[Standards] the meaning of "MUST be empty"

Kevin Smith kevin.smith at isode.com
Tue Jun 19 12:26:41 UTC 2018


On 19 Jun 2018, at 13:13, Guus der Kinderen <guus.der.kinderen at gmail.com> wrote:
> 
> 
> 
> On Tue, 19 Jun 2018 at 13:47, Kevin Smith <kevin.smith at isode.com <mailto:kevin.smith at isode.com>> wrote:
> On 19 Jun 2018, at 12:09, Bartłomiej Górny <bartlomiej.gorny at erlang-solutions.com <mailto:bartlomiej.gorny at erlang-solutions.com>> wrote:
> > 
> > Hi
> > 
> > If a XEP states that an attribute "MUST be empty", does it mean that it:
> > a) must be present and have a value ""
> > b) must not be there
> > c) can be either of the two
> > 
> > The question arose because of XEP-0313, which in point 5.1.2 says:
> > 
> >    "When sending out the archives to a requesting client,
> >    the 'to' of the forwarded stanza MUST be empty"
> > 
> > and then gives an example where forwarded stanzas have no 'to' attribute. We just hit a situation where there are conflicting implementations, and we want to sort it out The Right Way, hence the question.
> 
> Sounds like we messed up the text, sorry. The right thing is to not include a to, rather than including a to=“” (which is illegal).
> 
> 
> Is it (illegal)? It's valid in XML 1.0 and XML 1.1, if my Google skills are not failing me. Differentiating between not having a to attribute, and having a to attribute with an empty value seems needlessly complicated to me.

to= holds a JID, and a zero-length JID isn’t legal.

/K

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.jabber.org/pipermail/standards/attachments/20180619/71f822ff/attachment.html>


More information about the Standards mailing list