[xmpp] 3921bis: probe to full JID?

[Peter Saint-Andre <stpeter at stpeter.im>]

While reviewing the text on presence probdes just now in Section 4.3 of
3921bis, I noticed that we say a presence probe MUST be addressed to a
bare JID. However, in Section 8 there is no text about how to handle a
presence probe that (in contravention of the spec) is addressed to a
full JID. Therefore I think we need to do one of the following:

1. Continue to disallow probes to full JIDs but add rules for handling
such probes to Sections 8.2.1 and 8.2.2 (e.g., return a <bad-request/>
error or silently ignore the probe).

2. Allow probes to full JIDs and add rules for handling such probes to
Section 4.3.2.

I have heard arguments in favor of probing full JIDs (e.g., to help
prevent ghost occupants in chatrooms), but I think there are better ways
to solve the problems driving those arguments than to probe full JIDs.
Furthermore, currently we say that a presence subscription request MUST
be addressed to a bare JID, and that seems consistent with specifying
that probes are associated with subscriptions (not with directed
presence and the temporary presence sharing that results from directed
presence). Thus I favor Option #1 above. However, I didn't want to gloss
over this issue in case anyone has strong opinions in the matter.

As to the rules for handling probes to full JIDs if we decide to proceed
with Option #1, I suggest that the receiving server shall return a
<bad-request/> error. (We need to specify how the receiving server shall
handle subscription requests to full JIDs, and it seems reasonable to
return a <bad-request/> error in that case, too.)

Peter

-- 
Peter Saint-Andre
https://stpeter.im/

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature