Comment 11 for bug 358643

I was able to build wacom-tools just fine and I have tested both the
patch in comment 8 and the xserver-xorg-core (2:1.6.0-0ubuntu11.1)
from Timo's PPA.

With only the patched xserver-xorg-input-wacom (1:0.8.2.2-0ubuntu2)
there wasn't any difference from the unpatched one. I wasn't able to
reproduce the result from comment 7 where the xserver would crash
right away when the "Device" was set to "/dev/ttyS0" in either case,
and I don't know why.

The new xserver-xorg-core from PPA solves the problem with the corrupt
TTYs. As for the rest:

With xserver-xorg-core from PPA (without patched wacom):

If /dev/ttyS0:
No crash or hang!

If /dev/input/wacom:
On boot, the first X-session works fine. On logout, the next X-session
hangs. If I start with /dev/ttyS0 and during the X-session edit to
/dev/input/wacom, it hangs.
When it hangs, it the log ends with:
usbDetect: can not ioctl version
xf86WcmSerialValidate: bad magic at 3 v=c0 l=9
On first session after boot, the log is basically the same to this
point, but it continues and on logout the UnloadModule: "wacom" seems
to go fine and there is no Backtrace at the end.

With xserver-xorg-core from PPA and patched xserver-xorg-input-wacom:

If /dev/ttyS0:
No crash or hang!

If /dev/input/wacom:
Works mostly fine. Still has the "bad magic" lines in the log and
takes longer to get login screen (maybe it hangs for a while at the
"bad magic at 3"). It hung on me once (first session after boot), at
the same place where the second session after boot would hang every
time without the patch.

I'm still a little puzzled that it makes a difference whether I use
/dev/input/wacom or /dev/ttyS0 as "Device" in the "InputDevice"
section. Maybe I should try what happens if I use a bogus device name?
(but not tonight)