(sorry for terse toppost; phone) Yeah, I was concerned about switched paths, too, but haven't figured out a solid way to represent copy sources in an update. Maybe src_relpath is always repos-based, and dst is negotiated.
But that doesn't translate well to move's semantics. On Apr 9, 2012 5:03 PM, "Bert Huijben" <b...@qqmail.nl> wrote: > > > > -----Original Message----- > > From: Hyrum K Wright [mailto:hyrum.wri...@wandisco.com] > > Sent: maandag 9 april 2012 17:26 > > To: Greg Stein > > Cc: dev@subversion.apache.org > > Subject: Re: Ev2 paths (was: svn commit: r1310925 ...) > > > > For repos->wc drives, my initial thought was to have the "relpath > > root" be the working copy root, but there are scenarios where that > > doesn't play out too well. Commit test 28 "commit from disjoint > > working copies" is one example, though we'd still have to do > > repos_relpath->local_relpath->local_abspath translation on the client. > > For future reference in this thread: > > Note that this architecture would also break on switched paths. > > All this is why the update editor starts by describing the working copy > layout to the repository before the update. > > In the update/switch/diff/merge handling the layout is handled at the > repository, but when committing libsvn_client handles the working copy > layout. > > The repository also handles mixed revision situations (which might show up > as unrelated nodes), which we can mostly ignore for layout purposes on > commit as that would just produce an out of date error. > > Bert > > > > -Hyrum > > > > > > -- > > > > uberSVN: Apache Subversion Made Easy > > http://www.uberSVN.com/ > >
