richterm [Sat, 1 Aug 2009 09:45:28 +0000 (09:45 +0000)]
- adding the DAQ readout list block to the output of the trigger component
- adding some comments and documentation in the ReadoutList handler
- copy only the overlapping part of the list from an AliHLTEventDDL block
richterm [Sat, 1 Aug 2009 09:35:58 +0000 (09:35 +0000)]
adding data type definitions
kAliHLTDataTypeDAQRDOUT (DAQRDOUT)
kAliHLTDataTypeTriggerDecision (TRIG_DEC)
kAliHLTDataTypeGlobalTrigger (GLOBTRIG)
note: the interface version has not been incremented as it did not change apart
from those definitions
- Error estimation for the tilted Riemann Fit (this is necessary since their
roads for the 5th and 6th tracklet are defined according to the error
estimation of this fit)
- Add systematic uncertainty to the cluster error in y-direction in attach
cluster. This systematic error takes into account misalignment and
decalibration and is defined in the Reco Params (2cm for cosmics, 0cm for
simulation)
- In the seeding part replace condition for accepting clusters
(tracklet->IsOK()) by nclusters - nused -nNotInChamber > 6 (avoids a crash)
- Tracklet fit for the seeding tracklets and in case of HLT tracklet 5 and 6
- Reintroduce the "harmfull condition" since it was only harmfull because of
the missing tracklet fit
- Increase weight of the 4 consecutive seeding layers by factor 5. Interuptions
should not be the standard and only be allowed if something really bad
happened (i.e. chamber off)
- The function FitRiemanTilt will be replaced by a class for several reasons
(i.e. 2nd fit with only clusters in a 2-sigma resiual band around the track
in offline mode forseen). The class is created and a function that replaces
the old FitRiemanTilt is written, but not yet tested, so not enabled.
Bug detail:
1. memory reinitialization wasn't properly done for some mcms due to wrong
mcm number setting in the reseting loop.
2. when we meet the mcm header error and stop data reading, mcm number
counter wasn't increased and it prevented us to get the mcm header error
code of this problematic mcm.
modified AliTOFRawStream::GetDataBuffer and AliTOFRawStream::GetDataBuffer getters in order to return buffer pointer and not whole object. TOFda.cxx modified accordingly.
correcting the default HLT TPC reconstruction macro in order to use the Altro v3 decoder,
deleting obsolete CA tracking macro as this is incorporated in the standard macro
bugfix: the EsdCollector was not working after fix r33443 since it relies on the writing of
local ESD files by the HLTEsdManager. This has been made an optional feature, used by the
collector
Updates provided by Magali Estienne
1 - remove the coding convention violations you sent to me few weeks ago
2 - look for jets using charged + neutral information from AODs
- overload function FillTracksChecked in AliHLTTPCTrackArray.* to make it work with the new track structure
- modify the loops over clusters and tracks in AliHLTTPCTrackHistoComponent to have easier access to the used clusters
- update to the new track structure to enable the ntuple filling
bugfix: the macro has not been adapted yet to the developments of the last year, FillESD is required for HLT in order to merge the ESD output of the macro to the AliESDs.root
I improved (by a factor 2.5) the speed of the MatchToMC method
(there were unnecessary new/delete).
I changed accordingly also the class PWG3/vertexingHF/AliAODRecoCascadeHF.cxx)
[A. Dainese]
Using common HLT track data format for TRD tracks (Theodor)
- output of HLTTRDTracker is now standard HLT Tracks Struct
- changes in GlobalEsdConverter to follow HLTTRD changes
- GlobalTrackMerger now handles standard HLT Tracks Struct
Bug detail:
In case we use NextChamber(,..) function(which we does in our normal
reconstruction) in the current reader, we are missing the first ADC
channel data of every chamber except for the first chamber of a
supermodule. It happened since we use this first channel to check if we
meet new chamber without iterating back in case we return new chamber.
Theodor, once it is in svn, could you again compare the output of this
reader to those of HLTstreamer, and let me know if you still see the
problme? Thanks.