source: trunk/MagicSoft/Mars/Changelog@ 6265

 
Please  do  not write behind the end of line tag so that it is  possible
to   send  a   Changelog   file  to  a  line  printer  directly.  Thanks.

For your entries please follow EXACTLY this style:

_yyyy/mm/dd_[Your Name]
_
___*_filename:
_____-_description
_
___*_filename2, filename3:
_____-_description
_
_
_

While an underscore is a placeholder for a white-space or an empty line.


                                                 -*-*- END OF LINE -*-*-

 2005/02/04 Markus Gaug

   * mjobs/MJCalibration.cc
     - found a "the `impossible' happened" when running callisto with 
       a pure signal extractor --> MExtractTimeFastSpline has some 
       problems with the management of the internal arrays. Replaced it 
       by MExtractTimeHighestIntegral as default which is as good. 

   * callisto.rc
     - use MExtractTimeAndChargeDigitalFilterPeakSearch as default 
       instead of MExtractTimeAndChargeDigitalFilter. This extractor is 
       now tested and checks for the positions of the pulse within the
       DAQ-window.



 2005/02/04 Thomas Bretz

   * Makefile.rules:
     - fixed creation of main MDep.d

   * callisto.cc, merpp.cc, showlog.cc, showplot.cc, star.cc:
     - added commandline option --version

   * mbase/MFilter.cc:
     - changed default for 'GetRule'

   * mbase/MParContainer.cc:
     - simplified GetNewObject

   * mfbase/MF.cc:
     - if no conditional sign is found try to create
       MFilter-object

   * mraw/MRawEvtData.cc:
     - fixed the initialization of fABFlags in case maxid==0 which
       is the default. This caused MRawEvtPixelIter::HasABFlag
       to do unallowed access to memory for old root-files without
       the AB flag, because it assumed (due to size==1) that the
       data contains the ab flag

   * mraw/MRawEvtPixelIter.[h,cc]:
     - removed some obsolete debug comments
     - simplified HasABFlag

   * manalysis/MCerPhotPix.h:
     - removed the `!` again. It gave problems with the Mars
       display. We agreed that this is a minor problem because
       the structure of MCerPhot* will chnage soon in any case.



 2005/02/04 Abelardo Moralejo

   * macros/starmc2.C
     - added to output missing container MConcentration.



 2005/02/03 Thomas Bretz

   * mfbase/MFilterList.[h,cc]:
     - added new function AddToList(TCollection&)

   * mhbase/MFillH.cc:
     - print a warning if 'same'-option given without 
       a corresponding existing tab

   * mhflux/MHAlpha.[h,cc]:
     - made time bins settable by user

   * mfileio/MWriteRootFile.cc:
     - fixed ReInit and a warning output when writing to memory



 2005/02/03 Markus Gaug

   * mcalib/MCalibColorSet.[h,cc]
     - implement possibility to set colour explicitely (e.g. for MC).

   * msignal/MExtractTimeAndChargeDigitalFilter.cc
     - intialize fTimeShiftHiGain and fTimeShiftLoGain by default.
 
   * mjobs/MJPedestal.cc
     - in case, extractor is used, put the extractor into the tasklist, 
       otherwise ranges do not get initialized correctly (wrong memory 
       usage!). Use an event selector, however to not skip the Process.
       This is as fast as before and now safe for any kind of extractor
  
   * mcalib/MCalibrationCam.h
   * mcalib/MCalibrationTestCam.h
   * mcalib/MCalibrationTestCam.cc
     - initialize a missing array in InitAverageAreas
 
   * mraw/MRawEvtHeader.h
     - introduced Setter() of cal. pattern for Abelardo



 2005/02/02 Markus Gaug

   * mcalib/MCalibrationChargeCam.cc
     - fixed small bug in the display of mean charges when low-gain was
       used.

   * msignal/MExtractor.cc
     - print the missing overlap between high-gain and low-gain
       in the Print() function
     - remove setting of fHiLoLast in PreProcess, otherwise cloned 
       extractors do not work any more with windows exceeding the 
       high-gain range.

   * mhcalib/MHCalibrationTestCam.[h,cc]
   * mhcalib/MHCalibrationTestTimeCam.[h,cc]
     - set proper probability limits for the fits

   * mcalib/MCalibrateData.cc
     - catch the case in which no interlaced cal. events are being sent
       by the pulser box.

   * mjobs/MJCalib.[h,cc]
     - put the data type enum in a new base class

   * mjobs/MJCalibTest.[h,cc]
     - re-name old class MJExtractCalibTest to MJCalibTest and make it 
       compatible with the sequence-files.
   
   * mjobs/MJExtractCalibTest.[h,cc]
     - removed

   * mjobs/MJCalibrateSignal.cc
     - put a compiler warning about the default cal. to photo-electrons

   * callisto.cc
     - add new flag "--use-test" to enable the MJCalibTest facility
       (default: not used)



 2005/02/01 Raquel de los Reyes

   * mjobs/MSequence.h
     - Added "Get" functions for the TArrays variables.



 2005/02/01 Markus Gaug
  
   * mjobs/MJPedestal.[h,cc]
   * mjobs/MJCalibration.[h,cc]
   * mjobs/MJCalibrateSignal.[h,cc]
   * callisto.rc
     - put enum for data type. Can now set the data type (mc, raw, root)
       without running the data check.

   * mhcalib/MHCalibrationCam.cc
     - return kTRUE in PostProcess if GetNumExecutions is 0
   
   * mcalib/MCalibrationChargeCalc.cc
     - return kTRUE in PostProcess if GetNumExecutions is 0



 2005/02/01 Daniel Mazin

   * MPointingPosInterpolate.cc:
     - fixed bug to match the units of Ra and Decl in MPointingPos
     - fixed bug in handling of events with times outside of times from 
       MDrive tree



 2005/02/01 Thomas Bretz

   * showplot.cc:
     - another fix to make the batch mode working (you can use the batch
       mode to convert a root-file into another format)



 2005/01/31 Thomas Bretz

   * mbase/MStatusDisplay.cc:
     - fixed handling of batch mode in GetCanvas(const TString &name)
     - don't skip already existing tabs in Display()



 2005/01/31 Raquel de los Reyes

   * mjobs/MSequence.h
     - Added "Get" functions for some of the sequence.txt file 
       parameters.



