"Brian Beaudet" <bbeaudet@efficiencylab.com> writes:
> The only thing preventing me from going with a single repository is
> the global revision number. If I start with one project and observe
> the number increasing sequentially, no problem. But when I add a
> new project and start where the last revision number left off (plus
> one of course) then it just seems strange. I was hoping to use the
> revision numbers in my application's version number (1.0.56 or
> 1.1.234) or something like that. How else does one keep track of
> that if not by the revision number?
This is a classic complaint of the global revision number. Many folks
feel better knowing that if they have revision M in their working
copies, and revision N is the latest, then exactly N - M changes to
their project have occured.
And the only response I can offer is that the change must occur within
yourself. People seem to think that an increasing revision number
demonstrates something -- anything -- about the state of their
project. But this is a kinda shaky place to stand. And if you ever,
*ever* use the changes in this number in a progress report to upper
management, you should be shot. :-)
That metric is an illusion -- three commits could contain { small
change, reversion of that change, different small change }. But a
different single commit could contain { entire re-write of the program }.
The numbers just don't matter. A project's own "version" should be
managed externally to the version control system.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Mar 25 16:49:52 2004