Re: [gtk-list] Re: wheel support, was: Re: Plans for 1.3/1.4

[Paul Barton-Davis <pbd op net>]

>I don't think this is an X issue.  The FX does not send 4- and 5- clicks in re
>sponse
>to scrolling (That wouldn't work because it allows scrolling an any direction,
> not
>just vertical).  Button 4 sends a 4-click; it's the software that interprets i
>t to do
>the scrolling.  (The option is there in the Marble+ software too.  You can rem
>ap any
>button to the "Universal Scroll" behavior)  In Windows this was done at a driv
>er level
>because all Windows apps use a common toolkit, but in X it has to be done on a
> toolkit
>level, because the toolkit is the lowest level that knows anything about scrol
>ling.

I think this is wrong. It *is* an X issue. If the FX did sent 4- and
5- clicks (which it does not), and these were translated into motion
in two axes, you can scroll in any direction, although it would be
done via *two* events, not just one. 

However, the FX doesn't send 4- and 5- clicks, so the server needs to
map motion occuring while button 4 is pressed into the same thing.

The idea that you'd do this at the toolkit level when hardly any other
pointer devices do this seems odd to me. Its the server's job to
reduce the actual input hardware models to a common abstraction, and
the server can do this, it just doesn't do it right now.

Yes, the toolkit needs to be able to respond intelligently to the
translated button events, but thats a different (and much simpler,
although simultaneously more complex) story.

--p