[xorg-server-1.5.2] Mouse wheel trouble

[Tobias Jakobi]

Hi Peter!

Peter Hutterer wrote:
> On Fri, Oct 17, 2008 at 02:42:31AM +0200, Tobias Jakobi wrote:
>>       <merge key="input.x11_options.SendCoreEvents"
>> type="string">true</merge>
> 
> You don't need this option, it's the default anyway.
Thanks, going to remove this one then.

>>       <merge key="input.x11_options.RelHWHEELMapTo"
>> type="string">Buttons 5 4</merge>
> 
> this option isn't parsed by evdev, so it doesn't have any effect.
Hmm, are there currently any ways to configure evdev devices?
The manpages seems a bit outdated. The is nearly nothing there that has
to do with button mapping and such kind of stuff.


>> I have some settings in the /etc/hal/fdi/policy/10-x11-input.fdi.
>> I added the RelHWHEELMapTo setting after the device did not work, but
>> that didn't seem to fix it. If I remove the setting nothing changes. The
>> mouse still doesn't work after a restart of X (buttons 6 and 7 are
>> assigned to the wheel, which doesn't seem to be that what X expects) and
>> replugging makes the wheel work again.
> 
> xmodmap -pp when the buttons are screwed up. is the mapping 4:4 and 5:5?
Yeah, you're right.

When the mouse doesn't work I have a 4:6, 5:7, 6:4, 7:5 mapping. However
I haven't got any entries in my Xmodmap file.

How can I see where these mappings come from?



>> Testing ... (interrupt to exit)
>> Event: time 1224203149.075801, type 2 (Relative), code 8 (Wheel), value -1
>> Event: time 1224203149.075803, -------------- Report Sync ------------
>> Event: time 1224203149.363751, type 2 (Relative), code 8 (Wheel), value 1
>> Event: time 1224203149.363757, -------------- Report Sync ------------
> 
> the matching code in evdev is:
>    case REL_WHEEL:
>                 if (value > 0)
>                     PostButtonClicks(pInfo, wheel_up_button, value);
>                 if (value < 0)
>                     PostButtonClicks(pInfo, wheel_down_button, -value);
>                 break;
> 
> where wheel_up_button and wheel_down_button are hardcoded to 4/5. So I don't
> quite see what can go wrong there.
> 
> Cheers,
>   Peter
> 

Thanks,
Tobias