Does anyone disagree that 5327 is a bug which needs to be fixed?

[Deron Johnson]

So far, what I'm hearing is that no one disagrees with my assertion that
mouse wheel events should not be triggering grabs.

I agree that hardcoding '< 4' into this is a bad idea. However, I don't
know enough about the subject to know what the correct piece of general
code is. Would someone be willing to take an action item to determine
what the correct form of the test should be?

I also understand that some people feel that we shouldn't be
representing mouse wheel events as button events in the first place.
This may be a topic worth of further discussion, but to me it seems like
a slightly separate topic. At the present I am try to focus on fixing
the existing implementation, which clearly has a bug in it.

Lubos Lunak wrote On 01/19/06 08:57,:
> On Thursday 19 January 2006 15:58, Ely Levy wrote:
> 
>>On Thu, 19 Jan 2006, Matthias Hopf wrote:
>>
>>>However, we'll have to keep button events for a long time in order not
>>>to break 'older' (that is: current) applications.
>>>
>>>
>>>>Today most programs just guess it to be button 4-5 and I already saw a
> 
> 
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> 
>>>>lot of people complaining after suddenly IMPS/2 protocol changed it to
>>>>6-7 in xorg 6.9.
>>>
>>>Sure. 4/5 is standard for wheel 1 (y axis), 6/7 for wheel 2 (x axis).
>>
>>standard?!!?
>>How is it a standard?
> 
> 
>  That's how it's a (de facto) standard. Just run xmodmap -e "pointer = 1 2 3 6 
> 7 4 5" and watch your Qt/KDE apps, Gtk/GNOME apps and who knows what apps 
> suddenly scroll horizontally instead of vertically.
>