Re: source code reference points

From: Rene Brun (Rene.Brun@cern.ch)
Date: Thu Jul 22 1999 - 09:24:24 MEST


Hi Matt,
Could you clarify your 2nd point about "Incompatible changes in the
I/O subsystem". Version 2.22 IS BACKWARD COMPATIBLE with 2.21.
Files written by 2.21 are readable by 2.22 and no changes at all
are required in the code. Only the files written by 2.22 cannot be read
by 2.21. The reasons for this are explained in the release notes.
I would appreciate if you could tell us asap what is the problem
as we are getting mails from users worried by your message.

The changes in the Physics vectors classes were decided at the
FNAL workshop. The two classes TVector3 and TLorentzVector are now
based on the CLHEP implementation. These changes and the corresponding
documentation are also described in the release notes.

Concerning your original message about source versionning, we do stamp
each release (source and binaries) with a unique version ID.
The version numbers and the date are indicated. Because I am away
for a few days, Fons changed the source tar file without incrementing
the version number to incorporate a change in CINT affecting a few
implementations (AIX xlC in particular). I believe that the minor
inconvenience in not incrementing the version id had more advantages
than disadvantages and I hope that this procedure was an exception
to our normal rule.

Sorry for this late reply. I only have intermittent access to the
Internet.

Rene Brun


On Wed, 21 Jul 1999, Matthew D. Langston wrote:

> Hi Masaharu,
> 
> Masaharu Goto wrote:
> > 
> > Would you explain specifically about the incompatiblity problems?
> > I know this is not the revision management problem, but I'd like to
> > know.
> 
> The ROOT maintainer for the group at SLAC which is having the
> incompatibility problems (which is keeping them at an older version of
> ROOT) told me that there are two reasons why they cannot upgrade to a
> newer version of ROOT:
> 
>   1) Incompatible changes with the Vector classes.
>   2) Incompatible changes in the I/O subsystem.
> 
> Although only a small part of their code (the Calorimetry subsystem)
> depends on the Vector classes, the rest of the collaboration is unable
> to upgrade to newer versions of ROOT because of the IO incompatibility.
> 
> --
> Matthew D. Langston
> SLD, Stanford Linear Accelerator Center
> langston@SLAC.Stanford.EDU
> 



This archive was generated by hypermail 2b29 : Tue Jan 04 2000 - 00:43:36 MET