 2005/01/31 Markus Gaug 
 
   * mjobs/MJob.[h,cc]
   * mjobs/MJCalibrateSignal.[h,cc]
   * mjobs/MJCalibration.[h,cc]
     - move file type enum to base class

   * mjobs/MJCalibrateSignal.[h,cc]
     - allow also for direct setting of an MRunIter (not used in
       callisto).

   * mcalib/MCalibrationChargeCalc.[h,cc]
     - test also for pulser strength changes additionally to changes in 
       colour.



 2005/01/31 Abelardo Moralejo

   * mpointing/MSrcPosCalc.cc
     - Create only one instance of MAstroSky2Local for the 
       transformations, instead of two identical ones, as it was before.



 2005/01/30 Markus Gaug 
 
   * mjobs/MJCalibration.cc
   * mjobs/MJPedestal.cc
   * mjobs/calibrationref_Nov04.rc
   * mhcalib/MHCalibrationRelTimeCam.[h,cc]
     - updated correct display of reference lines



 2005/01/29 Markus Gaug 
    
   * msignal/MExtractFixedWindowSpline.[h,cc]
     - fixed a bug occurring with overlapping windows between high-gain
       and low-gain

   * msignal/MExtractFixedWindowPeakSearch.cc
     - take out call to fSignals->Clear() which is already done in 
       MTaskList.

   * msignal/MExtractTimeAndChargeDigitalFilterPeakSearch.[h,cc]
     - new extractor reducing the bias and solving the problem with the 
       jumping pulse readout.



 2005/01/28 Abelardo Moralejo

   * mpointing/MSrcPosCalc.cc
     - updated. Make it work as desired: obtain for each event the x,y 
       position on the camera of a source whose celestial coordinates 
       (J2000) have been set by the user in the container "MSourcePos" 
       of type MPointingPos, added to the parameter list. If the 
       container MSourcePos is not found, the center of the camera (or 
       other fixed position x,y set in the parameter list by the user - 
       in MSrcPosCam) is used as source position for all the events in 
       the loop.



 2005/01/28 Thomas Bretz

   * mjobs/MJCalibrateSignal.cc:
     - fixed that extract1,2,3 haven't been deleted if loop failed.
     - print tasklist also if loop failed.
     - added a undefined task (MTaskEnv "FinalFantasy") as last task to
       the event processing tasklist

   * callisto.rc:
     - added comments about "FinalFantasy" into call
     - added comment for obsolete resources as discussed yesterday
       evening: TimeUpperLimit

   * mbase/MMath.[h,cc], manalysis/MCerPhotPix.[h,cc], 
     mpedestal/MPedPhotPix.h:
     - replaced TruncatePrecision by ReducePrecision (its simply a better
       name)

   * mbase/MParList.cc:
     - added a check to make 200%(!) sure that the container added
       to the list is a MParContainer (everything else could crash
       almost everything in Mars!)

   * mjobs/MJCalibrateSignal.cc, mjobs/MJCalibration.[h,cc],
     mjobs/MJPedestal.cc:
     - simplified the code in the WriteResult functions by using
       the new WriteContainer in MJob
     - removed the TObjArray from the MParList! It is NOT a MParContainer
       there are a lot of circumstance how this can crash Mars. I don't
       know why the compiler allowed for this.
     - unified WriteResult and WriteHists
     - split GetOutputFile into two functions for simplification

   * mjobs/MJCalibration.h:
     - copy the extractorcam name when copying the contents

   * mjobs/MJob.[h,cc]:
     - new member function WriteContainer to simplify code in derived
       classes

   * mpointing/MSrcPosCalc.[h,cc]
     - fixed the header according to Abelardo's changes
     - added the initialization of the pointers to the constructor



 2005/01/28 Markus Gaug

   * msignal/MExtractTimeAndChargeDigitalFilter.cc
     - introduce a security check for the time fine adjustment in case, 
       it is too large.

   * mcalib/MCalibConstCam.[h,cc]
   * mcalib/MCalibConstPix.h
     - new classes to allow display of changing cal. constants with the
       interlaced cal. events

   * mcalib/MCalibrationChargeCalc.cc
     - removed stopping of PostProcess in case that the evtloop is
       finished with interlaced cal. events

   * mcalib/MCalibrateData.[h,cc]
     - fill MCalibConstCam with update results (once per calib. consts
       update == every 10 sec.)
     - introduce check for size of calibration constant. In case, it is 
       completetly out of range, set the MBadPixelsPix::IsUnsuitableRun

   * mjobs/MJCalibrateSignal.cc:
     - display one MHCamera more with changing conv. factors (only for
       interlaced cal. events)

   * mpedestal/MPedCalcFromLoGain.cc
     - set default for MExtractWinFirst from 15 to 17 because of the 
       late cosmics signals in December


 2005/01/28 Abelardo Moralejo

   * mtemp/mpadova/macros/trainsubsample.C
     - added # of events to be written in the output (instead of 
       fraction of total). Moved MSrcPosCam in output from RunHeaders
       to Events tree.



 2005/01/27 Daniel Mazin and Markus Gaug

   * mcalib/MCalibrateData.[h,cc]
     - derive this class from MCamEvent in order to be able to display
       evolution of conversion factors in case of interlaced cal. 
       events

   * mpointing/MInterpolatePointingPos.[h,cc]
     - new class to interpolate pointing positions, adapted from a class 
       by Marcos Lopez 

   * mjobs/MJCalibrateSignal.[h,cc]
     - enabled the calibration of data with interlaced calibration events     

   * msignal/MArrivalTime.[h,cc]
   * mcalib/MCalibrateRelTime.cc
     - removed Error of calibrated arrival time, as agreed on the last 
       software calibration meeting yesterday

   * mjobs/MJCalibrateSignal.cc
     - add display to show histogramming of conversion factors
     - modify default calibration to equiv. photo-electrons
     - allow for setting of MC, Raw or Root-file type in callisto.rc

   * callisto.rc
     - updated conf. file


 2005/01/27 Thomas Bretz

   * callisto.cc:
     - don't display starting pedestals when calibrating the signal

   * callisto.rc:
     - added some comments

   * star.rc:
     - according to Daniel M. we have agreed to change the default
       cleaning level to adapt it roughly to the new pedestal

   * manalysis/MCerPhotPix.[h,cc], mpedestal/MPedPhotPix.h:
     - as agreed in the current calibration meeting we truncate the
       precision of the stored floating point values to roughly 0.5%
       to gain storage space in the calibrated data (the gzip
       algorithm compresses this stuff nicely - we gain roughly 15%)

   * mhflux/MHFalseSource.cc, mjobs/MJCalibration.cc, mjobs/MJStar.cc:
     - added a comment

   * mjobs/MJCalibrateSignal.cc:
     - changed title of unsuitable plot
     - removed plot for calibrated fundamental pedestal as agreed in the
       current calibration meeting (we may change this to an option)

   * mjobs/MJPedestal.[h,cc]:
     - changed the name of the "PedDiff" plot to be consistent with the
       other tab-name
     - changed the options for the display
     - removed the dependancy of fDisplayType and the SetDataCheck
       option
     - added option kDisplayNone
     - renamed enums from k*Display to kDisplay*

   * mhist/MHCamEvent.cc:
     - trying to get rid of the Err_%p workaround.

