source: trunk/ReleaseNotes/Patch4.8.2-1.txt @ 1350

Last change on this file since 1350 was 1337, checked in by garnier, 14 years ago

tag geant4.9.4 beta 1 + modifs locales

File size: 9.2 KB
Line 
1
2                 Geant4 8.2 - patch-01 Release Notes
3                 -----------------------------------
4
5                                                           23 February 2007
6
7List of fixes included in this public patch since the public release 8.2:
8
9  o Geometry:
10    --------
11    + management:
12      o Added flag to G4VPhysicalVolume::CheckOverlaps() method for
13        controlling verbosity. By default, verbosity is ON.
14    + navigation:
15      o In G4Navigator, forcing call to CheckOverlaps() on the current
16        volume before event abortion for cases of stuck tracks.
17      o G4PropagatorInField: implemented ClearPropagatorState()
18        to clear track-dependent state (safety, zero-steps, loop flag. etc...).
19        Call to this method becomes mandatory in G4Transportation for each new
20        track to overcome side effects of safety inconsistency in presence of
21        geometry overlaps.
22    + solids/specific:
23      o Do not scale reducible-polygon in G4Polyhedra::Reset() method.
24        The 'original_parameters' structure offered as input through the
25        accessors are always assumed to be already scaled by the
26        'cos(0.5*phiTotal/theNumSide)' factor.
27        Addresses problem report #873.
28      o Fix in G4PolyconeSide::DistanceAway() for correct estimation of
29        distance for the case of Phi sections.
30        Addresses problem report #598.
31    + volumes:
32      o Added flag for controlling verbosity to CheckOverlaps() method for
33        G4PVPlacement and G4PVParameterised. By default, verbosity is ON.
34
35  o Global:
36    ------
37    + Changed date for patch-01 to release 8.2.
38
39  o Electromagnetic processes:
40    -------------------------
41    + standard:
42      o G4UrbanMscModel:
43        - Fix in the single scattering code for heavy particles, now the code
44          can be used with skin=1 e+/e- and for heavy particles as well.
45        - Removed unecessary call to G4Navigator location method.
46        - Added protection inside lateral displacement sampling against cases
47          of TrueStepLength < GeomStepLength.
48        - Added protections against NaN values in two places: sampling of
49          theta when tau=0 in lateral displacement, and in method
50          ComputeTrueStepLength() for the case of zero true length.
51      o G4MultipleScattering: fix to allow for reinitialisation of model
52        parameters between runs.
53      o Use SetEmModel(index) and added SetFluctModel() from
54        G4VEnergyLossProcess in eIonisation, hIonisation, ionIonisation,
55        eBremstrahlung.
56    + utils:
57      o Fixed a problem happening in interactive sessions when activating or
58        deactivating processes; the problem resulting in building wrong dEdx
59        tables for e+- and muons. Modified classes are: G4LossTableBuilder,
60        G4LossTableManager and G4VEmEnergyLossProcess.
61        New tables are added for e+- and muons ionisation additionally to
62        total energy loss tables.
63
64  o Hadronic Processes:
65    ------------------
66    + models/binary_cascade:
67      o G4BinaryCascade::Capture(): added protection against division by zero.
68    + models/chiral_inv_phase_space:
69      o Fixed cases of division by 0 in G4Quasmon and G4QNucleus.
70        and other cases of generation of NaN.
71        Reduces segmentation faults in CHIPS-based models and physics lists
72        such as QGSC.
73      o Fixed bug of generation of NaN in G4QElastic observed in complicated
74        event for the multi-projectile case.
75        Reduces segmentation faults in elastic scattering.
76    + models/de_excitation:
77      o Evaporation: added protection to exclude negative probabilities from
78        the "competitive channel sum". Together with the fix below for fission
79        this should only happen due to a precision error, but is protected
80        nonetheless.
81      o Fission: fixed probability to no longer be negative. Is now equivalent
82        to what is written in the Physics Reference Manual.
83        The above two fixes also affect the precompound model which invokes
84        these models.
85      o Fission: fixed several memory leaks in G4ParaFissionModel.
86    + models/low_energy:
87      o G4LElastic: added protection against sqrt() of negative value.
88    + models/management:
89      o For error reporting, added model name arguments and default names to
90        constructors for classes G4HadronicInteraction, G4InelasticInteraction,
91        G4VIntraNuclearTransportModel and G4VPreCompoundModel.
92    + models/neutron_hp:
93      o Fixed potential divisions by zero in: G4NeutronHPVector,
94        G4NeutronHPLevel and G4NeutronHPInterpolator.
95    + models/parton_string/hadronization:
96      o G4ExcitedStringDecay::EnergyAndMomentumCorrector(): fixed bug which
