Remove thread unsafe, incident-dependent variable in MCReconstructible
For the HltEfficiencyChecker tool (prints out/plots some efficiencies for the new HLT, see https://gitlab.cern.ch/lhcb/mooreanalysis/merge_requests/1) I use some tuple tools to write trigger decisions and basic kinematic information to a tuple. Out of the box, MCTupleToolReconstructed seg faults due to some old-school treatment of the m_TkInfo object of MCReconstructible. Before this MR, mc track info (in the m_tkInfo member variable) was populated once per event when you call the mcTkInfo() method here, then at the end of an event, in the olden days, this handle would be called that would reset the track info for the next event. But this incident handling has now gone, and so m_TkInfo just uses info from the first event, which probably contains reference to the first event which doesn't remain valid for subsequent events -> seg fault.
Also this is very thread unsafe.
I also fix a typo I spotted in Event/MCEvent/Event/MCTrackGeomCriteria.h.
TODO:
-
Remove the handle altogether? -
Switch to non-deprecated MCTrackInfoconstructor to remove the deprecation warnings in the builds.
Dependency of https://gitlab.cern.ch/lhcb/mooreanalysis/merge_requests/1, Analysis!576 (merged). Should also be merged with Moore!389 (merged), !2373 (merged), Phys!662 (merged) but could be merged before them all.
Merge request reports
Activity
added 13 commits
-
6ef0ffef...236b1e28 - 12 commits from branch
master - 5cfd8467 - Merge branch 'master' into rjhunter-make-MCReconstructible-Work
-
6ef0ffef...236b1e28 - 12 commits from branch
mentioned in merge request !2289 (merged)
mentioned in issue #72 (closed)
added 1 commit
- 911c0487 - Removed any reference to IIncidentListener and smart pointers->local variables for track info
Thanks for your help @graven - I've addressed all the points you made and opened a follow-up issue for the deprecated
MCTrackInfoconstructor. Unless there are any further comments, I think this is ready to merge.marked the checklist item Remove the handle altogether? as completed
added MC checking Simulation modernisation labels
removed Simulation label
added 7 commits
-
37715c6c...e577f4cf - 6 commits from branch
master - 6ba42938 - Merge branch 'master' into rjhunter-make-MCReconstructible-Work
-
37715c6c...e577f4cf - 6 commits from branch
-
let's put this in lhcb-gaudi-head
added lhcb-gaudi-head label
- [2020-01-31 00:06] Validation started with lhcb-gaudi-head#2519
- [2020-02-01 00:05] Validation started with lhcb-gaudi-head#2520
- [2020-02-02 00:05] Validation started with lhcb-gaudi-head#2521
Edited by Software for LHCbadded 29 commits
-
6ba42938...bdea3692 - 16 commits from branch
master - 2f37a199 - Reset m_tkInfo for each particle
- d054a313 - Better index searching
- 0886b20c - Fixed typo
- 3109a7a3 - Fixing memory leak: removed MCTkInfo funk and added instance each place when needed.
- 88f250ea - Revert change to MCProperty - wasnt causing the issue
- 4c92e65c - Fixed formatting
- ee327942 - remove incident handle
- 10848ca9 - Removed any reference to IIncidentListener and smart pointers->local variables for track info
- b00d58d8 - Fixed formatting
- 288d397f - Suffix line names with Decision in DecReports writer
- 050f39ac - Reverting because I want to check first if any chagnes to master
- d1839f81 - Merge branch 'master' into rjhunter-make-MCReconstructible-Work
- 1cb8be30 - Merge branch 'rjhunter-make-MCReconstructible-Work' of...
Toggle commit list-
6ba42938...bdea3692 - 16 commits from branch
Resolved by Ross John HunterLooks good to me
- Last reply by Ross John Hunter
removed lhcb-gaudi-head label
added 1 commit
- f42a61e5 - Use dataobjectreadhandle to pass around the track info
added 1 commit
- 7725aa92 - Make the mc track info names more meaningful
mentioned in merge request Analysis!576 (merged)
added lhcb-gaudi-head label