[cvsnt] Replacement for 'commit -r 2.0'?

Yongwei Wu wuyongwei at gmail.com
Sat Aug 12 13:06:23 BST 2006

Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.

On 8/12/06, Tony Hoyle <tony.hoyle at march-hare.com> wrote:
> Yongwei Wu wrote:
> > The official repository is on the Internet, but I want to have a local
> > repository for safety and in case I must work offline. OK?
> If you want to mirror a repository then fine but what has that got to do with
> the revision number?
> If you work offline you must ensure that no work is done on the main
> repository then sync back.  Revision numbers can't help there.

Imagine the scenario that most work is done on the outside repository.
Imagine that I need to view the whole history/diff info when working
offline. Imagine that some revisions are not committed locally
(possibly because the local sync is done much later than the creation
of the main repository) but I would like to increase the revision
number so that $Revision$ would not make diff show any difference.

Though it is trivial, I believe it is a valid use of 'cvs commit -r'.

> > Maybe it is good behaviour not to use this CVS feature. I simply fail
> > to see why it should be removed. Removing any existing features is a
> Because it has no legitimate use, encourages the misconception that revision
> numbers are somehow significant, and could potentially corrupt the repository.

See above. I would say it is legitimate use. And I have seen other
people in the list want 'cvs commit -r', though for a different
purpose. Whatever, removing an existing feature that still has some
use is not in my eyes the right thing to do.

The basic spirit of open source is belief in your user. Developers
give the solution, and users are expected to do the right thing.
Trying to educate users with every new release is something Microsoft
is likely to do. However, generally Microsoft does *not* break
backward compatibility.

> It was *never* a useful feature.  Bo is correct.  You shouldn't be using
> revision numbers for something that tags and branches are designed for.

You would be right if $Revision$ had never been invented, and CVSNT
did not have a related history with CVS. Since your need the revision
number for diff, and it is shown in $Id$, $Revision$, etc., it is
something the user will see and use, and may expect more when certain
circumstances arise.

Though not a show-stoppper, this kind of incompatibility is not a good
thing to convert people from CVS to CVSNT, either.

Best regards,

Wu Yongwei
URL: http://wyw.dcweb.cn/

More information about the cvsnt mailing list
Download the latest CVSNT, TortosieCVS, WinCVS etc. for Windows 8 etc.
@CVSNT on Twitter   CVSNT on Facebook