v01-04

v01-04

• 2017-10-17 Markus Frank (PR#248)

  VolumeManager Implementation

  A possibly important bug was fixed for the lookup of top level subdetectors in the VolumeManager by volume identifers of (sensitive) volumes. Due to a bug in the de-masking possible wrong top level subdetectors were returned. The default use cases typically do not use this call and hence should not be affected.

• 2017-10-17 Shaojun Lu (PR#247)

  • Fix C++11 pointer error by adding include for 'unique_ptr' (GCC 4.9).

• 2017-10-13 Marko Petric (PR#246)

  DDCMS:

  • Improve the CMS excercise. New examples etc.
  • Support for simulation using DDG4 (at least partially - since not all subdetector volumes are accepted by Geant4).

  DDG4:

  • Event reader returns EVENT_READER_EOF if EOF is detected rather than a generic IO error.
  • Add generator status word to the Geant4Particle object. Remove the extension mechanism, which is very heavy to just add one integer.

  General:

  • We need to distinguish the plugins using some namespace mechanism. I started to introduce the namespace separator "_". Hence all DD4hep plugins start with DD4hep_<plugin>. I hope this does not break everything. If it does, please notify me and we can undo.

• 2017-10-13 Whitney Armstrong (PR#243)

  • Added helper function getAttrOrDefault (defined in DDCore/include/XML/Helper.h)
    This function getAttrOrDefault(xml::Element e, xml::XmlChar attr_name, T default_value) will return the attribute name, converted to to type T but if it is not found it will return default_value. When building new detectors supplying this is useful for supplying default attribute values.

• 2017-10-19 Markus Frank (PR#249)

  • Improve the CMS tracker visualisation
  • Add DDG4 simulation example to DDCMS
  • Add some plugins to add visualisation attributes if required (not for the compact description)

• 2017-11-01 David Blyth (PR#254)

  • DDG4/python/DDG4.py: loadDDG4() changed to not raise exception if libraries are already loaded

• 2017-11-01 David Blyth (PR#252)

  • Added requirement of Python 2 in cmake/FindPYTHON.cmake. This makes clear the requirement of Python 2, and resolves the issue where CMake tries to build with Python 3 in a system where both exist.

• 2017-11-07 Frank Gaede (PR#256)

  • bug fix in BitField64::operator[std::string]()
  • make uses of TString in DocumentHandler.cpp compatible with clang9 (on Mac)

v01-03

v01-03

• 2017-10-12 Frank Gaede (PR#244)

  • allow for unbounded surfaces in DDRec
    • add new property SurfaceType::Unbounded
    • if a surface is marked unbounded Surface::insideBounds() ignores the volume boundaries (and only checks the distance to the surface)

• 2017-09-19 Whitney Armstrong (PR#233)

  • Added helper CellIDPositionConverter::cellDimensions(const CellID& cell)

• 2017-10-09 Frank Gaede (PR#242)

  • improve BitFieldCoder class
    • remove heap allocation of BitFieldElements
    • add move constructors for efficient filling of vector

• 2017-09-29 Frank Gaede (PR#238)

  • add new threadsafe class BitFieldCoder as replacement for BitField64
  • use as const everywhere
  • re-implement BitField64 using BitFieldCoder
    • is thread safe if used locally
    • can be instantiated from const BitFieldCoder*

• 2017-09-18 Markus Frank (PR#234)

  • Created a new example showing the CMS tracking detector
    • Get CMS going with their evaluation. Added a package DDCMS with the conversion plugins for the silicon trackers and the corresponding conversion mechanism for their xml structure.

• 2017-09-18 Frank Gaede (PR#232)

  • fix reading of stdhep/lcio generator files with generator statuses not in [0,3]
  • add G4PARTICLE_GEN_BEAM and G4PARTICLE_GEN_OTHER to DDG4
    • G4PARTICLE_GEN_BEAM is generally agreed to be used for beam particles (HepMC, LCIO)
    • all other status codes vary from generator to generator and we use OTHER
  • for stdhep or lcio input the true generator status is preserved in the lcio output, regardless of its value
  • create a vertex for every parent-less particle in LCIOEventReader
    • this allows for example to read GuineaPig files ( non-prompt pair particles) or special user created files with non-prompt particles
  • Resolves #101

• 2017-09-20 Markus Frank (PR#235)

  • A more complete version of the CMS tracker
    • Enhanced the CMS tracker example to be more complete.
    • Stopped at some point to convert all CMS algorithms. Hence, the tracker is not complete, but the remaining work looks to be purely mechanical.

