On Fri, Feb 29, 2008 at 2:27 PM, Justin Erenkrantz
<justin_at_erenkrantz.com> wrote:
> On Fri, Feb 29, 2008 at 2:22 PM, David Glasser <glasser_at_davidglasser.net> wrote:
> > In practice I think the only problem would be that the base props for
> > the roots of branches would not match the repo (one would have
> > svn:mergeinfo, and the other would have svnmerge-integrated).
> > Additionally, there'd be extra random svn:mergeinfo scattered in our
> > repositories. Perhaps somebody could write a script to do the same
> > transformation to working copies.
>
> That'd be far nicer than having to recheck out every WC. But, really
> what penalty is there if these props are out of sync? Won't they get
> fixed on the next update? (We treat properties as a single atomic
> value, don't we?) -- justin
That was my first thought. However, there are a few issues.
First, they may not get fixed "on the next update", since some of
these changes may be the deletion of svn:mergeinfo on random paths
that didn't need to have them; no dir-delta is going to decide to send
"delete this svn:mergeinfo that never existed in my version of the
world".
Second, I believe there is special-case code in libsvn_wc that in fact
*doesn't* treat svn:mergeinfo as an atomic value, but which does some
sort of merge.
--dave
--
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-03-01 00:11:41 CET