[dropping cross-post to ubuntu-x because it's subscriber-only] On Wed, Mar 24, 2010 at 18:23:04 +0200, Timo Aaltonen wrote:
> > Hi! > > I've backported support for some features from xserver 1.8 to 1.7.6: > > - xorg.conf.d > - inputclass > - udev (the one we have is slightly different) > - "various patches on top of these to fix some bugs or added features" > > and last, but not least, from bugzilla: > > - finally a proper fallback patch from SuSE to fix autoloading when there's a > conffile (which would've been always the case with xorg.conf.d) > https://bugs.freedesktop.org/show_bug.cgi?id=27229 > > Tested on lucid and seems to work fine, fallbacks are loaded even > when there's the evdev catchall-config in place: > > http://pastebin.com/PB4a80i4 > Hi Timo, There's one concern I expressed on IRC, which is that I don't much like shipping something as 1.7 which doesn't have the same ABI as 1.7 upstream. (And shipping with 1.8 is too late for lucid and quite a big jump even for squeeze at this point.) Removing the need for an extra parameter in NewInputDeviceRequest is probably not possible with the inputclass/inputattributes changes, so one option if we want to do the backport anyway would be to keep the old NIDR as-is, and use NIDR18 when we need to pass InputAttributes. Ideally wacom could keep using the old NIDR, since the inputclass matching could be done for the first device, hotplugged through config/udev, and the additional ones wouldn't need the InputAttributes. but I haven't looked at the code at all, so maybe not. Does this make sense? Any chance you could test it? Other than that, I think it would be good to include this, because the sooner we get configuration out of udev rules the better. Maybe also consider including Dan's patch to use the device name without double quotes. Thanks, Julien
signature.asc
Description: Digital signature