• 2017-10-02 Frank Gaede (PR#239)

  • add cell sizes to printout of LayeredCalorimeterData::layer
    • used in dumpdetector -d

• 2017-09-14 Frank Gaede (PR#231)

  • adapt LCIOEventReader for Pythia8 and Whizard2
    • add all parent-less particles to outgoing vertex
    • fixes #226 and closes #229
    • also used for stdhep files

• 2017-09-07 Daniel Jeans (PR#227)

  • Fix calculation of cell position in MegatileLayerGridXY
  • previously, returned position was the lower corner of the cell
  • after this bug fix, it's the cell centre

• 2017-10-05 Frank Gaede (PR#241)

  • remove deprecated and unused classes from DDRec

• 2017-10-05 Frank Gaede (PR#240)

  • add dd4hep::rec::FixedPadSizeTPCData.zMinReadout
    • needed to describe the cathode thickness

• 2017-08-21 Markus Frank (PR#221)

  • Document several classes in doxygen notation.
    • Aim is that there are (at least) no class headers without docs.
    • See documentation

v01-02

v01-02

• 2017-07-14 Daniel Jeans (PR#204)

  • add ncellsX/Y as a "parameter", allowing it to be set in compact description. This change is for easier use in the case of a MultiSegmentation. (Only a uniform segmentation can be defined in this way: for more complex cases, must set by driver.)
  • change from array to std::vector to store ncells information

• 2017-07-17 Markus Frank (PR#206)

  Implement ROOT persistency mechanism for detector descriptions (continuation of #202).

  Object extensions are now persistent.

  This is however not for free: it requires a dictionary for the extension itself and it requires a dictionary for the class holding the extension. These are:

  • dd4hep::DetElement::DetElementExtension<IFACE,CONCRETE> for DetElement extensions.
  • dd4hep::SimpleExtension<IFACE,CONCRETE> for simple extension managed by the user framework (user calls explicitly destructor).
  • dd4hep::DeleteExtension<IFACE,CONCRETE> for simple extension managed by dd4hep (dd4hep calls automatically destructor on hosting object destruction).
  • dd4hep::CopyDeleteExtension<IFACE,CONCRETE>. As above, but these extensions support calling the copy constructor of the embedded object and hence allow to copy also the hosting objects.

  Please note: to persistify these objects it was necessary to no longer use the type-info of the objects as an identifier, but rather a 64-bit-hash of the raw type-info-name. This relies that this name is identical across platforms. This typically is true for linux, but not enforced by any standard.

• 2017-08-11 Andre Sailer (PR#213)

  • CMake:: dd4hep_generate_rootmap: use CMAKE_INSTALL_LIBDIR if it is set. If the macro is called from other libraries this variable might be set and should be used for consistency. Fixes #212

• 2017-07-13 Markus Frank (PR#202)

  Implement ROOT persistency mechanism for detector descriptions

  The following commits allow to save and restore Detector objects (ie. the full detector description) to/from ROOT. Most changes affected the usage of void pointers.

  It is NOT possible to:

  • save object extensions. Another usage of void pointers and type-info objects, which both cannot be saved.
  • save callback objects. Callbacks store in memory pointers to member functions. These depend on the loading of libraries at run-time and hence may differ from application to application.
  • save conditions slices. This shall be a futute extension. There should be no fundamental problem doing so - it was simply not yet done.

  In the example area a new slot call "Persistency" was created with various tests:

  • Save and restore simple conditions
  • Save and restore the MiniTel detector (examples/ClientTests/compact/MiniTel.xml)
  • Save and restore the CLICSiD detector (examples/CLICSiD/compact/compact.xml)

  The detector examples also have a set of tests associated to check the validity of the restored information.

  Integrate DDSegmentatation into DDCore

  As a first step to start the cleanup of DDSegmentation the code was integrated into DDCore. For backwards compatibility a dummy library libDDSegmentation.so is kept, so that client cmake files directly referring DDSegmentation still work. This tough is only a temporary measure and users should remove references to the DDSegmentation library. Include files can be accessed as before. These are part of the DDCore include directory.

