On Thu, 2008-06-12 at 10:43 +0200, Erik Huelsmann wrote:
> On 6/12/08, Karl Fogel <kfogel_at_red-bean.com> wrote:
> > Julian Foad <julianfoad_at_btopenworld.com> writes:
> > > The attached patch simplifies code that walks a WC, by always using the
> > > "walk_entries" function, and no longer special-casing the single-file
> > > (or non-recursive directory) cases. The "walk_entries" function can
> > > handle those cases itself.
> > >
> > > This patch is NOT FINISHED: two merge tests fail, presumably due to some
> > > intentional difference between the single-item case and the recursive
> > > cases. I'll investigate and fix if there are no objections.
> > >
> > > Does anyone see a flaw in this approach?
> >
> > Not only do I not see a flaw in this approach, I see a big improvement
> > to our code...
>
> I had the same reaction, but didn't understand the diff (reading it
> from work). So, I didn't send my reaction. But a very big +1 to the
> concept.
Thanks.
Looking into the test failures, this reveals inconsistencies in the way
svn, especially property commands, treats items scheduled for deletion.
The patch changes some of this behaviour towards more consistency, but
this needs discussion.
I'll have to come back to that issue in a separate discussion and leave
that part of the patch aside for now. I might be able to apply the other
parts of it.
- Julian
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-06-12 13:00:44 CEST