Re: master 101bbd1392: Add support for pinch gestures to the XI2 build

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Cc: emacs-devel@gnu.org
> Date: Sun, 26 Dec 2021 18:39:09 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > Yes, but you could generate several Emacs events in make_lispy_event,
> > can't you?  Or maybe simply ignore the angle data?
> 
> I meant to say that those angle deltas only make sense as part of a
> pinch event, and they're not supposed to be treated as separate rotation
> events.  (Which we will want to expose to Lisp separately, once the
> XInput developers add support for them.)

Please explain how the angle deltas make sense as part of a pinch
gesture, I don't think I understand that.

> > I don't think I understand how read-event with a timeout will help
> > here: you still get a series of events, AFAIU.
> 
> Yes, but it solves the flickering problem.

That's not the only possible solution, and neither is it the cleanest
one, IMO.

> > Walking the event queue looking a for a series of pinch events is much
> > easier.
> 
> Yes, but limiting the frequency of these events is something not all
> commands might want.

We are talking about the situation that several pinch events were
reported to Emacs since the last time it entered the idle loop and
checked for input.  Which commands do you envision that would like to
process only some of that gesture, and leave the rest in the input
queue?  AFAIU, all it will do is cause some kind of "inertia", in that
the reaction to user gestures will lag after the gestures.  Why would
that make sense?