On 03/24/2017 02:06 PM, Peter Hutterer wrote:
> On Fri, Mar 24, 2017 at 11:52:22AM -0700, Jason Gerecke wrote:
>> On Thu, Mar 9, 2017 at 5:53 PM, Peter Hutterer <peter.hutte...@who-t.net> 
>> wrote:
>>> On Thu, Mar 09, 2017 at 07:37:39AM +0100, Sebastian Reuße wrote:
>>>> Hello Peter,
>>>>
>>>> Peter Hutterer <peter.hutte...@who-t.net> writes:
>>>>
>>>>> On Wed, Mar 08, 2017 at 07:25:53PM +0100, Sebastian Reuße wrote:
>>>>>> When running on Wayland or when using the Xorg libinput driver (which is 
>>>>>> the
>>>>>> default on some distributions), a call to XkbGetKeyboard always returns 
>>>>>> NULL
>>>>>> (causing xsetwacom to segfault) and is apparently not supported anymore 
>>>>>> (cf.
>>>>>> [1,2] and others). Instead, we are supposed to use XkbGetMap now.
>>>>>
>>>>> patch looks good for what it does, but the commit message is confusing.
>>>>> xsetwacom is a tool specific to the xorg wacom input driver. It won't work
>>>>> on wayland because xwayland tablets work completely different and it won't
>>>>> work on the xorg libinput driver because all the properties are different.
>>>>> Did you actually see a crash here, and if so what was the command and 
>>>>> setup
>>>>> to trigger that crash?
>>>>
>>>> Yes, I saw a segfault when running «xsetwacom --set $id Button 1 "key
>>>> $anykey"», which gdb revealed was due to XkbGetKeyboard always returning
>>>> NULL. The only references that came up in a search mentioned Wayland, so
>>>> my theory was that Archlinux now using libinput as a default was the
>>>> cause (I thought libinput somehow backs all the Xorg input, but I
>>>> realize now that I was mistaken). Using «XkbGetMap» instead fixed the
>>>> issue for me.
>>>
>>> yeah, two different problems, xsetwacom cannot work under wayland or with
>>> xorg-libinput. and I checked the code, it does the the property checking
>>> correctly.
>>>
>>>> However I just realized something entirely different may be the cause.
>>>> I’m using the AdNW [1] keymap (sort of a German dvorak) as installed by
>>>> running the script in [2]. When I load a standard keymap such as
>>>> «setxkbmap de», I can no longer produce the segfault using the unpatched
>>>> version of xsetwacom.
>>>>
>>>> I’m not really familiar with Xorg internals. Are you able to make sense
>>>> of this by any chance?
>>>
>>> yeah, my guess is because you're loading the keymap directly instead of
>>> going through the RMLVO you're missing the keymap names and things go boom.
>>> A more detailed analysis requires me swapping too much of XKB again, so I'd
>>> rather go with the oneliner patch :)
>>>
>>> Cheers,
>>>    Peter
>>>
>>
>> Just making sure this doesn't get lost. Does this only need an update
>> to the commit description to make it more accurate?
> 
> yes, that's all that's needed here.
> 
> Cheers,
>    Peter
> 

Poking Sebastian :)

-- 
Jason
---
Now instead of four in the eights place /
you’ve got three, ‘Cause you added one /
(That is to say, eight) to the two, /
But you can’t take seven from three, /
So you look at the sixty-fours....

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Linuxwacom-devel mailing list
Linuxwacom-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel

Reply via email to