On Thu, Feb 14, 2008 at 9:07 PM, Karl Fogel <kfogel_at_red-bean.com> wrote:
> "David Glasser" <glasser_at_davidglasser.net> writes:
> > Out of curiosity, why aren't you using svn_repos_has_capability in
> > mod_dav_svn (like the other two servers)?
>
> Hmm, no good reason, actually. Six of one, half dozen of the other?
>
>
> > That would also let you sidestep this issue for now...
>
> Well, no: whether we detect the incapability by getting an error from
> svn_whatever_get_mergeinfo() vs by asking svn_repos_has_capability(),
> is unrelated to how we report that incapability to the client.
>
> As you pointed out, we *could* have a non-error-ish way in the
> protocol of saying "mergeinfo isn't supported". However, since we
> explicitly have HTTP_NOT_IMPLEMENTED around for this sort of thing
> (not to mention SVN_ERR_UNSUPPORTED_FEATURE), doing an error
> transmission back to the client seemed like the cleanest fit to me.
>
> While I could be persuaded that the in-protocol answer would be
> preferable, the persuasiveness of any given argument on that topic
> would not be changed by how mod_dav_svn determined the presence or
> absence of mergeinfo support... Or so I think, but tell me if I'm
> missing something.
Sorry, I just realized I was quite confused. I didn't realize that
serf and neon were still implementing the capability via the model of
"try and see if it works" rather than by having mod_dav_svn say the
right answer in get_vsn_options in the first place. Somehow this led
to me making a completely unrelated point about a completely different
layer.
On a sort of related note, should SVN_DAV_NS_DAV_SVN_MERGEINFO be
explicitly documented as being about the server and not the
repository? And maybe a comment added to its line in get_vsn_options?
--dave
--
David Glasser | glasser@davidglasser.net | http://www.davidglasser.net/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-02-15 06:20:37 CET