   * manalysis/MCerPhotEvt.[h,cc]:
     - added some '!' signs in the comment line. This is quick hack
       to gain I/O speed and storage space. We will soon have new
       containers for the calibrated data and image cleaning
       information.
     - changed version number to 6

   * mhist/MHCamera.cc:
     - added member function SetAllUsed



 2005/01/26 Abelardo Moralejo

   * macros/mccalibrate.C
     - removed creation and writing out of MSrcPosCam. It is more 
       flexible to add it in starmc2.C if needed (for wobble mode for
       instance).

   * mtemp/mpadova/macros/RanForestDISP.C
     - added possibility to set the number of hadrons read in for the 
       training. Added some missing containers to output.
     - Fixed rule concerning the DISP, by adding parentheses. Remember
       that MHMatrix::AddColumn  does not respect the usual precedence
       of operations!

   * mtemp/mpadova/macros/resize.C
     - added to output missing container MRawEvtHeader

   * macros/starmc2.C
     - added missing container MImagePar. Added possibility to set the
       source position on the camera (for the wobble mode).



 2005/01/25 Abelardo Moralejo

   * macros/starmc2.C
     - modified to allow double (train / test) output, for g/h 
       separation studies. Added several missing containers to output.



 2005/01/25 Thomas Bretz

   * mcalib/MCalibrationIntensityChargeCam.cc:
     - fixed a lot of 'default argument'-warnings to make to code
       compile

   * mbase/MParContainer.[h,cc]:
     - moved the empty Reset() function into the source file and
       added a lot of comments
     - implemented new helper-functions: GetNewObject

   * mbase/MContinue.[h,cc]:
     - allow to use MFilter-classes as filter in ReadEnv

   * mbase/MTaskEnv.cc:
     - replaced some code by GetNewObject

   * mhbase/MFillH.cc:
     - handle DrawOption "same" to be able to draw to the same pad

   * mhbase/MH.[h,cc]:
     - added same-argument to DrawSame

   * mhbase/MH3.cc:
     - removed some obsolete comments
     - remove own drawing options before calling fHist->Draw

   * mimage/MHHillas.cc, mimage/MHImagePar.cc:
     - first try of implementing 'same' drawing option



 2005/01/25 Markus Gaug

   * mcalib/MCalibrationPattern.[h,cc]
   * mcalib/MCalibrationPatternDecode.[h,cc]
     - New Decoder task for the calibration pattern from the 
       digital modules
 
   * mraw/MRawEvtHeader.[h,cc]
     - remove decoding from this class   

   * mcalib/MCalibrationIntensityHiLoCam.[h,cc]
     - remove, because one needs anyhow much more than one data file
       to get a reasonable conversion factor

   * mcalib/MCalibCalcFromPast.[h,cc]
     - remove MCalibrationIntensityHiLoCam

   * mbadpixels/MBadPixelsPix.h
   * mbadpixels/MBadPixelsCam.cc
     - changed one function name to GetUnsuitableCalLevel()

   * mjobs/MJCalibration.cc
     - replace all Y-titles with units "FADC counts" by "FADC cnts" and 
       "FADC slices" by "FADC sl.", otherwise they range out of the 
       canvas.

   * mcalib/MCalibrationChargeCalc.cc
     - ensure to set the color correctly in MCalibrationBlindCam

   * callisto.rc
     - put space for 
       MJCalibrateSignal.ExtractSignal.HiGainFirst:     2
       MJCalibrateSignal.ExtractSignal.HiGainLast:     16


 2005/01/24 Abelardo Moralejo

   * manalysis/MMcCalibrationUpdate.cc
     - Adapted to changes in MCalibrateData from 2005/01/16: moved 
       qepix.SetAvNormFFactor(1.)  from Process() to ReInit(). Due to 
       this, for the last week you may have found trouble running 
       starmc.C and mccalibrate.C.



 2005/01/24 Thomas Bretz

   * callisto.cc:
     - removed non-existing option from output

   * manalysis/MEnergyEstimate.[h,cc]:
     - added ReadEnv member function

   * mbase/MContinue.[h,cc]:
     - implementes AllowEmpty option to support doing nothing 
       defined in resource file
     - added a first draft to allow usage of class-filters instead of 
       rules

   * mbase/MMath.[h,cc]:
     - implemented a function suggested by Florian Goebel which reduces
       the precision of a float to ~0.5% which is usefull for data
       compression by gzip

   * mdata/MDataChain.[h,cc]:
     - implemented ReadEnv member function

   * mfbase/MF.cc:
     - a small enhancement to the output in ReadEnv

   * mhbase/MH3.cc:
     - fixed call to FindObject in Paint which caused a profile to be
       displayed incorrectly in case it was read from a file

   * mhflux/MHFalseSource.[h,cc]:
     - removed old obsolete unsused Significance functions
     - implemented ReadEnv member function
     - set false source binning only if BinningFalseSource is not default
     - check parlist for BinningAlpha

   * mjobs/MJCalibrateSignal.cc, mjobs/MJCalibration.cc, 
     mjobs/MJPedestal.cc:
     - added an additional check and output if output-file could
       not be opened

   * mjobs/MJob.h:
     - added some Getter

   * mhbase/MBinning.[h,cc].
     - added nen member functions
       + SetEdgesRaw
       + ReadEnv
       + Print

   * mhflux/MAlphaFitter.[h,cc]:
     - added new member function ReadEnv

   * mhflux/MHAlpha.cc:
     - fixed a problem when fitting without off-data

   * mbase/MStatusDisplay.cc:
     - implemented a further fix in DrawClonePad to make sure
       that the X-stuff doesn't get confused while still
       GetDrawOption returns the correct value



 2005/01/24 Markus Gaug

   * msignal/MExtractTimeAndChargeDigitalFilter.cc
     - introduced security check for the times fine-tuning to catch 
       outliers because of deviating pulse forms.

