Re: [webfinger] Server Response language

In a webfinger query, the HTTP resource being requested is the full
"/.well-known/webfinger?resource=acct:user@domain.com".  If the host
doesn't have any metadata for "acct:user@domain.com", what else would they
return, aside from a 404?  I agree with Tim that simply referencing HTTP
should be sufficient here.

On Mon, Jul 8, 2013 at 1:41 PM, Stephane Bortzmeyer <bortzmeyer@nic.fr>wrote:

> On Thu, Dec 20, 2012 at 01:02:59PM -0500,
>  Paul E. Jones <paulej@packetizer.com> wrote
>  a message of 383 lines which said:
>
> > But there server did find something.  It found the "webfinger"
> > resource.
>

But that wasn't the HTTP resource that was requested (unless of course you
*actually* requested just "/.well-known/webfinger").  Section 3.4 of
rfc3986 makes it clear that the query string is part of the resource
identifier (as opposed to rfc2396 which left it somewhat ambiguous)

> The software that responds to the query has to then
> > decided what it returns.  It might be logical to some, but I'd argue
> > we need to state this to avoid confusion.
>
> It seems that this thread died without a conclusion? I see nothing in
> -15 about the server's return code (all examples are success, with a
> 200 code). Tim Bray claimed that it should simply follow the standard
> rules of HTTP but, as far as I know, there are no standard rules when
> /.well-known/webfinger is found *but* the resource is not. 200? 401?
> 404?
>
> GET /.well-known/webfinger?resource=acct%3ADOESNOTEXIST%40example.comHTTP/1.1
>
> _______________________________________________
> webfinger mailing list
> webfinger@ietf.org
> https://www.ietf.org/mailman/listinfo/webfinger
>