On 5 Jun 2015 08:44, "Florian Schmaus" <[email protected]> wrote:
>
> On 05.06.2015 09:36, Dave Cridland wrote:
> > On 5 June 2015 at 07:24, Florian Schmaus <[email protected]
> > <mailto:[email protected]>> wrote:
> >
> >     On 04.06.2015 09:39, Kevin Smith wrote:
> >     > On 3 Jun 2015, at 16:02, XMPP Extensions Editor <[email protected]
<mailto:[email protected]>> wrote:
> >     >> http://xmpp.org/extensions/inbox/nonza.html
> >
> >     > The definition here seems potentially useful. I would add a
‘generally’ to 4 so that it becomes “...they are generally used in a
more…”, so as not to be seen as prescriptive.
> >
> >     Good point, going to change it.
> >
> >     > None of the current nonzas are routed, but it doesn’t seem
impossible that one might be in the future, and I don’t see a reason to
forbid it here. Noting that they’re not expected to be routed seems useful
and sufficient, to me.
> >
> >     If you want to send something that is supposed to get routed, why
> >     wouldn't you use simply a Stanza instead? I consider it a security
> >     improvement if routing of Nonzas is explicitly forbidden.
> >
> >
> > I think the definition of a stanza is a routed top-level element, so an
> > extension that negotiated "routed Nonzas" is actually negotiating a new
> > stanza type. My reading of RFC 6120 seems to leave room for negotiating
> > new stanzas (and moreover, they needn't have the common attributes of
§8.1).
>
> I don't think so. It appears to me that Stanzas are very well defined in
> RFC 6120. See below.
>
> > However, I don't think that RFC 6120 actually defines what a stanza
> > *is*.
>
> From XEP-Nonza:
>
> Stanzas ... are specified in RFC 6120 [2] § 4.1 "Stream Fundamentals"
> and § 8. "XML Stanzas"
>

Ah, yes. Hadn't noticed the 4.1 definition. That's very much more
restrictive, and doesn't seem to leave room for new stanza types.

Moreover, it also suggests that XEP-0114 stanzas aren't actually stanzas.
Therefore under this proposal they would be unroutable nonzas.

> > 3) Some convenient term of art for first child elements of the stream -
> > ie, the collective term for both Stanzas and Nonzas.
>
> Top-level stream element?
>

We've used TLE in the past, I think.

> - Florian
>
>

Reply via email to