On Apr 3, 2012, at 9:51 AM, Marcel Holtmann wrote:
>>>> In the current Connman version the "state"  of the manager API is
>>>> either "offline" or "online".
>>>> If only one device is marked as connected the global state is "online".
>>>> 
>>>> So the global online state doesn't have the same signification for the
>>>> service state, it's a little bit confusing.
>>>> 
>>>> The service state makes the difference between "online" and "ready"
>>>> and detects well the Internet availability.
>>>> 
>>>> Why don't we notifies a global state "online" when we have Internet
>>>> and "ready" when we have an IP configuration ?
>>>> 
>>>> He could be usefull for the applications who are waiting for Internet
>>>> without checking the whole list of services state.
>>> 
>>> I rather have applications use the Session API, than this global state.
>>> 
>>> And maybe we just better remove it before 1.0 release.
>> 
>> -1. While it certainly has its limitations, the simplicity of the 
>> Manager.State element has value and, at this point, I'd rather avoid having 
>> to rewrite most of our application when transitioning from 0.76 to 1.0.
> 
> fair enough actually.
> 
> So should we keep it as "online" and "offline"? So pretty simple. Or
> should we introduce an extra state for when we have a connection, but
> not yet confirmed that it is actually a valid Internet connection.

To date, that's absolutely been sufficient. I could see an argument for 
mirroring the service state of "ready" in addition to "offline" and "online"; 
however, it hasn't been an application deal-breaker to date.

Reading through session-overview.txt and session-api.txt in the top-of-tree GIT 
repository reconvinces me that the session API is a whole lot of complexity 
that our application simply doesn't need.

Thanks,

Grant

_______________________________________________
connman mailing list
[email protected]
http://lists.connman.net/listinfo/connman

Reply via email to