On Thu, 25 Jan 2007, Giovanni Bajo wrote:
> On 25/01/2007 0.00, Daniel Rall wrote:
>
> >Implementing this behavior as an external program is somewhat
> >difficult, and leaves us little better off in terms of interface
> >between Subversion's libraries and the conflict resolution mechanism
> >than we are today with the external merge tool.
>
> I don't see exactly why. In svnmerge.py (and SVN 1.5) I think it would be
> reasonable to just stop at the first conflict and ask the user to retry. We
> just want to make sure that the property is correctly "flushed" before we
> bail out. The user can then fix the conflict in whatever form he pleases,
> and rerun the *same* command line, which will just continue the merge.
I was referring to actually implementing a conflict resolution
callback in the command-line client, as opposed to Peter's
recommendation to *not* implement such a thing. The amount of
communication from the callback back to Subversion's core libraries
may be clumsy to handle via process return codes (which is what we do
for external merge tools).
- application/pgp-signature attachment: stored
Received on Thu Jan 25 01:30:06 2007