97        did stop correction even if not yet converged; check on convergence
98        must check that scale-1 is close to 0, and not less than 0.
99    + models/parton_string/management:
100      o Debugged method Dump() added in G4InteractionContent to print object.
101      o Debugged output in G4VPartonStringModel checking E-p conservation of
102        string model.
103        The above two points provide improved, more detailed verbose output.
104    + models/parton_string/qgsm:
105      o Corrected E-p non-conservation in QGS. In G4QGSMSplitableHadron,
106        the smaller of the lightcone momenta Q+/Q- was ignored.
107      o Fixed division by 0 in G4QGSMSplitableHadron::SampleX().
108      o Added debugging output to several classes.
109    + models/util:
110      o Added method to G4Parton for setting momentum using both + and -
111        momentum component. Should help in improving shower shapes.
112      o G4Fragment: more info on warning output of negative excitation energy.
113    + processes:
114      o Added process names to G4HadronElasticProcess, G4HadronCaptureProcess,
115        G4HadronFissionProcess.  This will allow more detailed error reporting.
116      o Cleaned up commented-out code.
117    + stopping:
118      o G4MuMinusCaptureCascade: added sampling of random vector of e- in
119        decay; added factor 2 in potential energy - fixing decay spectrum.
120      o G4MuonMinusCaptureAtRest: fixed problem for the case when isotope
121        vector is undefined. Force to use round 'targetA'.
122        Addressing problem report #926.
123    + util:
124      o G4HadFinalState: corrected initialisation of energy deposition.
125        Fix affects elastically scattered neutrons and probably most
126        hadronic models. Reduces the number of events in which the summed
127        energy of final particles exceeds that of the incident particle.
128
129  o Materials:
130    ---------
131    + G4Material: fill fAtomVector in the case when material is built
132      from element mass fractions, allowing for low-energy package to
133      use NIST G4_WATER.
134    + G4NistMaterialBuilder: added liquids O2 and N2.
135    + G4NistElementBuilder: added protection for Z<=101.
136
137  o Physics Lists:
138    -------------
139    + Fix in QGSP_BIC_HP list: activated use of HP elastic.
140    + G4HadronElasticPhysics: set limits on low energy and QElastic to zero.
141    + Fixed special cuts for LBE physics list.
142    + Added set min/max Energy and Inelastic Energy to NeutronHP builder.
143    + G4EmStandardPhysics: use G4hMultipleScattering for muons and removed
144      extra printouts. Use default multiple-scattering setup in
145      G4EmStandardPhysics72.
146
147  o Run:
148    ---
149    + Fix in method G4RunManager::NumberOfEventToBeStored().
150      Addressing problem report #925.
151    + Added protection against missing particle types before
152      BuildPhysicsTable() is invoked. Addressing problem report #923.
153
154  o Transportation:
155    --------------
156    + G4Transportation: calling PropagatorInField to clear state for new
157      track. This addresses a reliabilitiy problem observed by ATLAS; the
158      fix is related to overlaps in the geometry, and consists in clearing
159      the safety value, which may become multivalued due to overlaps in the
160      geometry, and thus can become invalid for a new track in an overlap
161      region...
162    + Added protection against zero 'velocity' in method
163      G4Transportation::AlongStep().
164
165  o Configuration:
166    -------------
167    + Updated Configure script to prompt for new version of G4EMLOW data set.
168
169  o Data Sets:
170    ---------
171    + New data set G4EMLOW4.2: correction to Auger data.
172
173  o Examples:
174    --------
175    + Updated reference outputs.
176    + advanced/underground_physics
177      o Removed uninitialized 'UserLimitsForXenon'.
178    + extended/extended/TestEm7
179      o Cleanup of macros.
180    + extended/electromagnetic/TestEm16
181      o Added HistoManager and HistoMessenger and commands to modify
182        upper step limit. Minor fix to GNUmakefile.
183    + extended/radioactivedecay/exrdm
184      o Fixed usage of physics-lists and histograms
185    + extended/runAndEvent/RE02
186      o Fixed missing NULL pointer protection in ComputeMaterial() method.
187
188 ----------------------------------------------------------------------------
189
190                             Technical Notes
191                             ---------------
192
193  o  This patch should be applied on top of release 8.2
194  o  Technical notes distributed for release 8.2 are also applicable and
195     valid for this patch.
196
197The code and rebuilt binary libraries for release 8.2 are available
198through our "Source Code" Web page:
199   http://cern.ch/geant4/support/download.shtml
200
201Please refer to the Geant4 User Documentation:
202   http://cern.ch/geant4/support/userdocuments.shtml
203for further information about using Geant4.
Note: See TracBrowser for help on using the repository browser.