RE: svn commit: r1083805 - /subversion/trunk/subversion/libsvn_wc/update_editor.c

> -----Original Message-----
> From: Greg Stein [mailto:gstein_at_gmail.com]
> Sent: maandag 21 maart 2011 16:19
> To: dev_at_subversion.apache.org
> Subject: Re: svn commit: r1083805 -
> /subversion/trunk/subversion/libsvn_wc/update_editor.c
>
> On Mon, Mar 21, 2011 at 10:36, <rhuijben_at_apache.org> wrote:
> > Author: rhuijben
> > Date: Mon Mar 21 14:36:21 2011
> > New Revision: 1083805
> >
> > URL: http://svn.apache.org/viewvc?rev=1083805&view=rev
> > Log:
> > Following up on storing tree conflicts on the node itself, stop looking
> > for conflicts on nodes that don't exist when calling
> svn_wc__db_read_info().
>
> Hmm? Couldn't you have a delete/delete conflict? The node wouldn't
> exist, but a conflict would.

For this case svn_wc__db_read_info() just returns success and conflicted =
true, since a few months ago.

> How does this case not apply here? And assuming that is the case, then
> should additional comments be added to update_editor.c to explain the
> situation? (to avoid future questions like mine)

I added a comment in followup commit r1083843; would that be enough?

        Bert
Received on 2011-03-21 17:12:59 CET