Re: svn commit: r966774 - /subversion/branches/issue-2779-dev/subversion/libsvn_client/update. c

[Stefan Sperling <st...@elego.de>]

On Thu, Jul 22, 2010 at 09:53:02PM +0300, Daniel Shahaf wrote:
> Stefan Sperling wrote on Thu, Jul 22, 2010 at 20:11:23 +0200:
> > On Thu, Jul 22, 2010 at 07:59:42PM +0200, Bert Huijben wrote:
> > > Shouldn't we relocate the complete working copy, starting by the working copy root instead of just the operation anchor?
> > 
> 
> So "cd notes; svn up commit-access-templates" would relocate my entire trunk wc?
> 
> At the least, this departs from current semantics of svn.  I won't speak for
> "many users", but I, for one, would be surprised if svn started touching the
> parent of the dir I specified as target.
> 
> If this change is made, can we please add the appropriate big red letters
> somewhere?

This is svn switch --relocate, not svn switch.

If you switch --relocate a working copy, you are pointing to a different
server. It makes sense to relocate the entire working copy.
If you want nested working copies pointing to different servers,
you'd use an external.

Switched subtrees (i.e. without --relocate) are different, because we know
for certain that they are from the same repository.

Stefan