider methods generateSCMpionAbsorption()
        and generateSCMmuonAbsorption(), as they did not produce correct nucleon
        pairs when pion or muon was absorbed. Added quantum number conservation
        test added to method collide().
      o Fixed implementation of nuclear_rad_2par configurable key in
        G4CascadeParamMessenger.
    + models/coherent_elastic
      o Update to G4hhElastic to take into account recent TOTEM data for
        proton-proton at 13 TeV.
      o Added new class G4DiffuseElasticV2. The underlying physics model is
        the same as in G4DiffuseElastic, however, the code has been revised,
        cleanup and significantly improved in a number of places.
        The performance has been improved by a factor of ~2 with respect
        to the original implementation. The approximation of small angles
        has been dropped (replaced by exact calculation) as it was not
        bringing any visible benefits.
      o Set to 100 TeV (instead of 1 TeV) the upper limit of applicability
        for the model G4DiffuseElastic and G4DiffuseElasticV2 (the number of
        equally-sized logarithmic energy bins has been increased from 200 to
        250 to keep the same bin size, 0.04).
        The model has been validated with experimental data up to 300 GeV,
        but it should not give meaningless results well above this energy.
    + models/de_excitation
      o G4ChatterjeeCrossSection, G4KalbachCrossSection: moved implementation
        to source.
      o G4VEmissionProbability, G4EvaporationProbability: fixed integration 
        of probabilities for neutron emission for OPT=1,2.
      o G4DeexPrecoParameters: fixed checks on parameter values.
    + models/lepto_nuclear
      o G4NeutrinoElectronCcModel: added extra interfaces.
        Also fixed Coverity defect.
    + models/particle_hp
      o G4ENDFTapeRead, G4FissionProductYieldDist: fixed compilation warnings
        on gcc 8.1.0, regarding catching polymorphic type by value, instead of
        by reference.
    + models/parton_string
      o In the string fragmentation part of FTF, first implementation of
        rotating strings with introduction of Mt distribution of hadrons.
        Parameters are tuned using NA49 experimental data on pp-interactions.
        Good results are obtained for Pi+-, K+-.
        For P and Pbar <Pt> - xF correlations need still improvement.
      o Major revision of QGS model; implemented Reggeon cascading and
        associated "Fermi motion". Implemented new algorithm of parton
        kinematical properties determination and calculation of residual
        nucleus properties. Adjusted G4GammaParticipants to the revised
        model.
      o Improved energy-momentum conservation in the QGS model, especially
        in the case of diffraction interactions.
      o Major improvement to QGSM: Pomeron parameters are set up according to
        A.Kaidalov and M.Poghosyan (2009-2010). Non-vacuum reggeons: f and
        omega, are included in the calculations using parameters by A.Kaidalov
        and M.Poghosyan (2009-2010).
      o New class G4Reggeons for calculations of X-sections.
      o Cut non-vacuum reggeons are interpreted as a creation of a single
        string, or as a single string creation and one "stable" hadron.
        Such processes are described by RRR diagram of Reggeon phenomenology.
      o Introduced new class G4QuarkExchange for implementation of the
        processes.
      o G4QGSParticipants: implementation of Quark exchange simulation.
        Fix for nucleon state counter. Minor fix to avoid FPE observed only
        in optimized mode.
      o G4VPartonStringModel: throw exception for the case when string
        fragmentation does not produce hadrons.
      o Cleaned and improved G4SingleDiffractiveExcitation and
        G4QGSDiffractiveExcitation. Correction in G4SingleDiffractiveExcitation
        to treat h+N and Gamma + N diffraction.
      o Changed G4MesonSplitter probabilities UUbar DDbar (0.5, 0.5) to
        (0.8, 0.2) for gamma projectile.
      o G4QGSMSplitableHadron: fixed bug in the case of gamma projectile with
        P_minus = 0.
      o G4ExcitedStringDecay: bug-fix in FragmentStrings() method (access
        to the first element of an empty vector). Increased the number of max
        attempts to fragment strings from 10 to 100.
      o New tuning of FTF model parameters.
      o Introduced various <Pt^2> for diffractive and non-diffractive
        interactions in G4FTFModel, for NA49 and NA61/SHINE experimental data.
      o Fine tuning in G4DiffractiveExcitation for the FTF model, to describe
        NA49 experimental data, especially, <Pt> on Xf.
      o Fix to address the baryon puzzle of NA49 data; <Pt> of baryons 1/2 at a
        string fragmentation setting up at 435 MeV/c, and <Pt> for baryons 3/2
        production at 900 MeV/c.
      o Fine tuning of G4LundStringFragmentation. <Pt> of mesons created at 
        quark fragmentation is decreased. <Pt> of baryons created at quark and
        di-quark is increased. Corrected fragmentation functions of qq-> B(1/2)
        and B(3/2).
