[Council] JEP-0077?

Matthew A. Miller linuxwolf at outer-planes.net
Tue Aug 17 11:29:16 CDT 2004


All of the explanations look sound to me.

Can't wait to read them from http://www.jabber.org/jeps/jep-0077.html (-;


-  LW

Peter Saint-Andre wrote:
> On Tue, 17 Aug 2004 06:30:38 -0600, Matthew A. Miller
> <linuxwolf at outer-planes.net> wrote:
> 
> 
>>*  Section 5 seems to imply that x:data is used instead of the built-in
>>iq:register fields.  If this is the intent, it needs to be explicitly
>>stated.  If this is not the intent, then I would like that section to
>>contain language explicitly stating what takes presence (x:data form or
>>iq:register fields).
> 
> 
> A service could include both (iq:register for legacy clients that do
> not support x:data) or x:data only (e.g., if "extended",
> non-iq:register fields are required in the x:data form). If a client
> understands x:data, then the x:data form takes precedence (i.e., the
> receiving application must not mix and match iq:register fields with
> x:data fields).
> 
> 
>>*  Section 6 implies that x:oob is used instead of the iq:register
>>fields.  I would like this to be explicitly stated.
> 
> 
> First, this section is non-normative. However, I think that in most
> cases a given deployment would redirect potential registrants to a
> website because it already has a provisioning system running via the
> web and does not want users to register in-band via Jabber.
> 
> In fact this raises a further question of precedence. There are many
> possible combinations:
> 
> 1. iq:register only -- process as defined in the JEP.
> 
> 2. iq:register + x:data -- process only the x:data if you understand
> it, otherwise process the iq:register fields (which are probably
> included for "legacy" clients).
> 
> 3. iq:register + x:oob -- not recommended; if received, process the
> iq:register if you understand it, otherwise redirect to the URL
> 
> 4. x:data + x:oob -- process the x:data if you understand it,
> otherwise redirect to the URL
> 
> 5. iq:register + x:data + x:oob -- not recommended; if received,
> process only the x:data if you understand it; otherwise process the
> iq:register if you understand it; otherwise process the x:oob
> 
> So x:data is always first in the processing order if you understand
> it, iq:register is always second, and x:oob is always last.
> 
> If you don't understand any of these, you probably should not have
> sent the IQ get in the first place. :-)
> 
> If this makes sense, I will adjust the text accordingly.
> 
> Peter
> _______________________________________________
> Council mailing list
> Council at jabber.org
> https://jabberstudio.org/mailman/listinfo/council


More information about the Council mailing list