Re: svn_client_diff_peg4() returns inverse diff

On 7/30/07, Malcolm Rowe <malcolm-svn-devfarside.org.uk> wrote:
> On Sat, Jul 28, 2007 at 10:00:20AM -0700, Kyle McKay
wrote:
> > >> Stefan, can you provide an example with
the command-line using specific
> > >> WORKING revisions?  I tried the following
(which exercises
> > >> svn_client_diff_peg4), but was unable to
reproduce -- switching the START
> > >> and END revisions did reverse the diff
output.
> > >> svn di -r25820:25862
subversion/libsvn_client25862
> > >
> > > As I already mentioned, it's not possible to
give an example with the
> > > command line: it doesn't understand WC or
WORKING as revision arguments.
> > > The problem only occurs if one of the
revisions to diff is 'working copy',
> > > not a revision number.
> > >
> > > Stefan
> >
> >  FYI, issue 2683 <http://subversion.tigris.org/issues/show_bug.cgi?id=
2683>
> >  contains a patch to support WORKING revision
arguments.  It currently has a
> >  target milestone of "---".
>
> ... and one of the reasons it hasn't been committed is
because there are
> known bugs with some of the commands' handling of
rWORKING... for
> example, diff.
>
> However, we can't easily test and fix these bugs
without first having
> support in the command-line client - and meanwhile,
it's still broken
> for other users of the API.  Perhaps we should just
commit the patch
> to add rWORKING anyway, and try to shake out the bugs
afterwards?
>
> I might do that, if there aren't any objections.

Committing it will enable us to add tests too. +1.

It would help if we were to try and identify some of these
problems
right after you commit the change though.

bye,

Erik.

------------------------------------------------------------
---------
To unsubscribe, e-mail: dev-unsubscribesubversion.tigris.org
For additional commands, e-mail: dev-helpsubversion.tigris.org