On Mon, Jun 25, 2012 at 09:21:04AM -0400, C. Michael Pilato wrote:
> My days have begun the same for many months, now:
>
> $ cd ~/projects
> $ svn up *
>
> But it seems that recently a regression has been introduced into Subversion
> that causes this action to ultimately result in an error:
>
> $ svn up *
> Updating 'wc1':
> At revision 2.
> Updating 'wc2':
> At revision 5.
> subversion/svn/update-cmd.c:163: (apr_err=155007)
> subversion/libsvn_client/update.c:662: (apr_err=155007)
> subversion/libsvn_wc/conflicts.c:566: (apr_err=155007)
> subversion/libsvn_wc/wc_db.c:7417: (apr_err=155007)
> subversion/libsvn_wc/wc_db_wcroot.c:556: (apr_err=155007)
> svn: E155007: '/home/cmpilato/projects' is not a working copy
> $
>
> Since the error always happens at the end of the process -- after
> successfully updating all the children -- and because conflicts.c is in the
> stack, I'm guessing it has something to do with the recent change to run the
> interactive conflict handling post-update.
>
> I might get a chance to look at this later, but wanted to pop off email (if
> only as a reminder to myself to do so).
It's a silly bug I introduced where the new codes tries to resolve conflicts
in an unversioned directory, the parent of all update targets in your case.
Should be fixed as of r1353532.
Received on 2012-06-25 15:43:37 CEST