> -----Original Message-----
> From: news [mailto:news_at_ger.gmane.org] On Behalf Of Roman Donchenko
> Sent: maandag 12 oktober 2009 1:10
> To: dev_at_subversion.tigris.org
> Subject: Re: svn_client_status and path relativity
> Bert Huijben <rhuijben_at_sharpsvn.net> писал в своём письме Mon, 12 Oct
> 02:36:21 +0400:
> >> I'm talking about the path that's being passed to the receiver, not
> >> one that info itself uses.
> >> E.g.:
> >> > svn info -rBASE src-trunk\COMMITTERS
> >> Path: COMMITTERS <-- shouldn't this be src-trunk\COMMITTERS?
> >> ....
> >> > bin-trunk\svn info -rBASE testwc\trunk\README.txt
> >> Path: testwc\trunk\README.txt\README.txt <-- the same bug as in the
> >> bindings testsuite, caused by passing base_name to the receiver
> >> ....
> > -r BASE is also translated in the base revision of the path and then
> > send to the server with the url of the working copy path. Only if you
> > pass an unspecified revision (API) or no revision (CLI) you get the
> > wc-path version.
> BASE or HEAD is not the point here; point is that info currently
> returns a
> clearly malformed path (when the repo is queried).
Ok.. I think I found your issue.
In my steps I forgot a check that is crucial in this specific case where we allow paths and urls. Before appending the relative prefix, you must check that the path really starts with the absolute prefix.
(No time to check right now, but I think this is what you saw).
I will try to apply a fix to libsvn_client/deprecated.c later today...
Received on 2009-10-12 09:51:39 CEST