   * msignal/MExtractTime.cc
   * msignal/MExtractor.cc
     - lowered default high-gain saturation limit to 250
     - removed Clear() for two results containers at beginning of Process()

   * msignal/MExtractor.h
     - Take fHiLoLast into the Clone()

   * msignal/MExtractedSignalPix.cc
     - change sign of default initializer to negative.



 2005/01/23 Markus Gaug
  
   * mhcalib/MHCalibrationHiLoCam.[h,cc]
     - updated to new logain switch setting in MExtractTimeAndCharge
     - extract also high-gain vs. low-gain time difference

   * mhcalib/MHCalibrationCam.[h,cc]
   * mhcalib/MHCalibrationChargeCam.[h,cc]
     - updated reading of conf. values in ReadEnv

   * mhcalib/MCalibrationHiLoCam.[h,cc]
   * mhcalib/MCalibrationHiLoPix.[h,cc]
   * mhcalib/MCalibrationIntensityHiLoCam.[h,cc]
     - updated accordingly

   * mbadpixels/MBadPixelsCam.cc
     - removed some bugs in the Print() function



 2005/01/22 Markus Gaug

   * mhcalib/MHCalibrationRelTimeCam.cc
     - display in the averaged times display the event-by-event averaged
       absolute arrival times instead of the rel. ones which do not make 
       much sense.

   * mhcalib/MHCalibrationTestTimeCam.[h,cc]
   * mhcalib/MHCalibrationTestCam.[h,cc]
     - update to the latest changes in analysis chain.

   * mjobs/MJExtractCalibTest.[h,cc]
     - update to the latest changes in analysis chain.
    
   * mjobs/MJCalibration.cc
     - replace title of that times-averageing Tab from "RelTimes" by 
       "Times".

   * macros/calibration.C
     - update to changes in MJExtractCalibTest



 2005/01/21 Markus Gaug

   * msignal/MExtractTimeAndChargeDigitalFilter.cc
   * msignal/MExtractTimeAndChargeSlidingWindow.cc
   * msignal/MExtractTimeAndChargeSpline.cc
     - use the max. bin content for low-gain extraction
   
   * callisto.rc
     - add place for LoGainSwitch



 2005/01/21 Florian Goebel

   * mfbase/MF.cc:
     - include a (TString) cast to make it compilable with gcc2.95.3



 2005/01/21 Thomas Bretz

   * mpointing/MSrcPosFromModel.cc:
     - fixed crash in Process if no pointing model available

   * mpointing/MPointingPosCalc.h:
     - added missing initialization of data members

   * mastro/MObservatory.[h,cc]:
     - added ReadEnv member function

   * mbase/MTaskEnv.h:
     - added (=0) default to SetDefault



 2005/01/20 Markus Gaug

   * mcalib/MCalibrationIntensityChargeCam.[h,cc] 
     - added function to display fluctuations of a variable

   * mcalib/MCalibrationPix.cc
     - fixed small bug in High-gain mean error.

   * msignal/MExtractTimeAndCharge.[h,cc]
     - introduced high-gain maximum limit before extracting the low-gain



 2005/01/20 Thomas Bretz

   * mbase/MContinue.[h,cc]:
     - added ReadEnv member function to support resource files

   * mbase/MEvtLoop.[h,cc]:
     - added a second arguemtn to reading of the resource file by name
     - added some checks when reding the resource file and the
       parameter list hasn't been setup

   * mbase/MParList.cc:
     - removed an obsolete output from ReadEnv

   * mfbase/MF.[h,cc]:
     - added ReadEnv member function to support resource files
     - added kAllowEpty flag to support skipping by resource files

   * mfbase/MFilterList.cc:
     - added a comment

   * mbase/MStatusDisplay.[h,cc]:
     - fixed Write to be compatible with root 4.02/00
     - worked around a problem with reading the DrawOption correctly
       from a stored status display



 2005/01/19 Thomas Bretz
 
   * mbase/MMath.cc:
     - added a comment to SignificanceLiMa, made by Daniel Mazin
     - also check for b==0

   * mhflux/MAlphaFitter.[h,cc]:
     - fixed significance calculation in case of on-off data
     - added fScaleFactor

   * mhflux/MHAlpha.[h,cc], mhflux/MHFalseSource.cc:
     - handle scale factor in case of on-off observations

   * mjobs/MJPedestal.cc:
     - aligned screen output to 80col rule

   * mtrigger/MFTriggerPattern.[h,cc]:
     - assigned correct integers to enum
     - made sure that the enum is correctly handled in all function by
       applying a r-value to the condition



 2005/01/19 Abelardo Moralejo

   * mtrigger/MFTriggerPattern.[cc,h]
     - Changed Bool_t arguments of all setters to new type Prescale_t
       defined by N. Galante (see below)



 2005/01/19 Nicola Galante

   * mtrigger/MFTriggerPattern.[cc,h]
     - added in header enum Prescale_t { kUnPrescaled = kFALSE, 
       kPrescaled   = kTRUE }; 

     - fixed a bug in Process, both fMaskRequiredUnprescaled and
       fMaskRequiredPrescaled are checked simultaneously.



 2005/01/18 Markus Gaug

   * mcalib/MCalibColorSet.cc
     - fixed the remaining day of data without colour in project name
       neither in digital modules bits.

   * mpedestal/MPedCalcPedRun.[h,cc] 
     - catch the continuous light runs with MRawRunHeader::fRunType 
       == kRTCalibration to be recognized as effective pedestal runs.


 2005/01/18 Thomas Bretz

   * mbase/MMath.cc:
     - added a check for s==0 in SignificanceLiMa, because otherwise the 
       result is NaN

   * mfbase/MF.cc:
     - removed an obsolete debugging output from the latest changes

   * mhflux/MHAlpha.cc:
     - reset fHTime, too



 2005/01/18 Nicola Galante

   * mtrigger/MFTriggerPattern.cc:
     - added parentesis to ensure bitwise operators priority
       with respect to logical operators.