        Re-organized and cleaned up code ; the variant is tuned using NA49
        experimental data on pp->p/pi/K +X.
      o Introduced smearing of delta-isobar mass; delta-isobars are now treated
        as a kinetic track; improved parameterisation of the quark exchange
        process with excitation.
        NA61/SHINE exp. data on Pi+, Pi-, K+, K- and protons in PP interactions
        at 20, 31, 40, 80 and 158 GeV/c are now described. Now reproducing the
        ratio of <Strane Q>/<Normal Q> in PP interactions. Not describing kink
        in AA interactions.
      o Improved di-quark fragmentation into baryons in Lund string
        fragmentation. Improved description of NA49 and NA61/SHINE exp. data
        on p+p->p+X. Improvements in method SetMinimalStringMass(), relevant
        for Pbar P annihilation.
      o Improved process probability parameterizations for Pi - Nucleon
        interactions. Good results are obtained for HARP data (p/Pi+/Pi- + A at
        3, 5, 8, 12 GeV/c) and for Pi N inclusive reactions.
      o Corrected calculation of nuclear residual excitation energy.
        Fine tuning of the model parameters.
      o Set to false the switches for the A-dependence in nuclear destruction
        for baryon projectile in G4FTFParameters. Uncommented checks for
        these switches in the code. No changes in the random number sequence.
      o Improved versions and tuning of Lund and QGS fragmentations
        (used, respectively, by FTF and QGS string models).
      o G4LundStringFragmentation: bug fixes related to rotating strings.
      o G4VLongitudinalStringDecay: removed useless caching in the method
        FragmentationMass() which was causing rare irreproducibilities in
        gamma-nuclear reactions (i.e. involving QGS).
      o Improved QGSM and fixed issue of event reproducibility in
        gammma-nuclear.
      o Fixed leak in G4QGSParticipants::GetList() for targets not deleted
        after interaction.
      o G4QGSParticipants: introduced protection against negative squared mass
        in method PutOnMassShell(), which was causing some rare crashes.
        If a negative squared mass occurs, a "JustWarning" exception is thrown
        and the squared mass is set to 0, before calling the square root (which
        is now computed only once).
      o Moved G4PomeronCrossSection class to 'qgsm' module.
      o G4FTFAnnihilation: fix to get flat cos(theta) and phi distributions for
        antiproton annihilation at rest. Added also a protection for the
        evaluation of rapidity.
      o G4FTFModel, G4DiffractiveExcitation, G4FTFAnnihilation: code
        improvements. No changes in physics output, but with possible
        differences in the random sequence, due to numerical effects caused
        by re-ordering of operations.
    + models/pre_equilibrium
      o G4PreCompoundEmission, G4PreCompoundFragment, G4PreCompoundNucleon:
        fixed computations for OPT=1,2.
      o G4PreCompoundModel: fixed initialisation.
      o G4PreCompoundFragmentVector: moved inlined methods to source.
    + models/quasi_elastic
      o G4QuasiElRatios: removed unnecessary protection against heavy elements.
    + models/radioactive_decay
      o G4ECCapture, G4RadioactiveDecay, G4RadioactiveDecayMode: added electron
        capture channel from the N Shell.
      o G4ECCapture: added tables of PL2/PL1 PM2/PM1 and PN2/PN1 based on bound
        electrons radial wave amplitudes, taken from Bambynek et al., Rev. Mod.
        Phys., vol. 49, 1977.
        Compute partial probabilities of subshells PL1,PL2,PM1,PM2,PN1,PN2 from 
        the tables PL2/PL1, PM2/PM1, and PN2/PN1. Valid for allowed transitions.
