> > Ah cool, then I guess we can remove the cases for "4" and "8"? I'd be > nice to stick with decimal numbers in all cases for consistency. > Oh yeah we totally could, I'm not sure why the double cases were there in the first place - it might have been for another device included in the original patch. > > Also, buf[6] does not indicate the number of buttons (20?). > > Did you check with Wireshark if the Windows driver is doing something > different for your tablet? It'd be nice to avoid adding quirks but it > might not be possible :S > > We can ignore buf[12] and buf[14]. buf[0] indicates the size of the > descriptor (12), so the last two bytes are random memory. > I don't even think I have to drop into Windows, we have a userspace linux program for XP-PEN but I haven't tested it with this tablet yet. Yeah, I'd also like to avoid implementing device-specific quirks - hopefully I'll be able to shrink the init code but I might not be able to get rid of all of it. It's really weird that almost everything else (data frame wise) is the same, with the exception of init. And yes, this tablet contains 20 buttons and nothing in that buffer indicates button count :-/
