On Wed, Nov 24, 2010 at 07:30:22PM +0000, Stuart Henderson wrote: > On 2010/11/24 19:06, Antoine Jacoutot wrote: > > On Wed, 24 Nov 2010, Miod Vallat wrote: > > > > > > But is there any reason to keep these devices in uscanner? To my > > > > knowledge, sane is the only tool to access such devices. Is there > > > > other software that need uscanner? > > > > > > > > And more generally, is there any reason to keep uscanner? > > > > > > According to the manpage, it was written to provide a linux-compatible > > > scanner device. If nowadays' scanning applications are perfectly happy > > > with ugen(4), then I see no point in keeping uscanner(4). Unless I > > > misunderstood things... > > > > Well, I don't own many scanners. All I can say is that uscanner is > > deprecated in linux, they now use libusb. I also do so without issue > > but I cannot guess it'll be the same for all scanners around. > > The expected scenario is that all usb scanners *should* work with libusb > > where only some will also work also uscanner. > > > > This calls for testing from people who own such hardware. Personally I'd > > be happy to see uscanner move away, I wouldn't have to config(8) my > > kernel all the time. > > How about removing uscanner from GENERIC for now, then if nobody > has a problem with it, remove the code at a later date? (I would > suggest picking that date in advance so it doesn't sit around for > ages). >
I'd rather nuke it completely. Putting it back is not much more difficult than tweaking GENERIC and is more likely to flush out people actually using it. Who will change GENERIC, not say anything, and then get upset when the driver is removed. Just my 2 cents. .... Ken
