On Mon, 2005-06-06 at 09:25 -0400, John Peacock wrote:
> Erik Huelsmann wrote:
> > It's not about versions: it's about capabilities. You can't know in
> > advance all versions of all clients-to-be-developed and the versions
> > in which they will add certain capabilities.
> >
> > In other words, I'm against adding the version number to the stream.
>
> I wasn't advocating version numbers either, merely that there could be a
> good reason to add some client->server communication for the server to
> make better decisions about serving that particular client. I agree
> that a capabilities listing (much like SMTP et al use) is a much better
> way to handle it. However, the rub comes in defining those capabilities
> in a consistent yet extensible way.
HTTP provides a good way to do capability announcement through its
Accept* headers. Applying this strategy to ra_dav is straightforward.
Application to ra_svn might vary, since it uses a statefull protocol.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Jun 6 20:34:58 2005