marian [Thu, 15 Jul 2010 07:23:09 +0000 (07:23 +0000)]
AliTPCcalibDButil.cxx - make function public -possibility to debug from command line
AliTPCcalibSummary.h - adding trend info of relative change of pulser and CE timing - in respect with reference run
AliTPCcalibTracks.h AliTPCcalibTracks.cxx - adding tracklet residual histograms
AliTPCCalPad.h AliTPCCalPad.cxx - make calib pad from the histogram
Introducing class members for storing the p_T dependent DCA cuts, these are only for internal use and are modified via the setters of the string as before. One-time construction of TFormula instead of once per track
The monitoring task creates plots of the cut variables after different
cut
steps (0: just quality cuts, 1: invariant mass cut, 2: all cuts) and
maps the
quantities of tracks taken out by quality and V0 PID cuts in
corresponding
histograms.
In the AliTRDv0Info, the Armenteros-Polanski method was added and a few
graphical cuts adapted from the HFE V0 selection by Matus Kalisky.
The Kalman filter is in as a method, but not working properly yet (so
the cuts
on Chi2/NDF are by default not set).
using a sequence of 10 smaller propagations seems to work better than AliExternalTrackParam::PropagateTo in one go
enable updating of track parameters with TPCout and corrected propagation (Timur)
reverting revisions 41975 and 41935
The behavior of AliTPCcalibDB in the online system needs to be clarified first
bug https://savannah.cern.ch/bugs/?69885
Make sure all arrays are initialized properly in the ReadSOD member function.
Do _not_ vote on the settings. Instead write out useless values for the
effected cards. In this way, we can more easily detect problems.
Bug in the WriteToFile member functions meant that we _always_ wrote the
_same_ value for both inner/outer cards. The 'board' index was taken to
be just that, but in reality it is a half-ring index - very different.
Since we've been running with symmetric set-ups, this is not a problem.
However, this change should be proted to the release
Inners: Sector 0-> 9 mapped to board 0x00
Sector 10->19 mapped to board 0x10
Outers: Sectir 0->19 mapped to board 0x01
Sector 20->39 mapped to board 0x11
The correct mapping is the opposite:
Inners: Sector 0-> 9 mapped to board 0x10
Sector 10->19 mapped to board 0x00
Outers: Sectir 0->19 mapped to board 0x11
Sector 20->39 mapped to board 0x01
Note, that this _only_ effected the QA histograms, since it's the
only place where the Sector2Board member function is called. I found
this bug in our laborartory set-up.
general code audit, added documentation about AliESDtrack update sequence
- MC labels had not been set correctly if the ITS tracker did not return an associated label, in such a case the TPC label is correctly kept
still it needs to be verified why the tracker finds a track but does not assign any label
bug https://savannah.cern.ch/bugs/?69713
- AliESDtrack outer parameter was overwritten by wrong parameters coming from the ITS tracker,
this fixes partly bug https://savannah.cern.ch/bugs/index.php?69872
update with kITSout is skipped now, should be done for ITS standalone tracks only
- wrong kTPCout reference point set for the display, skip update with flag kTPCout
bug https://savannah.cern.ch/bugs/index.php?69875
the correct propagation to the last TPC cluster needs to be found