[Standards-JIG] Re: Re: JEP-45 requires JEP-79?

Gaston Dombiak gaston at jivesoftware.com
Wed Nov 23 21:07:02 UTC 2005


Hey Ralph,

Ok. Makes sense to me. Note that room presences can also be delivered to 
clients so clients should also ignore them. BTW, I guess you wanted to say 
JEP-160 instead of JEP-60.

Thanks,

  -- Gato

"Ralph Meijer" <jabber.org at ralphm.ik.nu> wrote in message 
news:20051123204929.GA3913 at ik.nu...
> On Wed, Nov 23, 2005 at 02:29:33PM -0600, Gaston Dombiak wrote:
>> Hey Nolan,
>>
>> Thanks for your reply.
>>
>> What about the "match-resource" condition? I think that we (the group
>> chat service) don't want the server to send a room message to another
>> user resource. I can see 2 possible scenarios where the "match-resource"
>> and "drop" action might be useful.
>>
>> The first scenario is when a user is connected from 2 resources and only
>> 1 has joined a room. And while there is some room activity the user
>> loses his connection from the client that was present in the room.
>> AFAIK, if we don't specify "match-resource" and "drop" in the messages
>> to broadcast then the server might sent the room conversation to the
>> remaining connected resource or if that resource has been disconnected
>> too then the message might be stored offline.
>
> I think this is really an edge case. Clients should ignore messages of
> type 'groupchat' from rooms they didn't join.
>
>> The other more unprobable scenario would be for clients that logged into
>> a server but never sent an available presence but they decided to send a
>> direct presence to join a room. AFAIK, clients that never sent an
>> available presence to the server should not get messages from other
>> clients or services since they are "unavailable". If this is correct
>> then the server might decide to store offline broadcasted messages of
>> the room.
>
> I don't think current server implementations store messages of type
> 'groupchat'. JEP-0060 lists this case as SHOULD NOT.
>
> -- 
> Groetjes,
>
> ralphm
> 






More information about the Standards mailing list