Mail Archives: geda-user/2011/10/01/21:10:39
On 10/01/2011 07:47 PM, Peter Clifton wrote:
>>> You might even be able to set the gains to negateive of their normal
>>> numbers to enable a kind of "natural scrolling" feature - I believe that
>>> is common on Mac?
>>
>> If that means "smooth scrolling" rather than "jumping in chunks", yes.
>
> Smooth scrolling is different to what I meant. I'm not familiar enough
> with GTK on OS-X to know whether we can get high precision scroll events
> from the "mouse".
Gotcha. I'm actually running Linux, though. I am using an Apple
desktop trackpad with Ubuntu 11.04 on x86_64.
> On Win32 (where I'm most familiar with the problems with the scroll
> wheel), there is a message parameter with the scroll event which tells
> you how many units the wheel has moved. When smooth scrolling is enabled
> (not sure how that happens), Windows will send you more messages.
>
> On X11 (and the Mac OS X version of PCB uses X11 of course), the wheel
> events are translated into button presses - not a valuator. This seems
> to imply that unless there is an API for getting at the high-res "wheel"
> data, we would have to fake the smooth scrolling - meaning things would
> still move in quanta - just that we might be able to animate it a bit.
Understood. That would be better than nothing; the erratic jumping
is pretty distracting.
> Another technique associated with smooth scrolling is "kinetic
> scrolling", where you assign some notional mass to the object being
> scrolled, and use that to implement accelerations and decelerations of
> the scroll animation. When this is got right, it works amazingly. It can
> be very infuriating when its got wrong though. (And on Apple's platform,
> where they put a lot of effort into getting these things right - we will
> have a high standard to keep up to!)
Yes, I miss that from OS X. Firefox does it under Linux, though.
>>> Try (scrollpan-steps -8) if you want that, the code comments seem to
>>> suggest this is possible.
>>
>> I tried that; I was expecting smaller steps to the scrolling, but it
>> didn't seem to have any effect. I'd dearly love a less "jumpy"
>> scrolling action. But frankly, having two-axis scrolling as above is a
>> huge deal in itself!
>
> If you increase the magnitude of the number, the scroll quanta will get
> smaller. (Sign sets the direction). But unless we can get more events
> from the input device. this will mean you scroll gets slower as well as
> less jumpy.
Ok, I will play with it a bit. Thanks!
-Dave
--
Dave McGuire
Port Charlotte, FL
- Raw text -