On 12/13/2011 01:42 AM, Peter Saint-Andre wrote: > On 12/12/11 1:43 AM, Sergey Dobrov wrote: >> On 12/12/2011 02:37 AM, Kevin Smith wrote: >>> On Fri, Dec 9, 2011 at 12:15 PM, Sergey Dobrov <[email protected]> wrote: >>> <snip/> >>>> 9. How A's server can know that it don't have to send statuses to B >>>> anymore? >>> >>> It doesn't - it continues to send A's status to B whenever A changes >>> it. (RFC6121, 4.4.2) >>> >>> /K >>> >> Even if A will logged out and then again logged in? I mean, once sent >> probe presence will mean that A's presence will be sent forever till >> subscription is active? > > When A logs out, A's server sends unavailable presence and then there is > no need for it to send presence to B until A logs in again.
This is an obvious thing since we just have no any new statuses to send until A logs in again but does it means that A server will send any changes forever to B since the moment it received probe? Maybe it's justified to introduce some presence in reverse to probe presence that will stop the process of presence translation? The problem I have now is to understand the process to understand how can be done the same process for the PEP. And I testing on two ejabberds (2.1.7/2.1.9) that installed in my local network for testing purposes and I see that if A logs out and then logs in when B is permanently online then B sees only first A's status and never sees if changes status even if A logs out again, B still see like A is online! Have I to post a but to ejabberd or this is some kind of the protocol bug? > > I think we might have a failure of communication here. Could you explain > your question in a bit more detail? > > Peter > -- With best regards, Sergey Dobrov, XMPP Developer and JRuDevels.org founder.
