> > Wouldn't going away from an event interface kill a possible direct link > > between the remote and X? Yes, it would. > > The way I see it, LIRC is an additional layer that may be one too many > > in most cases. From my point of view, it is a relative pain I could do > > without. But I may have tunnel vision by lack of knowledge. > > I agree with you. I'm more looking for a solution with existing things. > LIRC is not in kernel. I don't think we should do something specific, new. > If there is nothing which can be done with the event system I think we > should either extend it or just drop this idea. IMHO, the event interface does not match well with the reduced key set on a remote control. The keys are mapped in the driver which makes it difficult to customize. I'm not talking about the current problem with the mappings being hardcoded -- that could probably be solved without too much work. The problem I see with the mapping taking place in the driver, is that the interpretation of the key presses and releases should be application-specific. If I'm in MythTV, I want one interpretation. In Evolution, the same keypress should be interpreted differently. Firefox has a third set of mappings. Lircd solves this problem for me, while the event interface creates one static mapping. // J _______________________________________________ linux-dvb mailing list linux-dvb@xxxxxxxxxxx http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
