...
Fixes JIRA GRINF-35. When we moved to a GR v9 based EngineeringModel, MeritTuple was moved from the output sequence to the reconstruction sequence. In case of events with errors, we skip the reconstruction sequence and inserts an empty event in the root files and ntuples, thus preserving the 1-1 correspondence between all the data files. Unfortunately, because of this change the empty events were not inserted into the Merit ntuple, thus breaking the 1-1 correspondence. This new release fixes this bug.
Differences with respect to EngineeringModel v7r0913p10.
LatIntegration v2r65:
30-MAR-07 v2r65 hmk add Leon's brillant NtupleMaker Sequence to insure ntuple is written when recon is skipped.
System Tests for this version
System Tests results:
There are changes in MC tests, none of the major. Since Geant4 has changed this is probably not unexpected. In the data system tests there were changes (KS value different from 1), but none were significant.
I processed part of a data run with the new release and made digi and recon reports. They look normal.
No changes.
Fred version:
Pipeline version:
v1.4.67
GRITS tag (web browsing and task configuration)
...
svac pipeline code and tasks:
code/tasks v3r6p0v3r9p3:
pipelineDatasets v0r3
ISOC code and tasks:
...
Apps that run in pipeline:
eLog: v2r3p1v2r3p2
ConfigTables: v3r2p1v4r0p3
TestReport: v3r6p36 v3r6p41 (digi & recon reports): awb
- v3r6p36 Additional protection in time stamp code.
- v3r6p35 Fixed bug in <6 hits in the TKR and still a TKR trigger table.
- v3r6p34 Added MergeHistFiles.cxx and related lines to cmt/requirementsAdded line break between digi and recon files at the top of the report.
EngineeringModelRoot: v3r0p3 (SVAC tuple): Same as before