Re: picoLCD improved key input
Martin T Jones <[email protected]>
| Newsgroups | gmane.comp.sysutils.lcdproc |
|---|---|
| Message-ID | <[email protected]> |
On 13/10/11 18:14, Markus Dolze wrote: > On 19.08.2011 20:32, Martin T Jones wrote: > >> Hello, >> >> I agree, drivers should be self-contained. I only include main.h for >> PROCESS_FREQ which is used to convert the auto-repeat times to counts >> for use in get_key, perhaps it should be moved to a shared header. >> >> Initialisation& de-initialisation of libusb-1.0 is quite >> straightforward and only needs to move somewhere common to support >> multiple USB drivers. >> >> The only "FIXME" that needs some thought is the one in main. My code >> is very simple and works well for a single driver but this really >> needs to be done properly, >> the libusb-1.0 documentation is quite good, see my comment for where >> to look. >> >> Martin >> > Hi Martin, > > I got time to look at this again. > > The libusb-1.0 (de)initialization and processing of asynchronous > messages should really happen somewhere within LCDd if multiple drivers > use libusb-1.0. > > For now I'd like to commit it 'as is', but with handling of asynchronous > messages moved to picolcd_get_key(). Why? because this moves all > libusb-1.0 stuff into picolcd and get_key() is called as often as if it > were in LCDd core. > > The configure-stuff is non-portable as is the use of some libusb-1.0 > functions (libusb_detach_kernel_driver). But this only affects FreeBSD > and I may fix that later. > > Regards > Markus > > _______________________________________________ > LCDproc mailing list > [email protected] > http://lists.omnipotent.net/mailman/listinfo/lcdproc > Hello Markus, That seems a sensible step, it makes the libusb-1.0 code available for the picoLCD driver and allows integration into the core to be postponed until other drivers have been converted. Regards Martin