Administrator authored 2 years ago and Florian Uhlig committed 2 years ago

When uploading logs from failed tests, these logs are currently truncated
after 102400 bytes. For the long running weekly tests which produces much
larger log files it happens that the important part of the log file isn't
send to the CDash server.
Increase the log file size to 10MB.

Dominik Smith authored 2 years ago and Florian Uhlig committed 2 years ago

Refactored HitFinderTof. Two micro-optimizations of CbmTaskTofHitFinder which save about 10 percent execution time in total.

Dominik Smith authored 2 years ago and Florian Uhlig committed 2 years ago

Switched to using iterators in HitFinderTof.cxx. Fixed a bug in calibration. Added missing clear statement for digi vectors.

Fix c2f test times of long running weekly tests.
Remove tests for old obsolete mcbm setups.

Pierre-Alain Loizeau authored 2 years ago and Florian Uhlig committed 2 years ago

- If git argument given or none, script check all new geo files introduced by an eventual hash change compared to upstream/master
- if all argument given, check all geo file in local copy
- if main argument given, check all geo file in local copy except mcbm ones
- If other argument, use it as a tag to restrict the list of checked geo from local copy (e.g. using mcbm checks only all mCBM geometries)
- Also add gitignore for macro geometry folder
- Add instructions on new script in README file

Pierre-Alain Loizeau authored 2 years ago and Florian Uhlig committed 2 years ago

- Add single macro for all mCBM reco: CbmEvent/DigiEvent, TofTracking/L1 Tracking (based on the event_reco_l1 one)
- Split event building and reconstruction in all CI cases
- Make sure we test all 4 combinations of (CbmEvent, DigiEvent) x (TofTracking, L1 Tracking)

Pierre-Alain Loizeau authored 2 years ago and Florian Uhlig committed 2 years ago

- HadronAnalysis for Lambdas disabled in CI as far too long, O(30 mins) for first TS of 2391
- L1 tracking disabled in CI while pending validation in both reco modes

Correct target position. The macros were not adjusted for the shift of the
coordinate system. The origin is now in the middle of the magnet and not any
longer in the target. A shift of the target by -44 cm is needed.

- Fix calculation of integrated values of STS hist and Tracks over the TS
- Add nb MUCH/TRD/TOF hits per event in L1GlobalTracker debug printout
- Add runtime per event in end of run printout

Most changes done in CbmDigiData:
- Add CbmTzdDigiData class
- Change TZD storage class in CbmDigiData
- Bump Classdef of CbmDigiData to 3
- Add boost serialization version +  Bump to version 3 (default is always 0 in boost)

Dominik Smith authored 2 years ago and Pierre-Alain Loizeau committed 2 years ago

Changed event-time computation from mean of Tzd digis to earliest Tzd digi in CbmAlgoBuildRawEvents.