Pierre-Alain Loizeau authored 3 years ago and Pierre-Alain Loizeau committed 3 years ago

- Device using the central event builder algo in reco, with compatibility to output of the new unpack device
- Macro using these two new devices and setting the parameter files for the mCBM July 2021 data (optimized for run 1588)

Pierre-Alain Loizeau authored 3 years ago and Pierre-Alain Loizeau committed 3 years ago

- Split the InitUnpacker method of the UnpackConfig template into
  - InitOutput
  - RegisterOutput (Framework bound, to be replaced by method in Task class CbmRecoUnpack)
  - SetAlgo
  - initParContainer, moved to the Task Class CbmRecoUnpack
  - InitAlgo
- Move to the template version of these methods all blocks common to all derived config classes
- Whenever necessary, overload these methods in the derived config classes (including the common parts, no base method call)
- Bump Config classes version number
- Adapt the Sts Unpack algo classes to initialize the monitor classes
- Adapt the CbmRecoUnpack
- Adapt the CbmUnpackDevice to use the standard Unpack Config classes (compiles with full functionality but untested)

Expand CbmTaskMakeRecoEvent to all digi types in order to enable reconstruction and analysis from files taken with mCBM.

Administrator authored 3 years ago and Florian Uhlig committed 3 years ago

Add all data classes needed by libAlgo to libOnlineData.
Add all other referenced classses to libOnlineData.
Remove ROOT dependencies from all these classes.

Link libAlgo against libOnlineData.

Pierre-Alain Loizeau authored 3 years ago and Florian Uhlig committed 3 years ago

- Move implementation of CbmMuchAddress::SetElementId to cpp file
- Move Logger include from header to cpp
- Add missing Logger include to CbmMuchSignal cxx file
- Add missing Logger include to CbmAnaDimuonAnalysis cxx file (needed only for FS/FR jun19-v18.2)

Induced by inclusion in DigiData of the other detectors on top of STS (needed only for FS/FR higher than FS/FR jun19-v18.2)

Pierre-Alain Loizeau authored 3 years ago and Florian Uhlig committed 3 years ago

Add includes for new members of DigiData in algo and related task cmake files (to be stashed/fixup w/ change to DigiData)

Pierre-Alain Loizeau authored 3 years ago and Florian Uhlig committed 3 years ago

- Add copy constructor to CbmEvent
- Define DigiData types and add members for the other CBM detectors in CbmDigiData(!MUCH beamtime digi usage!)
- Add vector of CbmEvents, all DigiData types, CbmDigiEvent and vector of CbmDigiEvent to LinkDef to get ROOT streamers (needed to write to ROOT file)

Pierre-Alain Loizeau authored 3 years ago and Pierre-Alain Loizeau committed 3 years ago

In CbmMcbm2018RichPar, add explicit assignment OP constructor + replace map with vector when possible

- Introduce LoadInternalContainers method for setting all memberss derived from other members
- replace fToTshiftMap map with vector as continuous array with always successive keys (faster access)
- add the copy and assignment constructors
- Add a Print method

Cornelius Feier-Riesen authored 3 years ago and Florian Uhlig committed 3 years ago

and also change nofBins of hTofM2 histogram.

Add description for macros and scripts which were missing.

Dominik Smith authored 3 years ago and Volker Friese committed 3 years ago

TimeClusterTrigger: Fixed treatment of edge case where new digi has exactly the seperation 'deadTime' from last trigger.

Eoin Clerkin authored 3 years ago and Florian Uhlig committed 3 years ago

Outputs the structure of a root geometry binaries which comparison by standard unix text based diff tools.

Creates a GRAPHVIZ graph of the geometry structure.

Short README gives some basic command examples.

Applied clang format

Administrator authored 3 years ago and Florian Uhlig committed 3 years ago

With the latest change not all libraries were found any longer such
that not all libraries were tested.
The outcome of the test was such not reliable.

Administrator authored 3 years ago and Florian Uhlig committed 3 years ago

Build glue code for the new enumerator such that it can be properly used in
the macro.

Administrator authored 3 years ago and Florian Uhlig committed 3 years ago

Add missing library dependencies and link directories.
Remove obsolete library dependencies and include/link directories.

Administrator authored 3 years ago and Florian Uhlig committed 3 years ago

To resolve the circular dependency the KF library was split into 3 separate
libraries. KF contains the base part, the new KFQA lib now contains everything
which is related to quality assurance. The new lib KFParticleInterface is the
interface needed to use the external KFParticle library.
There wasn't any cleanup of the build system done yet.

Administrator authored 3 years ago and Florian Uhlig committed 3 years ago

The implementation of the function was in the source file. Since the
function was also used in a qa class no code code be created at that
place. This results in a missing symbol during linking.
Put the template code properly in the header solves the problem.

Administrator authored 3 years ago and Florian Uhlig committed 3 years ago

Add patch needed to for NicaFemto which adds missing library dependencies.

Add missing link directory and library dependency for fles logging library.

Use Vc from FairSoft if available
Only build the internal Vc version if none is found in FairSoft installation.
This solves the problem with failing macros due to incosistent versions
of Vc library and header files.

- Replace all legacy hard-coded numerical values for the TS_MSB cycle size by a constant derived in StsXyterMessage from the data format
- Remove deprecated special cases in the handling of the TS_MSB messages which are not present anymore in the current CRI data format
- Do not update TS_MSB and TS_MSB cycle from the MS header when the cycle happened exactly at the end of last MS
- Recompute the TS+MSB in TS offset from the MS heder only when changing the Cycle or the TS_MSB from the same header
- Add static method to convert Message Type into human readable string
- Slight improve/cleanup of debug printouts and debug levels

Was leading to errors and a crash in mCBM 2021 run 1588 files node8_3_0123 and node8_5_0002

Use FairSoft apr21p1 and FairRoot v18.6.4 for the test.

Before the time was calculated for each raw message, now it is calculated once per µSlice.

Due to some missed casting to double values, the UTC time stored in the CbmTrdRawMessageSpadic lost precision.
Now we also only store the time relative to the Timeslice in the 2021+ unpacker, such that a double is large enough to store the full information.