    + models/rpg
      o G4RPGAntiKZeroInelastic, G4RPGInelastic: fixed compilation warnings on
        gcc-8.1.0, regarding catching polymorphic type by value, instead of by
        reference.
    + models/util
      o First implementation of alpha cluster structure of carbon nuclei in
        G4Fancy3DNucleus, according to P.Bozek et al., Phys.Rev.C90, 064902
        (2014).
        Fixed issue of event reproducibility; added loop checks and general
        cleanup. In case of difficulties in the method ChooseFermiMomenta(),
        emit a "JustWarning" exception instead of writing to cerr.
      o Minor improvement to code in G4Nucleon and cleanup.
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.
    + processes
      o G4HadronElasticProcess, G4NeutrinoElectronProcess: fixed compilation
        warnings on gcc 8.1.0, regarding catching polymorphic type by value,
        instead of by reference.
        Use const pointers and "nullptr" for G4Material and G4Element.
      o Updated G4NeutrinoElectronProcess; modified constructor.
    + stopping
      o G4HadronStoppingProcess, G4MuonicAtomDecay: fixed compilation warnings
        on gcc 8.1.0, regarding catching polymorphic type by value, instead of
        by reference.
    + util
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.

  o Run
    ---
    + Replaced pthread-style <void*(void*)> function signatures with
      <void(G4WorkerThread*)> as it is now unnecessary
    + Removed WIN32 conditional for G4THREADCREATE in
      G4UserWorkerThreadInitialization, no longer needed.
    + Changed G4Thread to G4NativeThread when setting pin affinity in
      G4WorkerThread.
    + Updated use of thread-local static variables to function returning
      a thread-local static reference. New model for allowing support of
      multi-threading on Windows DLLs.
    + Added TiMemory auto-timer macros to some key functions in G4RunManager,
      G4RunManagerKernel, G4MTRunManager, and G4MTRunManagerKernel
    + Added TiMemory libraries and includes to CMake script for optional use.

  o Track & Tracking
    ----------------
    + G4SteppingManager: remove unnecessary checks of process manager for
      generic ion and muonic atom.
    + Updated use of thread-local static variables to function returning
      a thread-local static reference. New model for allowing support of
      multi-threading on Windows DLLs.

  o Visualization:
    -------------
    + Fixed compilation warnings on gcc-8.1.0.
    + management:
      o Implemented UI command /vis/viewer/set/colourByDensity: simple
        algorithm to allow colouring of detector elements based on materials
        density.
      o Introduced UI command "/vis/drawLogicalVolume", synonymous with
        "/vis/specify". Both commands create a new scene and invoke
        "/vis/scene/add/logicalVolume". Include overlap checking.
      o Moved time window parameters from G4OpenGLStoredViewer to
        G4ViewParameters. Now one can interpolate time window parameters (using
        /vis/viewer/interpolate) and make a movie of particles moving through
        time. To do this, see guidance on /vis/viewer/set/timeWindow/ commands.
      o G4ViewParameters: fix for /vis/viewer/save.
        Completed PrintDifferences(), operator<<() and operator!=() for time
        window parameters. Provide proper Get and Set functions for time
        parameters.
      o Moved time window commands to /vis/viewer/set/timeWindow/.
        Deprecated the equivalent /vis/ogl/set/ commands.
        Copied /vis/ogl/set time window commands to /vis/viewer/set.
      o G4VSceneHandler: reduced repetitive error messages.
      o Improved some UI commands.
    + modeling:
      o Implemented UI command /vis/viewer/set/colourByDensity.
      o G4LogicalVolumeModel: added overlap checking.
        Fixed Coverity warning for unnecessary check on pointer.
    + OpenGL:
      o Implemented G4OpenGLStoredViewer::AddPrimitiveForASingleFrame().
      o Updated G4CONDITIONWAIT(...) to new C++11 requirements.
      o G4OpenGLQtViewer: replaced explicit usage of G4Mutex locking/unlocking
        with G4AutoLock pointers initialized with std::defer_lock as this
        enables one to avoid any deadlock or non-locked condition_variable
        waits by using G4AutoLock::owns_mutex().
    + OpenInventor:
      o Implemented kernel visit for "colour by density".
    + RayTracer:
      o Use HyperStep of G4ParallelWorldProcess to draw volumes defined in a
        parallel world.
      o Updated use of thread-local static variables to function returning
        a thread-local static reference. New model for allowing support of
        multi-threading on Windows DLLs.

