[OpenSIPS-Devel] pres_rules: Does presence_xml uses "sphere" element?

[Iñaki Baz Castillo]

2009/7/16 Anca Vamanu <anca at opensips.org>:
> The RFC really does not suggest the processing you recommend.
> -----
>
>  Care must be taken in using <sphere> as a condition for determining
>  the subscription handling.  Since the value of <sphere> changes
>  dynamically, a state change can cause a subscription to be suddenly
>  terminated.  The watcher has no way to know, aside from polling, when
>  their subscription would be reinstated as the value of <sphere>
>  changes.  For this reason, <sphere> is primarily useful for matching
>  on rules that define transformations.
>
> ---
> Rules that define transformations refer to the rules that define the changes
> to be applied to the body before sending it in the Notify body. And from
> this paragraph I understand exactly that the check should not be that
> strict.

Well, I understand from the above text that playing with "sphere" is
dangerous so it's recommended just for some specific cases. However,
if sphere is present it should be matched as a stric condition (IMHO).



> But since you say that it does not work with eyebeam and there are not so
> many client implementing XCAP auth and maybe none other using sphere
> specifications, I will change the check the way you suggested.

I think it's the correct behaviour. It allows "private" usage of XCAP
(as EyeBeam does) without breaking interoperability. :)

Thanks a lot.



-- 
Iñaki Baz Castillo
<ibc at aliax.net>