• 2017-08-15 Yorgos Voutsinas (PR#216)

  • Adding variable "nocore" for beam pipe (with default value = false) In case the variable appears in the BP xml file as "true", the BP sections will have no core of beam material, in order that someone might add various BP walls made of different materials while avoiding G4 overlaps. Example of use:

  <detector name="BeBeampipe" type="DD4hep_Beampipe_o1_v01" insideTrackingVolume="true" nocore="true" vis="BeamPipeVis">                                                                                                                     

• 2017-08-15 Jan Strube (PR#215)

  • LCIOEventReader: adding parameter for the name of the MCParticle collection: MCParticleCollectionName

• 2017-08-14 Markus Frank (PR#217)

  Enhance ROOT Detector Description Persistency

  Conditions slices from the DD4hep conditions store can now be saved to ROOT files ans named objects. Conditions persistency examples are added to the examples/Conditions section.
  The examples are derived from the Telescope and CLICSiD example:

  • Conditions_Telescope_root_save
  • Conditions_Telescope_root_load_iov
  • Conditions_Telescope_root_load_pool
  • Conditions_CLICSiD_root_save_LONGTEST
  • Conditions_CLICSiD_root_load_iov_LONGTEST
  • Conditions_CLICSiD_root_load_usr_LONGTEST ...

v01-01-01

v01-01-01

2017-07-18 Frank Gaede

• apply changes from D.Jeans to MegatileLayerGridXY:
  - add ncellsX/Y as a "parameter", allowing it to be set in compact description. This change is for easier use in the case of a MultiSegmentation. (Only a uniform segmentation can be defined in this way: for more complex cases, must set by driver.)
  - change from array to std::vector to store ncells information

v01-01

v01-01

• 2017-07-06 Markus Frank (PR#201)

  DDCore: Changes to the VolumeManager interface

  Recent descrepancies showed that the call to lookup a placement
  from the volume manager may have an ambiguous meaning:
  It may (as used until now) be the placement of thge closest
  detector element - a functionality used by various tests
  or be the placement of the sensitive volume itself.
  So far, since each sensitive volume in the DD4hep tests
  is represented by a DetElement structure, both
  approaches returned the same placed volume.

  Since there is the possibility to have sensitive volumes, which are not
  directly connected to a single DetElement structure, this call was
  split to resolve this ambiguity:

  /// Lookup a physical (placed) volume identified by its 64 bit hit ID
  PlacedVolume lookupVolumePlacement(VolumeID volume_id) const;
  /// Lookup a physical (placed) volume of the detector element
  /// containing a volume identified by its 64 bit hit ID
  PlacedVolume lookupDetElementPlacement(VolumeID volume_id) const;

• 2017-07-06 Andre Sailer (PR#200)

  • Now will give a warning if multiple entities (e.g., constants) of the same name are defined in the XML

v01-00-01

v01-00-01

• 2017-07-04 Frank Gaede (PR#199)
  • bug fix for VolumeManagerContext::toElement() and VolumeManagerContext::placement()
    - set flag=true in VolumeManager_Populator::add_entry when
    a ContextExtension is needed, i.e. sensitive volume is not DetElement's volume
  • fixes problems in CellIDPositionConverter

v01-00

v01-00

• 2017-06-22 Marko Petric (PR#192)

  • Move AlignDet_Telescope_readback_xml to later in the pipeline since it depends on the output of AlignDet_Telescope_write_xml

• 2017-06-22 Andre Sailer (PR#191)

  • Surface: fix memory leak of transformation matrix
  • XML::Layering: fix memory leak of contained layers in the object

• 2017-06-23 Andre Sailer (PR#197)

  • Fix memory leaks for Tube, EllipticalTube and Polyhedron

• 2017-06-23 Andre Sailer (PR#196)

  • CMake: add Project( DD4hep ), needed to get the correct CMAKE_CXX_COMPILER_ID on macs due to CMP0025 (cmake policy)
  • CMake: fix treatment of linker flags, they are now properly set for Linux and Macs to error when undefined functions are encountered at link time
  • CMake: fix elif --> elseif when checking threading libraries

