On Tue, Jan 15, 2013 at 01:11:26PM +0100, David Henningsson wrote: > On 01/15/2013 12:51 PM, Damir Jeli? wrote: > >On Tue, Jan 15, 2013 at 09:42:11AM +0100, David Henningsson wrote: > >>On 01/14/2013 01:58 PM, Damir Jeli? wrote: > >>>On Fri, Jan 11, 2013 at 04:14:53PM +0100, David Henningsson wrote: > >>>>One comment here, btw: > >>>> > >>>>If we tell clients that the port volume changed, there should also > >>>>be a way for the clients to tell the port volume. I e, you need to > >>>>add the port volume in the PROTOCOL, introspect.h, and maybe a > >>>>client function to set it too? > >>> > >>>I wasn't sure if we want to confuse the user further with volume setting > >>>controls, but sure I can add that too that shouldn't be too hard anyways. > >> > >>You can also act incrementally; so that you don't send the card > >>change event in this patch (but possibly in a later patch). > >> > >>I mean, it's probably confusing for a client to get a card change > >>event, then go polling the card just to find that nothing (that it > >>can read) has changed. > >> > > > >The card change event is already needed for module-card-restore. I could > >move the event sending into the card-restore patch but I don't really > >see a need for this. > > Ok...this is not your fault, but maybe module-card-restore needs to > change from using the event notification to using hooks instead? > > From the problems we have in module-dbus-protocol, I like hooks > better, as they more precisely control when things happen. It is > easier to do things in the correct order. > There was some talk about this during the summer, but I never got around to actually fixing it. I'll add this to my TODO.