[Xorg] [PATCH] allow EmulateWheel to generate normal clicks too

[Andrew Pimlott]

I just saw that this patch was committed, so let me just add a couple
more comments.

I have been using Mathias's patch (the one committed, which works with a
timeout) for the last several weeks.  I have occasionally had the
experience of trying to middle-click, but holding the mouse too long and
therefor not getting a click.  (I have window delete bound to a
modifier-middle-click combinations, and I tend to pause before releasing
the mouse button to make sure I have the right keys pressed.)  The first
time this happened, it took me a few tries to realize what was
happening.  After that, I had no trouble remembering that I needed to
click a bit faster.  However, I'm not sure if this would occur to a
normal user who didn't know the implementation; it really needs
real-world testing.

I would add that if I try really hard, I can press the middle button,
move the cursor, and release the middle button quickly, and thereby
produce both a wheel event and a middle button event.  But I doubt this
could happen by accident.

Overall, I am comfortable with Mathias's patch and have felt no need to
switch back to mine.

On Mon, Aug 30, 2004 at 07:49:50PM +0200, Mathias Fröhlich wrote:
> I would now prefer my own timeout approach again for a simple situation which 
> happens to me often:
> I press the middle mouse button in advance to scroll the window I am watching 
> now. But I sometimes decide not to scroll and relese the middle button again 
> with little movement. With Andrews approach I then get the mouse button 
> event, which is defintly not what I wanted to have. This did not happen with 
> my timeouts.

In my own experience, I never do this.  (But maybe that's because I
subconsciously know the result with my patch!)  Regardless, if real
users will make this "mistake" (and they probably will if Mathias does),
it argues strongly in favor of Mathias's patch.  So I think applying his
was a good choice.

Finally, I want to offer that if any users complain about the behavior,
I will happily help to improve this feature until it is satisfactory.

Thanks, Alex, for following up on this.

Andrew

PS.  Please Cc: me on any future mail, as I will not be subscribed.