• 2017-06-23 Frank Gaede (PR#195)

  • fix crash in dd4hep::rec::Surface after changes in Handle assignment (PR #193)
  • fix use of deprecated dd4hep::rec::MaterialManager c'tor in Surface

• 2017-06-20 Frank Gaede (PR#185)

  • bug fix in material utilities
    • call MaterialManager( Volume v) with Detector.world().volume()

• 2017-06-20 Marko Petric (PR#184)

  • Reinstate the full test-suite on Travis

• 2017-06-20 Markus Frank (PR#183)

  • Unify header guards in DDCore
  • Add header to steer ignoring warnings of rootcling generated dictionaries.

• 2017-06-20 Frank Gaede (PR#182)

  • cleanup of namespace dd4hep::rec
    • remove obsolete bwd compatibility for DD4hep::DDRec
    • re-introduce [deprecated] warnings for unmaintained classes in DDRec/API
    • re-fix deprecated c'tor for MaterialManager in material utilities

• 2017-06-20 Markus Frank (PR#181)

  • Attack many warnings from:
    • -Wshadow
    • -Winclude-hygiene
    • -Woverlength-strings (int cling dictionaries)

• 2017-06-20 Markus Frank (PR#179)

  • Remove a bunch of shadow warnings and include-hygiene warnings.

• 2017-06-21 Marko Petric (PR#169)

  • Make boost explicit requirement for DD4hep and drop DD4HEP_USE_BOOST

• 2017-06-21 David Blyth (PR#168)

  • Added environment helper scripts thisdd4hep_only.(c)sh that only set up variables for DD4hep and not for dependencies.

• 2017-06-19 Markus Frank (PR#178)

  • Update documentation after reorganization of namespaces (put back previous docs).

• 2017-06-19 Markus Frank (PR#175)

  DD4hep namespace reorganization

  Re-organize namespaces according to the decisions of the DD4hep developers meeting from 16th June we have decided:

  1. all namespaces will be lower case and shorter
     • rename namespace DD4hep -> dd4hep
     • rename namespace DD4hep::DDRec -> dd4hep::rec
     • rename namespace DD4hep::Simulation -> dd4hep::sim
     • rename namespace XML -> xml and JSON -> json
     • rename all other namespaces according to this pattern
  2. The namespace DD4hep::Geometry:: will be incorporated into dd4hep::
  3. All utilities will be moved dd4hep::detail
  4. LCDD will be renamed to Detector and current Detector.h will be renamed to DetElement.h
  5. Examine if DDSegmentation can be incorporated into DDCore and make it volume aware
     • If this this cannot be achieved in whole place DDSegmentation into the right namespace

  DDParsers

  DDParsers are now a separate package. This does not make it yet standalone,
  but it is at the same level as e.g. DDSeqmentation. Any librarian is
  encouraged to externialize it fully.

v00-24

v00-24

• 2017-06-08 Markus Frank (PR#160)

  • Add a new class AlignmentsNominalMap, which behaves like a ConditionsMap and handles alignment entries.

  • The AlignmentsNominalMap is not a conditions cache per-se. This implementation behaves like a conditionsmap, but it shall not return real conditions to the user, but rather return the default alignment objects (which at the basis are conditions as well) to the user. These alignments are taken from the DetElement in question Alignment DetElement::nominal().

  • The basic idea is to enable users to write code "as if" there would be conditions present. This is important to ease in the lifetime of the experiment the step from the design phase (where obviously no conditions are taken into account) to a more mature phase, where alignment studies etc. actually are part of the "bread and butter work".

  • Added a corresponding example in examples/AlignDet:

  $>   geoPluginRun -volmgr -destroy -plugin DD4hep_AlignmentExample_nominal \
           -input file:${DD4hep_DIR}/examples/AlignDet/compact/Telescope.xml
  

  • Access the DetElement nominal conditions using the AlignmentNominalMap.
    Any use of DDCond is inhibited.
    1. We use the generic printer, which during the detector element scan accesses the conditions map.
    2. We use a delta scanner to extract the nominal deltas from the DetElement's nominal alignments
    3. We use a ConditionsTreeMap to perform the alignments re-computation.

