> On Wed, Jan 30, 2002 at 07:31:32PM +0100, Sander Striker wrote:
> > [...]
> > > But to answer your question: once we fix this use-case, yes, your
> > > algorithm is the right way to back out changes to a file. You would
> > > view the log for the *one* file, determine that you want to backdate
> > > the file to revision N, and then do something like:
> > >
> > > svn diff -r N -r HEAD <filename> | patch
> > > svn commit
> > >
> > > Or you could use 'svn merge' command, which will be something
> > > extremely similar.
> >
> > I can't remember what we defined rollback as, but if it would
> > look something like this:
> >
> > svn rollback [<path>] [--recursive] <revision>
> >
> > that would be very cool. Anyone care to refresh my memory?
>
> i believe the current plan is something like that, but there are also
> people who disagree, and think that rollback should physically remove
> changes from the repository itself. i'm not one of those people, but
> they do exist.
>
> -garrett
Actually, that would be quite nice. Saves you from having
to muck with a working copy.
svn rollback [(<url>|<path>)] [--recursive] <revision>
Sander
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:37:01 2006