  o Data Sets:
    ---------
    + G4EMLOW-7.4:
      o Updated microelec elastic cross-section files.
    + G4NEUTRONXS-2.0:
      o Updated data from G4NDL-4.5 for neutrons and from G4TENDL-1.3.2 for
        p, d, t, He3, He4. New data structure.

  o Examples:
    --------
    + Updated reference outputs, macros, READMEs and scripts.
    + Updated vis/UI setup. Removed no longer needed G4VIS/UI_USE.
    + Added inclusion of G4Types.hh in files for self-consistency.
    + advanced/composite_calorimeter
      o Created ActionInitialization class, in preparation for MT migration.
      o Removed CCalAnalysis class and employ g4analysis only.
      o Fixed compilation warning on gcc-8.1.0.
    + advanced/gammaray_telescope
      o Fixed issues reported by Valgrind.
        Addressing problem report #1981.
      o Fixed compilation warning on gcc-8.1.0.
    + advanced/hadrontherapy
      o Replaced obsolete use if G4VisAttributes::Invisible with
        G4VisAttributes::GetInvisible().
    + basic/B2
      o Modified default input macro exampleB2.in to reduce the output.
    + basic/B4
      o Replaced deprecated "/vis/ogl/set/" with "/vis/viewer/set/timeWindow/"
        command in visualization macros.
    + extended/electromagnetic/TestEm2
      o DetectorConstruction: improved method names, do not allow to clear
        geometry between runs.
      o StepMax: use thread safe instantiation, messenger merged with 
        PhysicsListMessenger.
      o Use MixMax default random number generator.
      o PhysicsList: make Opt0 default, removed local instantiation of G4Decay.
      o PhysListEm5DStandard: experimental PhysList with the new 5D gamma
        conversion model.
      o Corrected input macros to optimise output.
    + extended/electromagnetic/TestEm3
      o DetectorConstruction: improved method names, do not allow to clear
        geometry between runs.
      o StepMax: use thread safe instantiation, messenger merged with 
        PhysicsListMessenger.
      o Use MixMax default random number generator.
      o Removed radioactive.mac input macro,
      o PhysicsList: make Opt0 default, removed local instantiation of G4Decay.
      o Corrected input macros to optimise output.
    + extended/electromagnetic/TestEm5
      o PhysListEm5DStandard: experimental PhysList with the new 5D gamma
        conversion model.
      o Added test of new ion ionisation model, which is included in the new
        physics constructor PhysListEm19DStandard; added atima.mac macro file
        to compare results with known data.
      o StepMax: use thread safe instantiation, messenger merged with 
        PhysicsListMessenger.
      o Use MixMax default random number generator.
      o Removed obsloete class PhysListEmStandardSSM.
      o PhysicsList: make Opt4 default, removed local instantiation of G4Decay.
      o Corrected input macros to optimise output.
    + extended/electromagnetic/TestEm7
      o SteppingAction: added edep at end of step for photo-electric.
      o Fixed compilation warning on gcc-8.1.0.
    + extended/electromagnetic/TestEm8
      o Fix for thread safety in PrimaryGeneratorAction.
      o Disabled tracking cuts in input macro TestEm8.in.
      o Corrected input macros to optimise output.
    + extended/electromagnetic/TestEm11
      o SteppingAction: added edep at end of step for photo-electric.
    + extended/electromagnetic/TestEm15
      o Extended with gamma -> e+ e- 5D conversion code.
      o PhysicsList: added G4EmStandardPhysicsWVI, GS, SS.
      o Use G4RotationMatrix to convert to gamma reference system.
      o Use default random number generator.
    + extended/electromagnetic/TestEm18
      o Updated PhysListEmStandard, livermore, penelope.
    + extended/errorpropagation
      o Use implicit destructor in ExErrorMagneticField.
    + extended/exoticphysics/monopole
      o Changed G4MonopoleEquation to use G4MagneticField instead of
        not use G4ElectroMagneticField, as equations of motion do not use
        electric field.
    + extended/exoticphysics/phonon
      o Update references for caustic and timing.
    + extended/field/BlineTracer
      o Fixed compilation error for missing header inclusion.