• 2017-06-08 Markus Frank (PR#159)

  Implementation of the decisions made at the Conditions mini-workshop

  Access mechanisms of DD4hep conditions for utilities

  Access to conditions is solely supported using the interface class DDCore/ConditionsMap.

  • All utilities must use this interface.
  • Any concrete implementation using conditions/alignment utilities must implement this interface
  • Basic implementation using STL map, multimap and unordered_map are provided.
  • A special no-op implementation of this interface shall be provided to access "default" alignment conditions. This implementation shall fall-back internally to the DetElement::nominal() alignment.
    Known clients: VolumeManager (hence: DDG4, DDRec, etc.)

  Though this sounds like a trivial change, the consequences concern the entire conditions
  and alignment handling. This interface decouples entirely the core part of DD4hep
  from the conditions cache handling and the alignment handling.

  Based on this interface most utilities used to handle conditions, detectors scans
  to visit DetElement related condition sets, alignment and conditions printers etc.

  For details, please see:

  DDCore/include/DD4hep/AlignmentsPrinter.h
  DDCore/include/DD4hep/AlignmentsProcessor.h
  DDCore/include/DD4hep/ConditionsPrinter.h
  DDCore/include/DD4hep/ConditionsProcessor.h
  DDCore/include/DD4hep/DetectorProcessor.h
  

  Naming conventions for detector conditions

  • Condition are logically attached to DetElements

    • Condition names are: DetElement.path()+"#"+condition-name
      Example: /world/LHCb/DownstreamRegion/Muon/M5/M5ASide/R3ASide/Cham046#alignment

  • Condition keys are a int64 compound of two int32:

     union {
       int64 key;
       struct {
         int32 item_key;
         int32 det_key;     // Needs to be the high word to have a properly ordered map
       } values;
     };
     det_key  = hash32(DetElement.path())
     item_key = hash32(condition-name)

  Condition keys must be unique throughout the detector description.

  • Alignment conditions naming conventions:

    • Alignment-delta conditions are called alignment_delta.
    • Fully qualified alignment conditions are called alignment.
      DD4hep provided alignment utilities rely on this convention.

  • Other conditions can be named freely.

  Important Notice

  The Alignment conditions naming conventions are already used by several utilities involving alignments. If you plan to use these, do not freely ignore these recommendations. When the naming conventions are ignored, these utilities shall not work.

  Updates to DDCond

  DDCond implements a working conditions cache following the design criteria sketched above. The conditionsSlice object implements (though by forwarding to the ConditionsUserPool) a ConditionsMap interface.

  The DD4hep_ConditionsMapUserPool plugin implements in a very efficient way this interface using an ordered map. Using the above described key definition, this implementation allows very efficient scans of conditions/alignments etc. of individual detector elements, since conditions which belong to the same detector element are contiguous.

  Alignment handling/computations

  Using the conditions maps, the computation of (mis-)alignment data from deltas
  is no longer bound to the conditions mechanisms.

  A special utility called AlignmentsCalculator is put in place (see DDCore/include/DD4hep/AlignmentsCalculator.h) to facilitate the computation of a coherent set of alignments given a set of delta-parameters. This mechanism is much simpler, easier to understand and far less code intensive than the previously designed callback mechanism where alignments are obtained using conditions derivation.

  Update of the existing examples

  The example sets in DDDB, examples/Conditions, examples/AlignDet, examples/DDDB were updated according to the changed mechanism of accessing conditions. Here we can see the real benefits of the new approach: keeping same functionality, the examples became way off simpler. Simply count the number of lines of code.

• 2017-06-17 Marko Petric (PR#170)

  • Add clang flag to warn about using namespace directive in global context in header

• 2017-06-17 Frank Gaede (PR#167)

  • renamed the namespace DD4hep::DDRec to dd4hep::rec (see #166)
    - provide backward compatibility to outside world for now
  • moved the interfaces in namespace DDSurfaces to dd4hep::rec
    - provide backward compatibility to outside world for now

• 2017-06-15 Frank Gaede (PR#165)

  • started to cleanup DDRec
    - don't use LCDD::getInstance() in SurfaceManager and SurfaceHelper
    - deprecate unused(?) classes in DDRec/API and DDRec/Extensions
    - deprecate MaterialManager() using LCDD::getInstance()

v00-23

v00-23

• 2017-05-12 Marko Petric (PR#152)