 2005/01/17 Thomas Bretz
 
   * manalysis/MEnergyEstimate.[h,cc]:
     - changed to be more general. Energy is now estimated from a rule
       (MDataChain) The default behaviour is still the same as before
       to copy the energy from MMcEvt

   * mbase/MTask.cc:
     - changed to write out the descriptor istead of the task-name only
       in Pre/PostProcess

   * mhbase/MH.cc:
     - initialize fSerialNumber and fNumExecutions in constructor

   * mhflux/MHFalseSource.cc:
     - according to a suggestion by Markus Mayer I have changed
       \\sigma_{max}=f to S_{max}=f\\sigma

   * mjobs/MJCalibrateSignal.cc:
     - as discussed with Markus and Abelardo fixed the MFTriggerPattern
       call.

   * mtrigger/MFTriggerPattern.cc, mtrigger/MTriggerPattern.cc,
     mtrigger/MTriggerPatternDecode.cc:
     - added some comment about file version

   * mbase/MFilter.h:
     - implemented new virtual function GetDataMember like in MData

   * mbase/MTask.[h,cc]:
     - when filter is added ass GetDataMember to branch list (should be
       obsolete, because filter is in the tasklist, but it should not hurt)

   * mfbase/MF.[h,cc], mfbase/MFDataChain.[h,cc], mfbase/MFDataMember.[h,cc],
     mfbase/MFilterList.[h,cc]:
     - implemented new function GetDataMember



 2005/01/17 Markus Gaug

   * msignal/MExtractTimeAndChargeDigitalFilter.cc
     - remove a situation where the high-gain signal can be 0.

   * mtrigger/MTriggerPattern.h
     - fixed a bug in defining the trigger bit pattern (Have to start
       from 0 instead of 1)

   * mjobs/MJPedestal.cc
   * mjobs/MJCalibrateSignal.cc
     - fixed default extraction ranges.

   * callisto.rc
     - fixed default weights files for pedestal extraction for cosmcis

   * mcalib/MCalibrationChargeCalc.cc
     - removed program exit on occurrance of multiple colours without 
       intensity calibration



 2005/01/17 Nadia Tonello and Markus Gaug

   * mcalib/MCalibrationChargeCam.cc
     - some fit improvements in the function GetAveragedConvFADC2PhotPerArea



 2005/01/16 Markus Gaug

   * mhcalib/MHCalibrationCam.[h,cc], mhcalib/MHCalibrationRelTimeCam.[h,cc],
     mhcalib/MHCalibrationChargeCam.[h,cc]
     - put overflow management and results camera initialization into 
       base class
 
   * mcalib/MCalibrateData.[h,cc]
     - added possibility to re-normalize signal to equiv. photo-electrons
       ( SetSignalType(MCalibrateData::kPhe) ). 
       Default is old version in photons
     - speed up Process by storing pre-calculated calibration constants
       in arrays (needed 40% of CPU time of the eventloop before, now: 23%) 
     - added new pointer to different MCalibrationChargeCam used for 
       updates
     - added possibility to update the conversion factors using the 
       MCalibrationChargeCam stored in fCalibUpdate. 
     - added Print() function

   * mcalib/MCalibCalcFromPast.[h,cc]
     - committed final version, now also in the Makefile

   * mjobs/MJCalibrateSignalFromOutside.[h,cc], macros/calibration.C
     - committed final version, now able to calibrate interlaced cal. 
       events.



 2005/01/14 Hendrik Bartko

   * msignal/calibration_weights_UV.dat
     - fixed a small bug in the noise autocorrelation of the AB-noise

   * msignal/calibration_weights_blue.dat       
     - new weights file for blue calibration pulses

   * cosmics_weights.dat
     - fixed a small bug in the noise autocorrelation of the AB-noise



 2005/01/14 Markus Gaug

   * msignal/MExtractTimeAndChargeDigitalFilter.cc
     - added a small acceleration of the program, tested by Hendrik and
       me.

   * mpedestal/MPedCalcFromLoGain.cc
     - fixed one warning output which was not a real warning.

   * mcalib/MCalibrationChargeCam.cc
     - improved the function to calculate a mean conversion factor over 
       the whole camera including a fit to the values.

   * mpedestal/MPedCalcPedRun.[h,cc]
     - undo some changes introduced by TB on 2/12/04 which had made the 
       use of interlaced pedestal events impossible again. 
     - Use now the new class MTriggerPattern directly to recognize an 
       interlaced pedestal event.     



 2005/01/14 Thomas Bretz
 
   * callisto.cc:
     - fixed a typo in output

   * manalysis/MGeomApply.cc:
     - removed some old code already in comments
     - added possibility to switch off automatic processing (used
       if two different geometries around)

   * manalysis/MMultiDimDistCalc.cc, mhbase/MHMatrix.[h,cc],
     mranforest/MRanForestCalc.cc, mtools/MHSimulatedAnnealing.[h,cc],
     mtools/MSimulatedAnnealing.h:
     - added include of TVector because in newer root version it is 
       not included from TMatrix anymore

   * mbase/BaseLinkDef.h, mbase/MMath.[h,cc]:
     - to stick to the convention of TMath in newer root versions MMath
       changed from a class to a namespace

   * mbase/MEvtLoop.[h,cc]:
     - to support newer root version added a second Write-function
       with a const qualifier.

   * mdata/MDataChain.cc:
     - changed access to TMath for newer root versions

   * mfileio/MChain.[h,cc]:
     - implemented a workaround for teh changed LoadTree definition
       in newer root-versions

   * mfileio/MReadTree.cc:
     - fixed a bug in CheckBranchSize caused some warning to be
       displayed

   * mhflux/MHAlpha.cc:
     - removed check for fResult in PostProcess (it is always there)

   * msignal/MExtractTimeFastSpline.cc:
     - fixed delete --> delete[]

   * mbadpixels/MBadPixelsCalc.cc:
     - slight change to some output

   * mbase/MTaskEnv.h:
     - set name of default task when pointer is given

   * mjobs/MJCalibrateSignal.cc:
     - fixed handling of extraction range for various cases
     - fixed handling in case no arrival times are calculated

   * mjobs/MJCalibration.cc:
     - write at least a default arrival time calibration container
       to make further processing easier

   * mjobs/MJPedestal.[h,cc]:
     - fixed handling of extraction range for various cases



 2005/01/14 Abelardo Moralejo

   * mtemp/mpadova/macros/trainsubsample.C, RanForestDISP.C
     - added. Two macros used in the analysis of real wobble mode data.



