On Mon, 2013-01-07 at 14:14 +0100, Mikel Astiz wrote: > On Sun, Dec 30, 2012 at 2:21 PM, Tanu Kaskinen <tanuk at iki.fi> wrote: > > On Wed, 2012-12-19 at 13:58 +0100, Mikel Astiz wrote: > >> + d->audio_state = PA_BT_AUDIO_STATE_DISCONNECTED; > >> + > >> + if (parse_device_properties(d, &iface_i) < 0) > >> + return -1; > > > > The broken device is now left in y->devices. I think it would be better > > to only add the device to y->devices if parsing the properties succeeds, > > and otherwise free the device here. > > device_info_valid==-1 seems to serve exactly this purpose, so I'd > avoid changing this behavior in this patch and leave it for later. Ok, you don't need to change it in this patch. If you change it later, could you also change device_info_valid to be a normal boolean variable? -- Tanu