  • Update CI to GCC 7.1 and LLVM 4.0 and include Geant4 10.3

• 2017-05-22 Frank Gaede (PR#154)

  • protect against NANs in Guineapig pairs files in Geant4EventReaderGuineaPig
  • make INFO printout more consistent with DD4hep style

• 2017-06-07 Frank Gaede (PR#157)

  • bug fix in test_cellid_position_converter
    • with this no tests for position from cellID lookup should fail
  • re-implement CellIDPositionConverter::cellID(pos)

• 2017-06-08 Marko Petric (PR#156)

  • Mark all fallthroughs in case statements with attributes to suppress warning

• 2017-06-01 Frank Gaede (PR#155)

  • add new class DDRec::CellIDPositionConverter
    - replaces DDRec::IDDecoder
    - implement positionNominal(CellID id) and cellID(position)
    • prepare for using alignment map by separating transforms to DetElement and daughter volume
    • do not use deprecated methods/members in VolumeManager
  • add test_cellid_position_converter.cpp
  • add VolumeManagerContext::toElement
    • transform from sensitive volume to next DetElement

v00-22

v00-22

• 2017-04-28 Markus Frank (PR#148)
  Improvements to the compact xml processing

  Elements may now be specified within compact in 2 ways:

  1. old way: create an effective element:
     <element Z="4" formula="Be" name="Be" >
<atom type="A" unit="g/mol" value="9.01218" />
</element>

  2. create element by defining an isotope mixture:
     <isotope name="C12" Z="6" N="12"/>
<atom unit="g/mole" value="xxxx"/>
</isotope>
.....
<element name="C">
<fraction n="0.9893" ref="C12"/>
<fraction n="0.0107" ref="C13"/>
</element>

  Improved debugging of compact xml conversions
  The following tags allow to enable increased prinout depending on additional tags in the compact file:
  <debug>
<type name="isotopes" value="1"/>
<type name="elements" value="1"/>
<type name="materials" value="0"/>
<type name="visattr" value="0"/>
<type name="regions" value="0"/>
<type name="readout" value="0"/>
<type name="limits" value="0"/>
<type name="segmentation" value="0"/>
</debug>

  Disable the ROOT TGeo element table from the compact xml.
  Note: ALL elements must then be specified in the XML material database.
  <geometry>
<clear name="elements"/>
</geometry>

  By default compact accepts exactly one input file (others may be included therein).
  In order to process iteratively multiple input files, the opening and the closing of the
  geometry steering tags may be added to the compact description. This example
  gives the default behaviour:
  <geometry open="true" close="true"/>

  Please note: per compact file exactly ONE geometry tag is allowed.

  Debugging the DDG4 geometry conversion mechanism

  New boolean properties of the Geant4DetectorGeometryConstruction object,
  which result in debugging printouts (defaults are OFF):
  DebugMaterials
  DebugElements
  DebugShapes
  DebugVolumes
  DebugPlacements
  DebugRegions
  PrintPlacements
  PrintSensitives

• 2017-04-28 Ben Couturier (PR#146)

  • Trivial fix for the DDDB converter to create paramphysvol3D volumes, which are otherwise ignored.

• 2017-04-20 Andre Sailer (PR#145)

  • LCIOOutput: Add setting of ProducedBySecondary bit for SimTrackerHits if the hit is produced by a particle that is not stored in the MCParticle collection, needs lcio 2.8

• 2017-05-05 Andre Sailer (PR#150)

  • Always create DD4hepConfigVersion.cmake in CMAKE_INSTALL_PREFIX and cmake folder
  • Create DD4hepConfig.cmake also in cmake folder
  • renamed Cmake Macro GENERATE_PACKAGE_CONFIGURATION_FILES to DD4HEP_GENERATE_PACKAGE_CONFIGURATION_FILES so it does not clash with the macro of the same name in ilcutil/cmakemodules

• 2017-05-07 Andre Sailer (PR#151)

  • Use cmake to create Version.h file to contain DD4hep version information and macros
  • Change the way DD4hep package version is defined and set standard cmake variables for this purpose