On Thu, 22 Feb 2001 21:04:35 -0500 Konstantin Olchanski <olchansk@ux1.phy.bnl.gov> wrote concerning ": Re: New Class BrTPCSequenceAdder": > > > > > > I think you misunderstood me. I ment to say that I did not bump up the > > > brat version number (which currently is 1-14-1). And I have set: > > > ClassDef(BrTPCSequenceAdder, 0). If I misunderstood you please notify > > > me again. > > > > I didn#t misunderstand you, I just warned you that non-persistent > > classes should not have version number different from zero. I > > understood what you wrote as: "if the class had affected other > > classes, then I would have bumped the version number", which would > > have been wrong, since the class is non-persistent, and so I though I > > could catch the error in the up-running, before it would cause > > grief. > > > It might be a good idea to write a perl script to check the version > numbers for consistency- enforce version number 0 for transient objects > and make sure the version number changes when the object changes. > Ideally, the version number should "change itself", maybe tied > to the CVS revision number? That would not be possible, and overkill. What matters is the data members of a class and thier order, not the methods. Coming ROOT 3, this will be even less of a problem, as long as you use the '*' argument for ROOTCINT. 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 2b29 : Fri Feb 23 2001 - 07:19:06 EST