Re: pvcs to cvs and magic branches

[Jim Gallagher]

On Thursday 12 July 2001 13:56, you wrote:
> [ On Thursday, July 12, 2001 at 13:31:01 (-0400), Eric Siegerman wrote: ]
>
> > Subject: Re: pvcs to cvs and magic branches
> >
> > It's CVS, not RCS, that wants branch numbers to be even.  I'm not
> > sure why.  (Exception: the vendor branch is 1.1.1 by default.)
>
> The "magic" in CVS branch numbers is that they do not ever point to real
> revisions.  I.e. they are "magic" in the sense that they are virtual and
> indicate that a tag is a branch tag and they point to where a branch
> revision will be created when, and if, a delta to that file on that
> branch is created; and they point to the branch from the top revision
> will be extracted when CVS deals with operations on a sticky branch.
>
> CVS magic branch numbers facilitate "lazy branching" in CVS and thus
> make CVS often far more efficient at branching than RCS (or SCCS) alone.


I'm sorry, but I still don't understand. What would be affected in CVS if odd 
numbered branches were created during the conversion from PVCS? Would I not 
be able to get working copies or commit? Is this a 'dusty corner' that I 
probably will not need to know about?