IMPORTANT (was Re: BrDCClusterFinder does not compile!!!)

From: Christian Holm Christensen (cholm@hehi03.nbi.dk)
Date: Sun Aug 05 2001 - 07:50:16 EDT

  • Next message: Christian Holm Christensen: "Re: updated BrZdcRdoModule"

    Hi Kris, Peter, Pawel, et al, 
    
    On Sat, 4 Aug 2001 00:17:23 +0200 (CEST)
    "Peter H. L. Christiansen" <pchristi@nbi.dk> wrote
    concerning ": Re: DC and TPC track mathing":
    > NB! :
    > I wanted to look at some of this stuff but a certain Monsieur Staszel has
    > made a unhealthy brat-update. You forgot to update BrDCHitPosition, so
    > brat does not compile...
    
    On Sun, 05 Aug 2001 01:46:33 -0500 (CDT)
    Kris Hagel <hagel@cyclotronmail.tamu.edu> wrote
    concerning ": BrDCClusterFinder does not compile!!!":
    > Hello,
    > I updated to the "latest" version of BRAT and compiled and saw to my great
    > astonishment that BrDCClusterFinder does not compile.  It was looking for 3
    > undeclared methods.
    > 1. BrDCHitPosition::GetTdc()
    > 2. BrDCHitPosition::SetTdc(Int_t)
    > 3. BrDCHit::SetTdc(Int_t, Int_t, int)
    
    
    I took what I believe to be Pawels latest changes from his working
    directory and commited the files 
    
      BrDCHit.h 
      BrDCHit.cxx 
      BrDCHitPosition.h 
    
    I also cleaned-up the class a bit: 
    
      1) Members was named something like ftdc, rather than fTdc 
      2) Obfuscated inline code was moved to after the class definition
         with inline linkage. 
      3) Class version of BrDCHitPosition bumped to 2 due to the addition
         of the data member fTdc. 
    
    I've said this to people a thousand times, it's written in The Guide,
    and numerous emails on the list contains detailed information, so
    there's absolutly no excuse: 
    
      WHEN CHANGING THE CLASS DEFINITION OF A PERSISTENT CLASS, THAT IS
      FOR A CLASS WITH CLASS VERSION GREATER THAN 0, THE VERSION NUMBER
      MUST BE INCREMENTED!!! 
    
    This is really bloody important!!!  Failure to follow that procedure
    is a serious problem that causes tremendous havoc.  If I were to
    compare it to a crime, it's be nothing less than treason!!!  I'm not
    kidding. 
    
    Ok, so you may have caught this if you were to commit the changes, but
    seeing how you forgot to commit the subdirectory all together, I
    speculate that what you intented to do was to commit what you had, as
    it were when I found it. 
    
    Always make sure you've commited all your changes.  The best way to do
    that is to cd to the top of the source tree and do 
    
      cvs -n update -A -P -d 
    
    This will tell you if any files is modified in your working directory
    compared the one in the CVS respository.  Also, if you've added or
    removed files, that will also show up.  
    
       R        File is removed in working directory, but not in CVS -
    	    needs to be commited
       A        File is added to working directory, but not in CVS -
    	    needs to be commited
       M        File is modified in working directory, but not in CVS -
    	    needs to be commited
       U        File is modified in CVS, but not in working directory -
    	    needs to be retrieved from CVS 
       ?        File is not registered in the CVS. 
    
    It's far from the first time this has happened, so I'd like to direct
    your attention to The Guide.  There's a section there that describes
    the procedure for commiting changes to BRAT (and BREG, BRAG, BROP, and
    so on), and it should not be taken as guidiance but as a rock-fast
    procdure to be followed.  Under each section describing the different
    class groups, there's a short description on what to do with the class
    version number.  
    
    Pawel, the class BrDCHit is non-persistent (can not be written to
    disk) - is that really the intention?   Wouldn't you like to be able
    to store BrDCHits on disk, just like you like to store BrTpcCluster
    objects on disk.  Please look into that.  Also please check that I had
    the right version of the classes.   That means you have to move the
    ones you've changed out of the way (rename them to BrDCHit.h.back,
    BrDCHit.cxx.back, and BrDCHitPosition.h.back) and do an update.  This
    is important, because otherwise you may overwrite my changes. 
    
    I'd like to make it absolutly clear that I'll not make it a habit of
    checking in your changes for you.  However, I needed to look at
    something over the weekend and I needed the latest BRAT for that, so I
    bloody well had to be able to build BRAT.  Hence the commit. 
    
    >From ChangeLog: 
    
    2001-08-05  Christian Holm Christensen  <cholm@hilux12.nbi.dk>
    
    	* test/.cvsignore: Corrected line from chkparams to test/chkvparams
    
    	* data/track/BrDCHit.h, data/track/BrDCHit.cxx:
    	Took what I beleive to be Pawels latest changes (see the corresponding
    	email to brahms-dev-l) and commited that.  Also cleaned up the class.
    
    	* data/track/BrDCHitPosition.h:
    	Took what I beleive to be Pawels latest changes (see the corresponding
    	email to brahms-dev-l) and commited that.  Also cleaned up the class.
    	Most importantly, I incremented the class version.
    
    2001-08-04  Djamel Ouerdane  <ouerdane@hilux12.nbi.dk>
    
    	* modules/calib/tof/BrTofTdcOffsetCalModule.cxx:
    	minor update in the Begin method
    
    2001-08-03  zdc  <zdc@hilux12.nbi.dk>
    
    	* modules/vertex/BrVertexModule.cxx:
    	The functions names for getting times from ZDC have been re-named in BrVertexModule.cxx file, accordingly to changes in BrZdcRdoModule class
    
    	* data/calib/Makefile.am, data/calib/Include.h, data/calib/LinkDef.h, data/calib/BrCalibrationParamsZDC.h, data/calib/BrCalibrationParamsZDC.cxx:
    	ZDC calibration class BrCalibrationParamsZDC.h has been added to BRAT
    
    	* data/rdo/BrZdcRdo.h, data/rdo/BrZdcRdo.cxx, modules/rdo/BrZdcRdoModule.h, modules/rdo/BrZdcRdoModule.cxx, modules/rdo/BrTileRdoModule.cxx, modules/rdo/BrMultRdoModule.cxx:
    	The methods to read low-gain ZDC ADC have been added to BrZdcRdoModule
    
    2001-08-03  Pawel Staszel  <staszel@hilux12.nbi.dk>
    
    	* modules/util/BrDriverDC.cxx: some correction for planes position.
    
    	* modules/track/dc/BrDCClusterFinder.h, modules/track/dc/BrDCClusterFinder.cxx, modules/track/dc/BrDCTrackingModule.h:
    	low and up limits on number of hits added.
    
    	* modules/track/dc/BrDCTrackingModule.cxx:
    	up and low limits on hits number has been added.
    
    	* modules/track/dc/BrDCTrackingModule.cxx:
    	Fill input-node with detector tracks has been added.
    
       
    Yours, 
    
    Christian  -----------------------------------------------------------
    Holm Christensen                             Phone:  (+45) 35 35 96 91 
      Sankt Hansgade 23, 1. th.                  Office: (+45) 353  25 305 
      DK-2200 Copenhagen N                       Web:    www.nbi.dk/~cholm    
      Denmark                                    Email:       cholm@nbi.dk
    



    This archive was generated by hypermail 2b30 : Sun Aug 05 2001 - 07:53:05 EDT