- James
On 3/14/06, Michael Bernstein <[EMAIL PROTECTED]> wrote:
On Tue, 2006-03-14 at 22:14 -0800, James M Snell wrote:
>
> Michael Bernstein wrote:
> > On Tue, 2006-03-14 at 21:12 -0800, James M Snell wrote:
> >> If the server responds indicating 204 Created, the entry should appear
> >> within the feed.
> >
> > It was my understanding that generally speaking, the contents of a feed
> > (like introspection docs) can vary with various factors including user
> > authentication, in which case a conformant implementation still might
> > not immediately show a successfully POSTed entry to the same user that
> > created it, at least not until some other condition was met (such as a
> > second user's approval).
>
> You know what, you're right, I'm wrong. Consider POST only feeds. I
> may have the authority to POST and entry to the feed, but not be able to
> turn around and read it. That doesn't change the 202 Accepted
> discussion above, however.
Even if some other user *can* see the Entry in that same collection?
I'm not talking about a POST-only feed (although that's an interesting
use-case), but a feed where only *some* entries are visible to the
POSTing user, when (for whatever reason) the newly posted Entry (though
it *does* exist in the collection) is not (yet?) one of them.
In other words, though I think the POST should succeed with a 204 and
return a location header, a GET (by that same user) to the returned
Entry location should result in a '403 Forbidden', and *not* a '404 Not
Found'.
Since the user cannot (yet?) GET the newly created Entry (though it *is*
demonstrably there), I think it would be valid *not* to list it (for
that user) when a GET is done to the collection.
It seems to me that this behavior conforms with the spec, and
constitutes a credible and useful use-case.
- Michael
--
- James Snell
http://www.snellspace.com
[EMAIL PROTECTED]
