[cvsnt] Re: "cvs commit -r " problem

Rick Genter rgenter at silverlink.com
Fri Jul 22 19:57:44 BST 2005


> -----Original Message-----
> From: cvsnt-bounces at cvsnt.org [mailto:cvsnt-bounces at cvsnt.org] On
Behalf Of Tony Hoyle
> Sent: Friday, July 22, 2005 2:48 PM
> To: cvsnt at cvsnt.org
> Subject: Re: [cvsnt] Re: "cvs commit -r " problem
> 
> Rick Genter wrote:
> 
> > 
> > Personally, I like the system that subversion uses: 1, 2, 3, 4 :-).
> 
> Theirs ends up like 1,2,24,48,92, due to there being one version for
the 
> entire repository, so individual files jump around a lot.

It's a nit, but actually even though a particular file changed on
repository revisions 1, 2, 24, 48, 92, etc., revisions 3 through 23, 25
through 47, 49 through 91, etc. of the file exist, they're just the same
as the previous revision. So it's not an error to fetch revision 5 of
the file in question - you just get the same content as if you fetched
revision 2 (or revision 9 or revision 23).

I haven't used subversion yet - just read about it a bit - and I'm not
keen on how it implements tags and branches. I know that you can get the
status of a file to see at which repository revisions it changed, so you
don't have to go digging too far to find the "relevant" revision numbers
for a file, but it's not as trivial as CVS's 1.1, 1.2, 1.3, etc. (Which,
technically, you can't trust anyway because someone could have hacked
the repository with admin -o...).

Rick
--
Rick Genter
Principal Engineer
Silverlink Communications



More information about the cvsnt mailing list