This widget is too narrow to make touch interaction tricky enough, so
don't add the penalty of having the slider run farther from the touch
coordinates if it happens to miss the slider.
Instead of just stopping the acceleration source ID, check whether
it's overshooting, and let it snap back if needed after ::grab-notify
and ::button-release-event
Velocity calculation has been refactored out of the captured motion events
handler, and also has more tolerance defore determining that a drag is actually
still.
This is so submenus stay open as the parent menu item is pressed/released,
as the user would typically lift the finger in order to select a submenu
item.
_gtk_widget_set_device_window() is suppose to make accounting of
the topmost widget under the device at each time, so avoid setting
it on virtual crossing events as the device is already in another
window.
gtk_scrolled_window_set_kinetic_scrolling() now takes a set of flags,
GTK_KINETIC_SCROLLING_CAPTURE_BUTTON_PRESS makes the "capture button
press and maybe replay later" vs "let button press go through, but
trust in ::grab-broken to undo things" an opt-in, by default that
flag is set, which is the most conservative approach.
This is so the grab doesn't break the implicit grab on the
child widget's window, which avoids that the button press and
release are possibly sent to different windows, and after the
grab was actually broken.
Anytime a touch device interacts, the crossing events generation
will change to a touch mode where only events with mode
GDK_CROSSING_TOUCH_PRESS/RELEASE are handled, and those are sent
around button press/release. Those are virtual as the master
device may still stay on the window.
Whenever there is a switch of slave device (the user starts
using another non-touch device), a crossing event with mode
GDK_CROSSING_DEVICE_SWITCH may generated if needed, and the normal
crossing event handling is resumed.
This last slave device (stored per master) is used to fill
in the missing slave device in synthesized crossing events
not directly caused by a device event (ie. due to configure
events or grabs)
Fixes the situation where overshooting in scrolled windows with a
small viewport could end up overshooting right to the opposite
side, and then back again indefinitely.
Request automatically more motion events in behalf of
the original widget if it listens to motion hints. So
the capturing widget doesn't need to handle such
implementation details.
If the widget which events are captured listens to motion hints, there
are situations where neither the scrolled window or the child widget
request more motions.
The maths being used didn't resemble much about velocities or
friction/deceleration, so reimplement it in terms of velocity
vectors and decelerations, measured in pixels/ms^2.
Overshooting is also handled within the deceleration effect,
turning into a constant acceleration vector in the opposite
direction so it returns elastically within the boundaries.
An extra GdkWindow has been added, this window is the parent
of the child widget, and is the one getting resized/moved when
overshooting.
The unclamped adjustments' values are also stored in
GtkScrolledWindowPrivate as a separate value, overshooting is
pretty specific to GtkScrolledWindow and it isn't worth to
expose API in GtkAlignment for this single purpose.
This method allows GtkScrollable children to be blissfully
unaware of overshooting, as otherwise they'd have to handle
rather odd adjustment values themselves.
The innermost scrolled window always gets to capture the events, all
scrolled windows above it just let the event go through. Ideally
reaching a limit on the innermost scrolled window would propagate
the dragging up the hierarchy in order to keep following the touch
coords, although that'd involve rather evil hacks just to cater
for broken UIs.
It now returns a GtkCapturedEventFlags which tells whether the
widget captured the event, and additionally whether the event
should be stored for later replay.
gtk_widget_release_captured_events() has been added too so
all stored events are released onto the widget that was initially
to receive the events.
Press and hold couldn't reasonably work if nested widgets
handle ::captured-event, once the widget inits press-and-hold,
it'd better also handle possible cancellation on motion and
button release, which isn't guaranteed with ::capture-event.
Also, tentatively start press-and-hold by default on the
grab_widget, and before event capturing happens, this avoids
awkward situations like the scrolled window preventing/delaying
press-and-hold to happen on the child textview for example.
When clicked again close to the previous button press location
(assuming it had ~0 movement), the scrolled window wil allow
the child to handle the events immediately.
This is so the user doesn't have to wait to the p-a-h timeout
in order to operate on the scrolledwindow child.