 2005/01/14 Daniela Dorner

   * macros/sql/filldotrun.C
     - added new arehucas-verions and changed code accordingly



 2005/01/13 Abelardo Moralejo

   * macros/starmc.C
     - Added (commented) lines for the case of gamma MC source off-axis 
       (wobble mode). Changed default extractor to 
       MExtractFixedWindowPeakSearch.



 2005/01/13 Raquel de los Reyes

   * callisto.rc
     - introduced several configuration lines 
       (format: "class".DataCheck: No) to allow callisto 
       to run with rawfiles. Necessary for the datacheck.



 2005/01/13 Markus Gaug

   * mcalib/MCalibColorSet.cc
     - fixed runs 45605-45609

   * msignal/MExtractor.h
     - take out the variable fHiLoLast from the streamer. Will be 
       initialized in the ReInit(), if needed

   * msignal/MExtractor.cc
     - Initialized fHiLoLast to 0. Otherwise, a second eventloop with 
       the same extractor may have that variable set to a certain 
       value, even if not desired. The ReInit() fixes fHiLoLast, if 
       different from 0.

   * mpedestal/MExtractPedestal.cc
   * mpedestal/MPedCalcPedRun.cc
   * mpedestal/MPedCalcFromLoGain.cc
     - fix the output of the print to the correct value for the case 
       that the extractor is used.


 2005/01/13 Thomas Bretz

   * callisto.rc:
     - fixed some typos

   * mbase/MLog.cc:
     - unset DEBUG defined unintentionally

   * msignal/MExtractTimeAndChargeDigitalFilter.cc:
     - removed printing Weights in InitArrays, because otherwise
       the output of a full data calibration is totally crowded
       with numbers. Having the weights-file name on the display
       is perfectly enough.



 2005/01/12 Markus Gaug

   * mcalib/MCalibrationIntensityRelTimeCam.cc
     - added fit to improve averageing of the obtained results



 2005/01/12 Thomas Bretz

   * manalysis/MMatrixLoop.cc:
     - changed name to MRead because it now derives from MRead

   * mbase/MEvtLoop.[h,cc]:
     - implemented an option not to cascade the status display to the
       parameter list in PreProcess

   * mhflux/MHAlpha.cc:
     - changed name of "Significance" to "MinimizationValue"
     - return -significance to have a value being minimized

   * mhist/MHCamera.cc:
     - fixed output of text in legend

   * mhist/MHEvent.[h,cc]:
     - implemented new options for calibrated pedestals

   * mmain/MEventDisplay.cc:
     - fixed to support latest changes in MJCalibrateSignal

   * mpedestal/MPedPhotCam.cc:
     - added new options in GetPixelContent (4 and 5) to support
       a correct camera display

   * mjobs/MJCalibrateSignal.cc:
     - fixed display of calibrated pedestal and pedestal rms
     - changed order of plots. Now all interpolated values 
       (means all calibrated values which are bad-pixel treated)
       are displayed behind the unsuitable pixels.

   * mjobs/MJ*.cc:
     - corrected all wrong %06d to %08d to fit the number of
       digits in run-numbers



 2005/01/11 Abelardo Moralejo

   * mjobs/Makefile
     - added missing include -I ../mtrigger



 2005/01/11 Thomas Bretz

   * callisto.cc, callisto.rc:
     - changed such, that all three kinds of pedestals are calcualted 
       (usefull for studies)

   * manalysis/MMatrixLoop.[h,cc]:
     - derived from MRead
     - implemented some basic functions from MRead

   * mfilter/MFCosmics.[h,cc]:
     - allow setting of name of MPedestalCam

   * mhbase/MBinning.[h,cc]:
     - added SetEdges taking MParList as input

   * mhflux/MHAlpha.cc:
     - use new function from MBinning

   * mjobs/MJCalibrateSignal.[h,cc]:
     - changed to calculate all three kinds of pedestals
     - added filters for trigger pattern and illuminated events

   * mjobs/MJPedestal.cc:
     - give different names to tabs with different pedestals

   * mjobs/MJStar.cc:
     - on request changed used pedestal from "FromExtractor" to
       "FromExtractorRndm"

   * mtrigger/MTriggerPattern.h:
     - added Reset function

   * mtrigger/MTriggerPatternDecode.[h,cc]:
     - added using file format version number



 2005/01/11 Markus Gaug

   * mbadpixels/MBadPixelsPix.cc
     - updated documentation

   * mcalib/MCalibrationChargeCalc.cc
     - make pixels with deviating number of phes unreliable instead of 
       unsuitable (this was already thought so, but missed to modify
       at one place).

   * mhcalib/MHCalibrationCam.[h,cc]
     - add possible flag 'kInterlaced': default: kFALSE (like before)
     - added pointer to MRawEvtHeader to deny cosmics events in case
       of 'interlaced' flag is set. No changes for 'normal' calibration



 2005/01/10 Benjamin Riegel
 
   * mreport/MReportStarguider.cc, mreport/MReportStarguider.h
     - added two classes for implementing starguider-data 
 
   * mreport/ReportLinkDef.h
     - added class MReportStarguider to be linked:
       #pragma link C++ class MReportStarguider+;
     
   * mreport/Makefile
     - added file MReportStarguider.cc
   
   * merpp.cc
     - added two containers ( MReportStarguider, MTimeStarguider ) and
       added MReportStarguider to the list:
       w->AddContainer("MReportStarguider",  "Starguider");
       w->AddContainer("MTimeStarguider",    "Starguider");
       r->AddToList("MReportStarguider");
       
       
       
 2005/01/10 Markus Gaug

   * mcalib/MCalibColorSet.cc
     - return kFALSE for CL-run instead of kTRUE

   * mcalib/MCalibrationChargeCalc.cc
     - stop program in case that multiple colours are found in same file
       without intensity calibration

   * mbadpixels/MBadPixelsPix.h
     - added two new possible UnCalibrated_t flags for the hi-lo 
       calibration

   * macros/calibration.C
     - add the calls to:
          MArray::Class()->IgnoreTObjectStreamer();
          MParContainer::Class()->IgnoreTObjectStreamer();
       in main.

   * msignal/MExtractor.h
     - added Getter for fOffsetLoGain (was still missing)

   * mhcalib/MHCalibrationChargeCam.[h,cc]
     - remove pixels with histogram overflow only if their percentage 
       passes a certain limit (default: 0.5%). Before, every overflow 
       lead to an unsuitable pixel.

