[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [xmpp] 3921bis: delivery of messages addressed to bare JIDs



On Fri, Jan 29, 2010 at 5:39 PM, Jonathan Schleifer <js-xmppwg at webkeks.org> wrote:
Am 29.01.2010 um 00:18 schrieb Peter Saint-Andre:


Currently some XMPP services broadcast messages addressed to a bare JID
instead of delivering those messages only to the highest-priority
resource(s). The highest-priority rule is a legacy of the original
jabberd 1.x codebase and I don't see a strong reason to forbid other
message delivery styles. Therefore I propose a change to Section 8.3.1.1
of 3921bis.

The relevant text in -03 reads:

 For a message stanza of type "chat" or "normal", the server SHOULD
 deliver the stanza to the highest-priority available resource.

(Followed by text about delivery ties.)

I propose the following text for -04:

 For a message stanza of type "chat" or "normal", the server MUST
 either (a) deliver the stanza to all available resources or (b)
 deliver the stanza to the highest-priority available resource.

(Again followed by text about delivery ties in case (b).)

Thoughts?

The proposed text seems to suggest delivering to all available resources is recommended (it's the first option). Is that intentional?

I would have thought something like "SHOULD deliver the stanza to the highest-priority available resource, but MAY deliver to all available resources" would have done the trick.
 

Hm, this sounds to me like killing resources all together. Why not let the user decide? For example, if the user gives all resources the same priority, all resources get the message. If the user decides to use different priorities, only the one with the higehst gets it (if 2 are sharing the highest prio, both should get it).

 
This doesn't kill resources, this just explicitly allows servers to be configured such that priority doesn't affect message delivery.

--
Waqas Hussain


Note Well: Messages sent to this mailing list are the opinions of the senders and do not imply endorsement by the IETF.