cvetan [Thu, 1 Oct 2009 22:18:47 +0000 (22:18 +0000)]
1. Automatically enable alien connection of the workers in case it is enabled on the client. 2. Enable schema evolution in order to avoid glitches in PROOF reco in case AliReconstruction class versions on the client and proof cluster are different.
cvetan [Thu, 1 Oct 2009 21:43:06 +0000 (21:43 +0000)]
Adding new URI for raw data - 'raw://run<#>', where <#> is the run number. The URI is handled by AliRawReaderChain. The corresponding chain is created by querying the alien FC. The URI is supposed to be used mainly within the PROOF reco.
jgrosseo [Thu, 1 Oct 2009 14:46:47 +0000 (14:46 +0000)]
adding function to create default tpc-only cuts
adding function to get reference multiplicity from tpc-only tracks
removing functions that are deprecated for more than 6 months
marian [Thu, 1 Oct 2009 13:34:43 +0000 (13:34 +0000)]
AliTPCCalibRaw.cxx.diff Change to fill only sparse info to save space, Add functions to create histos for L1 phase distribution
AliTPCCalibRaw.h.diff Change to fill only sparse info to save space, Add functions to create histos for L1 phase distribution
mtadel [Thu, 1 Oct 2009 11:20:30 +0000 (11:20 +0000)]
Try to open galice.root directly before calling AliRunLoader::Open():
- gSystem->AccessPathName() does not work for alien any more,
always returns "file=exists".
[Reported, but won't be fixed soon.]
- AliRunLoader::Open() calls Fatal() if the file does not exist.
Adding a trigger decision object for every CTP trigger class to the input of the HLTGlobalTrigger
This allows to simply use the CTP class name in the logic of the TriggerMenu
Making the TArrayL64 member variable of the AliHLTGlobalTrigger base class a dynamic member.
There have been problems in the contructor and potential memory corruption due to this class
See https://savannah.cern.ch/bugs/?56302
adding the propagation of the ECS parameter data block to AliHLTSystem
- AliHLTSystem initializes one dummy CTP_TRIGGER_CLASS for the moment
- extended ControlTask functionality, list of blocks to be sent is easily extensible
- filter out warnings for data flow steering blocks in the AliHLTDataSource component
I changed the method MatchToMC to include the possibility to check
on the PDG of the daughters and tightened the momentum conservation
check (this is now possible because the momentum is not any more stored
with limited precision 1/4096 in AliAODMCParticle).
(A. Dainese)
- put baseline subtraction into the clusterizer and remove all other
places (rawStreamers) where this could be done
- use gain calibration outside of TailCancallaion
- inline AliTRDarrayADC::SetData and GetData
1) Removed the obsolete class AliHLTPHOSRcuProperties.cxx/h
as the functionality of this class is now performed
by the specification and data origing variables
of the HLT framework.
2) removed the reference to the classes that was using
the class AliHLTPHOSRcuProperties and modified the code accordingly.
1) Removed AliHLTPHOSAltroConfig.cxx/AliHLTPHOSAltroConfig.h
since the altro configuration is now available from the
RCU trailer.
2) With the removal of the above AliHLTPHOSBase.cxx(h becomes
a man inthe middle class so removed.
3) removed all refernces to AliHLTPHOSBase from other classes
adding a member for the active CTP trigger mask to the CTP data object and setting it for each event
bugfix: the version number of the class was 0, so it was never streamed and thus not in the printout
of the HLT decision
Fix from Roberto: when changing monitoring policy one first has to logout in order to prevent the monitoring from hanging. Should be ported to the release as it affects the trigger selection in the online reco/display.
reorganization of TRD PID reference maker classes. Data management has
been moved to the base abstract class AliTRDpidRefMaker. NN and LQ
implementations have been derived in daughter classes.
Status : compiles and run without problems
Todo : check data quality and algorithm performance for NN. Implement
kdTree approach for reference LQ representation