Re: renaming under CVS

[Noel Yap]

--- "Greg A. Woods" <address@hidden> wrote:
> [ On Sunday, February 24, 2002 at 20:46:05 (-0800),
> Noel Yap wrote: ]
> > Subject: Re: renaming under CVS
> >
> > I'm starting to think about a scheme where CVS
> would
> > go through a filename mapping if the usual archive
> > file isn't found (I think this is how Meta-CVS
> works).
> >  The mapped archive file would only exist if
> there's
> > been a rename or move.  I don't think such a
> scheme
> > would _fundamentally_ break backwards
> compatibility.
> 
> If an old version of CVS cannot recover a release,
> or indeed if plain
> RCS commands on the RCS files cannot revcover a
> release, then you've
> broken backwards compatability.

1. No, I've broken forward compatibility.
2. The internal representation of the repository is
just that, internal, which means no other tool should
rely on it being a certain way.  This has been further
supported by the fact that CVS has internalised the
RCS code.

>  If you want to
> rename things in that
> way then you'll have to do it post
> checkout/update/export.

This is not the right place to do it.  Essentially,
you'd be wrapping all the commands, which is what you
propose now (and can't seem to break away from). 
Instead, if an added level of indirection were
created, all commands should still work and they
should work seemlessly with regards to renames.  Of
course, the devil is in the details so some command
changes will probably need to be made as well (eg,
what'll now happen when updating under these
circumstances).

Noel

__________________________________________________
Do You Yahoo!?
Yahoo! Sports - Coverage of the 2002 Olympic Games
http://sports.yahoo.com