Hi

(1) I tested more. It is due to the difference between bria 3 and jitsi while handling the adding/removing contact:

if  contact is added/removed:

for jitsi, it send out:

PUT /xcap-root/resource-lists/users/sip:101@192.168.122.32/index
PUT /xcap-root/pres-rules/users/sip:101@192.168.122.32/presrules


for bria 3, it only send out

    PUT /xcap-root/resource-lists/users/101@192.168.122.32/contacts-resource-list.xml HTTP/1.0.
    PUT /xcap-root/resource-lists/users/101@192.168.122.32/contacts-resource-list.xml

    it does NOT send out the pres-rules.

    Is it the good behavior ( from RFC point of view) ?
    should I call the pres_update_watchers even though its is for resource-lists?

(2)  BTW,   what is the exact purpose of watcher table?

    e.g: on 101, if delete 103, watcher table become:

        id    presentity_uri                  watcher_username    watcher_domain    event    status    reason    inserted_time
        1    sip:103@192.168.122.32    101    192.168.122.32    presence    1                            1340096051
        2    sip:101@192.168.122.32    103    192.168.122.32    presence    3 terminated         1340096181

    it is item 2 become terminated instead of item 1. If watcher table is for subscription, then seems item 1 should be terminated?
     Look at the code, it seems  somehow serve as  mixed role of subscription/auth/xcap purpose, not for subscription state?

    and is the active_watcher  mainly for the state machine of    http://tools.ietf.org/html/rfc3857#section-4.7.1?

thanks.


min

 

On 06/19/2012 10:41 AM, Daniel-Constantin Mierla wrote:

Hello,

On 6/18/12 2:11 PM, Andreas Granig wrote:

Hi,

On 06/15/2012 05:25 PM, Min Wang wrote:

|  1 | sip:103@192.168.122.32 | 101              | 192.168.122.32 |
presence |      2 |        |    1339772803 |

    then I deleted the 103 from the contact list, the watcher table still
shows the same.

For local storage, I'd expect an unsubscribe (subscribe with Expires=0)
from 101 to 103. The obvious work-flow would be to set it to
"terminated" in watchers table, and in case of a subsequent re-subscribe
it should be changed back to "pending" state, although the state-machine
doesn't indicate a transition from "terminated" back to "pending". Isn't
this the case?

For xcap storage, there are other ways to block/remove a contact on/from
the list. As Iñaki pointed out in
http://lists.opensips.org/pipermail/devel/2009-August/003868.html, the
xcap server needs to notify the sip server about the change, which in
turn will notify the other party (103) that it's no longer allowed to
see 101's state. If the xcap_server module of kamailio is used, there is
the following code snippet in some examples floating around on the internet:


switch($rm) {
   case "PUT":
   xcaps_put("$var(uri)", "$hu", "$rb");
   if($xcapuri(u=>auid)=~"pres-rules") {
     pres_update_watchers("$var(uri)", "presence");
     pres_refresh_watchers("$var(uri)", "presence", 1);
   }

So, shouldn't this update the watchers accordingly? Anyways, also in
this case the watcher state should change to "terminated", and in case
of a re-subscribe it should go back to pending if xcap rules are
allowing this.

Maybe someone with good xcap_server/presence insights could elaborate on
that?

to summarize, you say that when the contact is removed from the buddy list, the watcher table is not updated to state terminated by pres_update_watchers(...)?

Cheers,
Daniel

-- 
Daniel-Constantin Mierla - http://www.asipto.com
http://twitter.com/#!/miconda - http://www.linkedin.com/in/miconda
Kamailio Advanced Training, Seattle, USA, Sep 23-26, 2012 - http://asipto.com/u/katu
Kamailio Practical Workshop, Netherlands, Sep 10-12, 2012 - http://asipto.com/u/kpw