Re: [PATCH] Cancellation editor discards abort_edit

"David Glasser" <glasser_at_davidglasser.net> writes:

> On Wed, Apr 2, 2008 at 6:53 PM, Karl Fogel <kfogel_at_red-bean.com> wrote:
> > Hmm. I thought we deliberately designed all our structures now to take
> > additions at the end, so that old code referring into them would never
> > miss. In fact, I thought this was one reason we insist that
> > 'svn_delta_editor_t' be allocated by 'svn_delta_default_editor() instead
> > of by apr_p[c]alloc() ? (We seem to have neglected to document that,
> > but I remember it being The Plan.)
> >
> > This strategy is obvious with opaque structures, of course, but I
> > thought we were doing it for some non-opaque ones too...
>
> Yeah, I believe that you *can* add new callbacks; you just can't add
> new callbacks that must ~always have a non-default implementation.

I'm not sure how well that's going to work out in practice,
especially with something like the editor interface, but OK. Is
this what you had in mind?

Index: svn_delta.h
===================================================================
--- svn_delta.h (revision 30158)
+++ svn_delta.h (working copy)
@@ -922,6 +922,8 @@
   svn_error_t *(*abort_edit)(void *edit_baton,
                              apr_pool_t *pool);
 
+ /* Be sure to update svn_delta_get_cancellation_editor and
+ * svn_delta_default_editor if you add a new callback here. */
 } svn_delta_editor_t;

-- 
Eric Gillespie <*> epg_at_pretzelnet.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org