On Fri, 2013-03-15 at 15:43 +0100, David Henningsson wrote: > On 03/14/2013 04:17 PM, Tanu Kaskinen wrote: > > On Thu, 2013-03-14 at 15:02 +0100, David Henningsson wrote: > >> On 02/20/2013 07:24 PM, Tanu Kaskinen wrote: > >>> The flag will be used in the upcoming pa_device_port_update_proplist() > >>> function. > >> > >> If the flag's function is to protect against an initial change > >> notification, does it really work? init_eld_ctls is called after > >> pa_card_new in module-alsa-card.c. > > > > Good observation. It looks like we need to separate pa_card_put() from > > pa_card_new(), because init_eld_ctls() needs the card object, so moving > > it before the pa_card_new() call. > > > >> If we need a flag at all, I'd prefer to keep it in the card struct > >> rather than the port. That seems cleaner to me. > > > > OK, I don't necessarily agree with this, but there aren't any big issues > > with reusing the card state as the port state. > > Since it's the card we actually send a notification for, following the > card's state seems more logical to me. That's a native protocol specific quirk. The D-Bus protocol would send the notification for the port, if it was implemented (the D-Bus protocol doesn't currently send any notification). I would prefer to do the same in the native protocol too, if changing that wouldn't have compatibility issues. -- Tanu
