Re: BRAT 1.12.5

From: hagel@comp.tamu.edu
Date: Tue Jan 23 2001 - 15:46:38 EST

  • Next message: Christian Holm Christensen: "Re: BRAT 1.12.5"

    Hello,
    #3 seems to be inconsistent with #2, but I am sure everyone gets the point.  From
    my point of view, I am lukewarm about a) in the short term because we are now
    trying to get out a paper very quickly and a change to root 3.0 would be
    disruptive.  Once we get to that point though, I would strongly suggest b)
    instead of c) because I don't think we should be in a position to have to rewrite
    all of the root files.
    
    Kris
    
    Konstantin Olchanski wrote:
    
    > Let me summarize the I/O situation, plase correct me if I am wrong:
    >
    > 1) until now we were/are creating root files unreadable with root 3.0
    > 2) with the new BRAT, we started creating root files readable with root 3.0,
    >    and we still can read the old root files.
    > 3) we cannot start using root 3.0 because it cannot read our root files.
    >
    > ?From this follows:
    >
    > a) everybody should immediately switch to the new "root 3.0" compatible BRAT,
    > b) root 3.0 should be fixed so it can read our old root files, or
    > c) we need a program to convert our old root files to a root 3.0
    >    compatible format.
    >
    > Correct?
    >
    > K.O.
    >
    > On Tue, Jan 23, 2001 at 06:10:05PM +0100, Christian Holm Christensen wrote:
    > ? Hi all,
    > ?
    > ? I've commited a few changes to the CVS, and bumped the version number
    > ? to 1.12.6 (CVStag BRAT-1-12-6).
    > ?
    > ? Pleae ignore BRAT 1.12.5 (it's buggy). Yury that's the one you
    > ? got. Update as soon as you get this mail. Oh, and I do think we should
    > ? make ROOT 2.25/03 "pro" (Uh, I thought it was).
    > ?
    > ? Changes are:
    > ? * BrModule::fStatus can be one of kOk, kStop, kFailure, kAbort, there
    > ?   meaning as outlined by Konstantin in previous mail. They are also
    > ?   descripted in the class documentation of BrModule
    > ? * BrModuleContainer changed to react according to fStopLevel, set by
    > ?   SetStopLevel.
    > ? * BrIOModule::fStatus renamed to BrIOModule::fIOStatus to prevent
    > ?   confusion with BrModule::fStatus. That also affects BrEventIO,
    > ?   BrRawDataInput, BrRawDataOutput, and BrGeantInput.
    > ? * BrModule and BrIOModule streamers written by hand, and class version
    > ?   bumped to 2 (why it was 1 - should be 0 - in the first place is
    > ?   beyond me).  This means that new files created, will not contain all
    > ?   the data members of BrModule and BrIOModule (which they of course
    > ?   shouldn't). However, this has some side-effets. We can not use ROOT
    > ?   version 3.00/01+ to read old files. In other words, if you're
    > ?   intrested in reading old files made with ROOT 2.2x/xx then you
    > ?   should not compile BRAT with ROOT 3.00/xx. However, files written
    > ?   with BRAT 1.12.6 + ROOT 2.25/xx will be readable by BRAT 1.12.6 +
    > ?   ROOT 3.00/xx.
    > ?
    > ?   Here's a table:
    > ?
    > ?
    > ?     Creator/     | BRAT ?1.12.6  | BRAT ?=1.12.6 | BRAT ?=1.12.6
    > ?     Reader       | ROOT ?3.00/00 | ROOT ?3.00/00 | ROOT ?=3.00/00
    > ?   ---------------+---------------+---------------+---------------
    > ?    BRAT ?1.12.6  |               |               |
    > ?    ROOT ?3.00/00 |   GOOD        |    BAD        |    BAD
    > ?   ---------------+---------------+---------------+---------------
    > ?    BRAT ?=1.12.6 |               |               |
    > ?    ROOT ?3.00/00 |   GOOD        |    GOOD       |    BAD
    > ?   ---------------+---------------+---------------+---------------
    > ?    BRAT ?=1.12.6 |               |               |
    > ?    ROOT ?3.00/00 |   BAD         |    GOOD       |    GOOD
    > ?   ---------------+---------------+---------------+---------------
    > ?
    > ?   As you can see, there's only one place where we break backward
    > ?   compatibility (and I hope - with help from Brun - to resolve that
    > ?   one).
    > ?
    > ? Oh, and I'd like to direct your attention to the application
    > ? dbapps/BramCalibSession.cxx by Flemming, which is an excellent
    > ? example of a well structured BRAT application. This should serve as an
    > ? inspirration to everyone.
    > ?
    > ? 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
    > ?
    > ?
    >
    > --
    > Konstantin Olchanski
    > Physics Department, Brookhaven National Laboratory, Long Island, New York
    > olchansk@bnl.gov
    



    This archive was generated by hypermail 2b29 : Tue Jan 23 2001 - 15:47:42 EST