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

The code was copied at some time from an external project into the lx
directory. Since it fails the license header check the code was moved
to the externals directory. Also a second copy of the headers were deleted.

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

Bug 2: In both Legacy and current unpackers, a signed integer was used at some point to store/return the raw timestamp in clock cycles.
While the last bit would probably never be reached in standard usage, it could lead to negative values and crazy results during the
conversion through double (full precision time in ns) and back to unsigned integer (ns precision time) for data coming from badly
synchronized setups

Bug 3: In the current unpacker, the subtraction of the Timeslice start time from the TS_MSB value was not protected against the case of
badly synchronized data where the TS_MSb does not match the Timeslice index. It now triggers a Fatal as these data are not recoverable
with this unpacker.

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

Bug 1: Size was too small which led to some digis with TS_MSB above 65555 escaping the duplicates filter.
No effect on the digi themselves (time is computed from the original 64b/32b value).

In the legacy case we can stay with 32b (instead of the original 64b or the old 16b) as the field in the
raw message is 29b, while in the current version we need 64b to account for potential TS_MSB cycles and
large Timeslice start offset.

Alexandru Bercuci authored 3 years ago and Florian Uhlig committed 3 years ago

protect against missing par files
update steering macro for the new TRD-2D interface
incorporate parameter calibration into the FASP read-out data
protect the code for the case of par un-consistency (found e.g. for TRD-2Dh)
add user interface to the Asic param set
rename modules in the TRD setup
extend the TRD module type index above one digit
adding class definition to allow SPADIC info generation during unpacking

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

The wrong license headers currently block the merging of !470.

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

- Add vector constructor to CbmSourceTsArchive
- Call string constructor of FLES IPC TimesliceAutoSource if vector of size 1
- Make vector input default in run_unpack macro
- Add alternative call with single string to run_unpack_macro to keep legacy interface available

Dominik Smith authored 3 years ago and Florian Uhlig committed 3 years ago

Moved CbmStsDigitizeQa, CbmStsDigitizeQaReport and CbmStsFindTracksQa to qa subfolders. Updated CMakeLists.txt accordingly.

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

- Add Binning FW support to the Print method of the StsXyterMessage, enabled by default
- Fix the duplicate digis filtering in the STS unpacker and make it flag controllable
- Add an optional full time sorting to the Unpackers steering class
- Related mCBM macros changes

Sergey Gorbunov authored 3 years ago and Volker Friese committed 3 years ago

remove resets of the random seed from the compiled code. We should keep the ability to reproduce all the results when a non-zero seed is set in the main macro

Jan de Cuveland authored 3 years ago and Florian Uhlig committed 3 years ago

This change corresponds to Pascal's request at
!455 (comment 13399)

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

Somehow (at least on some systems) the link order was such that the
first directory was $SIMPATH/lib. This was even the case when the order was
defined differently in the CMakeLists.txt. The problem isn't yet understood
completely.

This results in the problem that libVc.a was taken from $SIMPATH/lib
were it is installed by ROOT/FairSoft. As long as the version provided with
CbmRoot and the version provided by ROOT were the same the problem wasn't
seen. When updating our internal Vc version this results in a creash when
loading libryries due to a version mismatch between the library version and
the header file version.

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

Implemented monitoring for CbmStsUnpackAlgoLegacy. Minor fix to CbmStsUnpackConfig and some cleanup of CbmStsUnpackAlgo and CbmStsUnpackAlgoBase.

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

Bugfix: Replaced ECbmModuleId::kHodo by ECbmModuleId::kT0 in CbmTaskBuildRawEvents and CbmEvbuildRawTask.

Adrian A. Weber authored 3 years ago and Volker Friese committed 3 years ago

remove sorting from Algo and add check if time is greater 0. Mkissing backplanes in data could produce negative values and bee an issue. These digis are not stored

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

With macosx 10.13 (Apple clang 10.0) there is a runtime problem when
loading libL1. The problem is due to the usage of a static constexpr
in the source file which can be overcome when doing a static cast
of the variable.