On Thu, Apr 24, 2008 at 3:21 PM, Branko Èibej <brane_at_xbc.nu> wrote:
>
> David Glasser wrote:
>
> > On Thu, Apr 24, 2008 at 12:17 PM, <brane_at_tigris.org> wrote:
> >
> >
> > > Modified: trunk/notes/wc-ng-design
> > > URL:
> http://svn.collab.net/viewvc/svn/trunk/notes/wc-ng-design?pathrev=30774&r1=30773&r2=30774
> > >
> ==============================================================================
> > > --- trunk/notes/wc-ng-design Thu Apr 24 12:07:54 2008
> (r30773)
> > > +++ trunk/notes/wc-ng-design Thu Apr 24 12:17:08 2008
> (r30774)
> > > @@ -387,6 +387,8 @@ in this document, or elsewhere in the su
> > > * Modularization
> > > * Agree on a UI to pull in other parts of the same repository
> > > (NOT svn:externals) [relates to issue #1167]
> > > +#####XBC I submit this is a server-side feature that the client
> > > + (i.e. the WC library) should not know about.
> > >
> > >
> >
> > Why?
> >
> > In my mind, it's exactly the opposite. (Hell, SVK has a project-spec
> > feature *today* which essentially does this entirely client-side.)
> >
> > This feature is about "letting the user dictate what their working
> > copy looks like" (where one possible choice is "base it on a shared
> > and versioned set of settings"). That's about as client-side as you
> > can get.
> >
> >
>
>
> Because then every client that wants to be comparable to others has to
> implement that. Even those that don't or can't use the working copy library;
> such as, for example, the dav-autocommit "client". Doing this on the server
> side (with a kind of hardlink equivalent) would be completely transparent to
> all clients.
>
> Sure, that would change the server-side revision from a tree to a DAG; but
> no-one except the server would notice that.
But then it would keep us in the broken state we are today where
cross-repo "externals" are second-class citizens.
--dave
--
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/
Received on 2008-04-25 01:02:32 CEST