Hi; Pls see inline Regards Diego B
Brett Tate wrote: >> Recently I met a strange server, which is responding to the >> REGISTER requests (adding or refreshing a binding) with the >> 200 OK response *without* any Contact header fields. >> > > Since it is considered an abnormal situation, the device can act as it > wishes as long as it doesn't continually retry without some reasonable > time delay. > > Why do you think this is an abnormal situation? A response without contact can have normal meanings ( i.e. no bindings for the Registered AOR ) > There is likely something related to the Contact which makes it unusable > (scheme, transport, malformed, decoding issue, etcetera) or too low in > priority (q too low, too many contacts, etcetera). > > For the above situations, the REGISTRAR may response with an apropiate error code, and not just OK with an empty contact list. > >> I understand that such responses are incorrect, but I must >> choose the best way how my softphone should handle them. >> Two opposed variants are: >> 1) interpret 200 w/o contact as signal from the server that >> it did not registered my contact (i.e. handle it the same way >> as 403 response); >> 2) interpret 200 w/o contact as successfull registration of >> the contact which the softphone tried to register. >> Which of these variants is more correct and user-friendly? >> > > > Both of your variants are acceptable. It is hard to say which is best > since it really depends upon why the Contact is being lost and if the > phone is really still usable. If some of the Contacts are working, > option 2 might be better. If none of the phones contacts are working, > option 1 might be better. > > I think the UA may consider itself not registered and retry with an appropriate delay. Option 2 is not acceptable here, since the registrar did not list *any* contact registered for the AOR. > _______________________________________________ > Sip-implementors mailing list > [email protected] > https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors > _______________________________________________ Sip-implementors mailing list [email protected] https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