   * msignal/MExtractTimeAndChargeDigitalFilter.[h,cc]
     - implemented changes suggested by T.B. to speed up calculations
     - take fNameWeightsFile into the streamer, otherwise the Clone() 
       will always create an extractor with cosmic's weights. 
     - set fOffsetLoGain to a slighly higher value. The DF seems to 
       have a different offset since it uses different pulse shapes 
       and thus sightly different time definitions.



 2005/01/10 Thomas Bretz

   * mhvstime/MHVsTime.cc:
     - fixed a compiler error for gcc 2.95.3, see Mantis bug # 157

   * callisto.cc:
     - changed such, that the calculated pedestal is not random
       anymore
     - added code to calculate the pedestal also with randomization,
       but this code is for the moment commented out.

   * callisto.rc:
     - fixed naming of pedestal calculation

   * mbadpixels/MBadPixelsCalc.cc:
     - added a second check which compares the pedestal rms with
       the variance of the pedestal rms in the same sector

   * mjobs/MJCalibrateSignal.[h,cc]:
     - changed to calculate the pedestal from extractor without 
       randomization
     - added code to calculate the pedestal also with randomization,
       but this code is for the moment commented out.

   * mjobs/MJCalibration.cc:
     - made sure that pedcam always has the correct name
       (FIXME? Print a warning if it has the wrong name)

   * mjobs/MJPedestal.[h,cc]:
     - changed such, that it allows the calculation of pedestals
       in three different ways (fundamental, fromextractor, 
       fromextractorrndm)
     - therefor removed SetExtractorResolution and added
       + void SetExtractionFundamental()
       + void SetExtractionWithExtractorRndm()
       + void SetExtractionWithExtractor()

   * mjobs/MJStar.cc:
     - changed/fixed the input name of the MPedPhotCam used in the
       image cleaning

   * mraw/MRawFileRead.[h,cc]:
     - changed return type of OpenNextFile
     - in some cases return kERROR in OpenNextFile to abort the Processing
       with an error

   * mraw/MRawRunHeader.cc:
     - call a wrong file format version ERROR now

   * mhflux/MAlphaFitter.[h,cc]:
     - make sure that resetted values are returned if fit failed
     - set default scaling too kOffRegion

   * mhflux/MHAlpha.cc:
     - some small fixes to handling of MHAlphaOff

   * mhflux/MHFalseSource.[h,cc]:
     - implemented possible usage of MHFalsSourceOff for On-Off-Analysis
     - updated display accordingly
     - updated fit accordingly
     - removed catalog temporarily from plots until I found the
       reason for the strange crashes

   * mbadpixels/MBadPixelCalc.[h,cc]:
     - implemented the possibility to check the pedestal rms against the
       variance of the pedstal rms, too. The option is curretly switched
       off.
     - enahanced informal output in PreProcess

   * mbase/MTaskList.cc:
     - a little change to some output



 2005/01/10 Wolfgang Wittek

   * mfbase/MFEventSelector2.cc
     - check whether filter exists before calling 
       filter->CallPreProcess(parlist)



 2005/01/08 Markus Gaug

   * mhcalib/MHCalibrationHiLoCam.[h,cc]
   * mcalib/MCalibrationHiLoCam.[h,cc]
   * mcalib/MCalibrationHiLoPix.[h,cc]
   * mcalib/MIntensityCalibrationHiLoCam.[h,cc]
   * mhcalib/Makefile, mcalib/Makefile, mhcalib/HCalibLinkDef.h, 
     mcalib/CalibLinkDef.h
     - new classes for calibration of conversion high-gain vs. low-gain 

   * msignal/MExtractedSignalPix.[h,cc]
     - added data member fChisquare, set to -1. by Clear() and Reset()
     - raised version number by 1

   * msignal/MExtractedSignalCam.[h,cc]
     - added data member fNdf, set to -1 by default
     - new function: GetProb(const Int_t pixidx) to be called to 
       calculate the joint probability of Chisquare of pixel "pixidx"
       and fNdf
     - added documentation of data members
     - raised version number by 1

   * mhcalib/MHCalibrationChargeCam.cc
     - strengthened default histogram limits a bit, should accelerate
       the calibration by 20%.

   * mjobs/MJCalibrateSignalFromOutside.cc
     - fixed two lines to compile MCalibrateData with the recent changes
       by TB

   * msignal/MExtractTimeAndChargeDigitalFilter.cc
     - fixed two bugs introduced yesterday by TB.

   * msignal/MExtractTimeAndChargeSpline.cc
   * msignal/MExtractTimeAndChargeSlidingWindow.cc
     - adopt part of TB's speed accelerations into these classes

   * msignal/MExtractTimeAndCharge.cc
     - fix the low-gain extraction window for the case that the high-gain
       is saturated. Was skipped so far.



 2005/01/07 Markus Gaug

   * msignal/MExtractFixedWindowPeakSearch.[h,cc]
     - catch the case that more than 90% of the pixels have saturated 
       signals and fix the extraction window from the occurrance of 
       saturation, then. Important for high-intensity calibration events.


   * msignal/MExtractor.cc
   * msignal/MExtractTime.cc
   * msignal/MExtractTimeAndCharge.cc
     - added some documentation

   * msignal/MExtractTime.[h,cc]  
     - moved the empty function FindTimeHiGain() and FindTimeLoGain() to
       the header-file.
      



 2005/01/07 Abelardo Moralejo

   * mbase/MLog.h
     - Added #include "TSystem.h", otherwise gSystem is not recognized
       in Mac OSX.

   * manalysis/MMakePadHistograms.cc
     - Changed badcalc.SetNamePedPhotContainer to 
       badcalc.SetNamePedPhotCam



 2005/01/07 Thomas Bretz

   * mbadpixels/MBadPixelsCalc.[h,cc], macros/ONOFFAnalaysis.C:
     - replaced PedPhotContainer by PedPhotCam for consistency
   
   * mbase/MLog.[h,cc]:
     - return *this in Seperator()

   * mbase/MPrint.[h,cc]:
     - use new separator manipulator in PreProcess
     - fixed a type: separator instead of seperator

   * mfileio/MReadTree.[h,cc]:
     - added a check for branch->IsAutoDelete() in Notify()

   * mpedestal/MExtractPedestal.[h,cc]:
     - added new option 'fRandomCalculation' be used to switch on-/off-
       randomization in extractors, the default should be like before
     - added Process() calling Calc()

   * mpedestal/MPedCalcFromLoGain.[h,cc], mpedestal/MPedCalcPedRun.[h,cc]:
     - renamed Process() to Calc()

   * Makefile:
     - removed macros/calibration.C, which seems to have been
       unintantionally introduced.

