Hi, On Mon, 2009-01-12 at 21:08 +0200, Ville Syrjälä wrote: > On Mon, Jan 12, 2009 at 07:03:05PM +0100, Niels Roest wrote: > > I understand that in some input device cases, a keymap will not present. > > Using an external utility doesn't seem very portable to me. I mean, you > > would want DirectFB to solve this in the library (or the keyboard input > > module), rather than linking an external app against DirectFB. See it as > > extended configuration.. > > I agree that loading it from a file may not be the best way, but I would > > rather solve it by changing it to a more 'portable' API.
IMO DirectFB should only provide the API that is needed to implement the configuration file parser. It should allow applications to configure the keymap in case that the kernel doesn't do this for a particular input device. > Hmm. Yeah, you're right. I didn't think about that scenario. But the > current code doesn't really solve that issue either since AFAICS there's > no way to load the keymap via directfbrc, but that should be simple to > fix. So it seems that it is best to keep the keymap parsing code in the > library instead of moving it. Parsing a keymap file appears somewhat out of scope for DirectFB. IMO it would be best if dfbinput would contain such code, serving as an example that shows how the DirectFB keymap API can be used. Sven _______________________________________________ directfb-dev mailing list directfb-dev@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
