On Wed, 2009-01-21 at 16:29 -0500, Saint, David (CAR:9D60) wrote: > I don't yet a good proposal for an architectural change, but I had also > considered the approach of sending two objects to the Subscribe Server > (fullState=TRUE and fullstate=FALSE), but the problem with this is that > when the Subscribe Server automatically responds to the SUBSCRIBE > message the version number in the NOTIFY RLMI is not incremented so the > message would likely get discarded. It looks like the RLS component is > currently responsible for incrementing the RLMI version number. > > For example: > <- NOTIFY( RLMI version=36 fullState=FALSE) -- from RLS > <- NOTIFY( RLMI version=37 fullState=FALSE) -- from RLS > -> SUBSCRIBE > <- NOTIFY( RLMI version=37 fullState=TRUE) -- from Subscribe > Server > <- NOTIFY( RLMI version=38 fullState=FALSE) -- from RLS
That's ugly. I suppose we could claim that the full and partial "version=37" *states* were the same, even if the NOTIFY bodies were different. But I don't expect many phones would handle it that way. Perhaps it is time to fix the Subscribe Server to insert the version numbers, rather than having the application do that. Because we now have the application insert the version numbers, most subscribers do not see the version start at 1 (although I've not seen any phones that care about that). Dale _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