   * manalysis/Makefile:
     - removed unintantionally introduced wrong include

   * msignal/MExtractTimeAndChargeDigitalFilter.cc:
     - a small acceleration discussed with Hendrik by replacing a
       counter in a loop by pointer arithmetics

   * mraw/MRawEvtPixelIter.cc:
     - fixed a bug in GetMaxIdxHiLoGain which caused the loop to
       read one byte behind the end of the array
     - fixed HasLoGain() which really did weired things, means accessing
       sometimes memory somewhere. This could theoretically result
       in assigning the wrong lo-gains to the wrong pixels.

   * mtrigger/MTriggerPattern.[h,cc], mtrigger/MTriggerPatternDecode.[h,cc],
     mtrigger/MFTriggerPattern.[h,cc]:
     - added

   * mtrigger/Makefile, mtrigger/TriggerLinkDef.h:
     - added MTriggerPattern, MTriggerPatternDecode and MFTriggerPattern

   * manalysis/MEnergyEstParam.h:
     - added a new function SetVariables calling SetCoeff

   * mcalib/MCalibrateData.[h,cc]:
     - removed data members for container names
     - added new lists for container names
     - allow to calibrate more than one pedestal container at the same
       time

   * mjobs/MJCalibrateSignal.cc:
     - changed accordingly



 2005/01/05 Thomas Bretz

   * mcalib/MCalibrationBlindPix.cc, mcalib/MCalibrationChargeCalc.cc,
     mhcalib/MHCalibrationRelTimeCam.cc:
     - made some output lines fit into convinience window of 80 col

   * mhist/MHCamEvent.cc:
     - make sure that x-axis has correct title

   * msignal/MArrivalTimePix.cc:
     - init error to -1 which is a obviously non valid value
     - fixed IsValid() which gave wrong results

   * msignal/MExtractTimeAndCharge.cc, msignal/MExtractedSignalPix.cc:
     - made sure that LoGain is flagged non-valid if no lo-gain exists
     - push time back at the edge of the allowed limits if outside

   * mbase/MGList.[h,cc]:
     - introduced new functions Delete and DeletePictures
     - SetBit(kMustCleanup) in constructor. Otherwise it is not
       correctly removed from ListOfCleanups

   * mbase/MLog.cc:
     - SetBit(kMustCleanup) if adding a plugin to make sure it is cleaned
       from the list

   * mbase/MParContainer.[h,cc]:
     - replaced return type of GetDescriptor by TString. The old version
       was terribly wrong! The returned memory wasn't removed as soon
       as the function returned.

   * mbase/MStatusDisplay.[h,cc]:
     - for security (thread safty) replaced all Form() by MString::Form()
     - implemented a more secure scheme of Closing (Clone()) this was
       a hint given by valgrind
     - set kMustCleanup for all new object in ClonePad not only kCanDelete
     - fixed some lines in pulldown-menus
     - add TPads to ListOfCleanups in ClonePad (workaround!)

   * mbase/MString.[h,cc]:
     - implemented a new static Form() function returning a copy of
       the result

   * mbase/MTask.[h,cc]:
     - replaced return type of GetDescriptor by TString. The old version
       was terribly wrong! The returned memory wasn't removed as soon
       as the function returned. 
     - Made GetDescriptor thread safe using MString::Form

   * mfileio/MWriteRootFile.cc:
     - Use GetDescriptor().Data() when using any kind of printf function

   * mhbase/MFillH.cc:
     - replaced string concatenation in all constructors by a Form 
       (also suggested by valgrind)

   * mhcalib/MHCalibrationRelTimeCam.cc:
     - made some output fit into 80col boundary




 2005/01/03 Thomas Bretz (HAPPY NEW YEAR)

   * mfht:
     - added new directory for hexagonal fourier transformation

   * mfht/HftIncl.h, mhft/HftLinkDef.h, mhft/MGeomCamMagicXT.[h,cc],
     mhft/MGeomCamMagicEnhance.[h,cc], mhft/MHexgaonalFT.[h,cc], 
     mhft/MHexagonalFreqSpace.[h,cc], mhft/MHexagonalFTCalc.[h,cc],
     mhft/Makefile:
     - added

   * macros/extendcam.C: 
     - added: tutorial for MGeomCamExtend

   * macros/hft.C: 
     - added: tutorial for MHexgonalFTCalc

   * macros/dohtml.C: 
     - updated

   * macros/rootlogon.C: 
     - updated

   * Makefile:
     - mhft added

   * mbase/MParContainer.cc, mbase/MLog.cc:
     - added a pre-compiler option for destructor debugging

   * mbase/MTaskList.[h,cc]:
     - added new function Replace

   * mdata/MDataChain.cc, mfbase/MF.cc:
     - shortened some output

   * mdata/MDataArray.[h,cc]:
     - added new function FindRule

   * mfileio/MReadTree.[h,cc]:
     - enhanced functionality to be able to read a TTree stored in memory
       instead of a file. Therefor introduced fTree which is now the pointer
       used and replaced fChain where TTree and TChain have common behaviour.
       Where they behave uncommon new code for the new case has been added.

   * mfileio/MWriteRootFile.[h,cc]:
     - enhanced functionality to be able to write a TTree into memory
       instead of a file. This were maily bugfixes for the case fOut==0, which
       must be handled correctly. Becarefull a tree filles without a file
       is not deleted automatically!
    
   * mhbase/MBinning.h:
     - added copy constructor
     - fixed buggy SetEdges(MBinning&) which didn't work

   * mhbase/MHMatrix.cc:
     - when adding a new rule check first if the same rule already exists

   * mhflux/MAlphaFitter.[h,cc]:
     - added new functionality to support on-off analysis
     - didn't touch existing code

   * mhflux/MHAlpha.[h,cc]:
     - added new functionality to support on-off analysis
     - replaced Fit() calls by newer function calls
     - added DrawAll - PRELIMINARY
     - fixed a bug when projecting - may be the reson for some strange
       crashes

   * mcalib/MCalibColorSet.cc:
     - fixed a stupid bug introduced last year, which caused MCalibColorSet
       in some circumstances to stop the eventloop when it shouldn't

   * mjobs/MSequence.cc:
     - expand path names