    + extended/hadronic/Hadr00
      o DetectorConstruction: cleanup geometry.
      o Use FTFP_BERT as default physics-list.
    + extended/hadronic/Hadr01
      o Cleaned up geometry definition.
    + extended/hadronic/Hadr07
      o SteppingAction: added edep at end of step for photo-electric.
    + extended/medical/DICOM
      o Enabled CMake package configuration. Added DICOM_USE_DCMTK and
        DICOM_USE_HEAD options backwards compatible with environment variables.
      o Updated READMEs.
    + extended/medical/DICOM2
      o New example demonstrating how to inherit from DICOM example.
    + extended/medical/dna/chem1
    + extended/medical/dna/chem2
    + extended/medical/dna/chem3
    + extended/medical/dna/chem4
    + extended/medical/dna/mfp
    + extended/medical/dna/neuron
      o Use G4RegisterPhysicsConstructors inline definitions to ensure
        G4PhysicsConstructorRegisty is filled when using static libraries.
    + extended/medical/dna/pdb4dna
      o Use G4RegisterPhysicsConstructors inline definitions to ensure
        G4PhysicsConstructorRegisty is filled when using static libraries.
      o Fixed analysis code.
    + extended/medical/dna/slowing
      o Updated input script and code cleanup.
    + extended/medical/dna/spower
      o Use G4RegisterPhysicsConstructors inline definitions to ensure
        G4PhysicsConstructorRegisty is filled when using static libraries.
    + extended/medical/fanoCavity
      o ActionInitialization: added printout of Run::EndOfRun().
    + extended/optical/LXe
      o LXeDetectorConstruction: moved DefineMaterials() to constructor.
      o Replaced local physics configuration with FTFP_BERT + G4OpticalPhysics
        Deleted LXeEMPhysics, LXeGeneralPhysics, LXeMuonPhysics and
        LXePhysicsList.
      o Removed class LXeSteppingVerbose.
      o In input macros, removed /LXe/detector/update and added /run/initialize.
        General cleanup.
      o Addressing problem report #2042.
      o Replaced local physics with FTFP_BERT + G4OpticalPhysics; deleted
        custom lists LXeEMPhysics, LXeGeneralPhysics, LXeMuonPhysics and
        LXePhysicsList; use G4EmStandard_option4 EM physics.
      o Added LXeRun to record and print results at the end.
      o Removed LXeSteppingVerbose and cleanup of macros.
      o Removed LXeUserEventInformation, use LXeEventAction instead;
        removed LXeRecorderBase, replaced with LXeHistoManager.
    + extended/optical/OpNovice2
      o New example OpNovice2 for investigation of optical properties and
        parameters. Details of optical photon boundary interactions on a surface
        are recorded, as well as the details of optical photon generation and
        transport.
    + extended/parallel/ThreadsafeScorers
      o TSPhysicsLists no longer declares G4VPhysicsConstructors in header.
      o Added TiMemory timing to TSPrimaryGeneratorAction and G4Run.
      o Added TiMemory ASCII output and serialization to main().
      o Fixed shadowing warnings in G4atomic.hh header.
      o Removed G4MUTEX_INITIALIZER from constructors of
        G4TAtomicHits{Collection,Map}.
    + extended/persistency/gdml/G02
    + extended/persistency/gdml/G03
      o RunAction: removed visualisation command.
    + extended/persistency/P02
      o Updated selection.xml file with new members of G4MaterialPropertiesTable.
      o Fixed crash on reading due to duplication of materials in .root file.
        Addressing problem report #2065.
    + extended/physicslists/shared
      o Copy script works whether include/src directories exist or not.
    + extended/radioactivedecay/rdecay02
      o PhysicsList: in constructor, set nuclide table half-life threshold to
        0.1 picosecond and level tolerance to 1.0 eV, to not miss short-lived
        nuclides
      o HistoManager, TrackingAction: added two histograms to look at 0-0.1 MeV,
        and 0-1 MeV ranges in particle emission spectra.
      o Input macros: updated to latest UserEvapData_z83.a210,
        UserRadData_z82.a210, UserEvapData_z102.a252 database formats.
        Removed energy unit (keV) from call to /gun/ion in u238b.mac,
        po201m.mac and UserEvapDataBiTest.mac.

 ----------------------------------------------------------------------------