Hi,
On Wed, 2014-01-15 at 16:45 +0100, Daniel Wagner wrote:
> On 01/15/2014 04:11 PM, Patrik Flykt wrote:
> > On Wed, 2014-01-15 at 15:55 +0100, Daniel Wagner wrote:
> >> IIRC we discussed the benefits of having a __connman_auto_disconnect()
> >> and we agreed that it is not brilliant idea. Instead it would be
> >> better when the auto connect part handles this.
> >
> > I can't remember anymore what we discussed (f2f?), but it would be
> > better to disconnect the service when it is known that a session
> > connected it and the last session using it has closed.
>
> http://permalink.gmane.org/gmane.comp.handhelds.moblin.connman/15856
>
> My understanding was that you didn't like to have an explicit call to
> disconnect. Instead it should be part of auto connect which makes really
> sense. In this series, auto connect tries first to connect something
> (and my plan for later would also like switch from one service to
> another if needed) and in the second step to disconnect unused services.
In the above link I didn't like an explicit function for disconnecting
that is called at some undefined time. I have been trying to say that
disconnects should happen immediately after the last session is not
interested in the service anymore and that the service was connected
only because a session caused it to be connected in the first place.
I still don't want a *_disconnect(void) function that can be called
anytime "later" and leaving the session requested service idling for an
extended period of time.
Cheers,
Patrik
_______________________________________________
connman mailing list
[email protected]
https://lists.connman.net/mailman/listinfo/connman
