On Apr 23, 2012 1:36 PM, "Hyrum K Wright" <hyrum.wright_at_wandisco.com> wrote:
> On Mon, Apr 23, 2012 at 12:18 PM, Greg Stein <gstein_at_gmail.com> wrote:
> > On Apr 23, 2012 1:12 PM, "Hyrum K Wright" <hyrum.wright_at_wandisco.com>
> >> On Mon, Apr 23, 2012 at 12:08 PM, Greg Stein <gstein_at_gmail.com> wrote:
> >> >
> >> > On Apr 23, 2012 8:00 AM, "Hyrum K Wright" <hyrum.wright_at_wandisco.com>
> >> > wrote:
> >> >>...
> >> >
> >> >
> >> >> I added an Ev2 capability to svn_ra.h on the ev2-export branch,
> >> >> we could use to check the ability of the client and/or server to
> >> >> receiver Ev2-encoded deltas. Feel free to poach the one-line patch.
> >> >
> >> > I'm not sure it is needed. How would the client use it?
> >> To determine if the server is Ev2 capable? We've got to have *some*
> >> method of doing so, and the existing capabilities mechanism was built
> >> just for that purpose.
> > But is that an RA constant, or something under the covers? ie. does that
> > need to be in svn_ra.h for client consumption, or can RA determine that
> > privately and compensate?
> The string is globally defined, but each server (and client?)
> advertises the capabilities it supports as part of the initial
> connection set up. These are then cached by the client (and server?)
> for other libraries to consume. In practice, these means that
> libsvn_ra can fall back to alternative implementations of
> functionality if talking to older servers.
> > For example, ra_dav can do everything but the symlink operations and
> > rotate(). Those will need to query the server to determine whether to
> > new/old marshalling. But to the svn_ra user, it is always available.
> > For ra_local, it is always available at both internal/client levels.
> Both correct, and the capabilities mechanism is the way for the ra
> layers to determine if the other side supports the new marshalling
Alright. It sounds like svn_ra.h has two purposes: stuff for libsvn_client,
and stuff for all RA layers to use internally. Ugh.
I'd rather see internal stuff moved to a private header, since it is not
part of the public API...
IOW, if/when that capability hits trunk, can we put it into svn_ra_private?
(and do so, going forward)
Received on 2012-04-23 19:53:15 CEST