Alt-TAB should have different semantics from the APP_SWITCH key
or long-press on HOME. Accordingly, remove the fallback action
for Alt-TAB and initiate the task switching behavior directly
in the policy.
Modified RecentApplicationsDialog to be more precise about the
initial modifiers that it considers to be holding the dialog.
The dialog is now dismissed by a second press on the APP_SWITCH
key or by a second long press on HOME.
Change-Id: Idf4d803f51103819057cb655ff3b770b7729e4be
The idea is to assist with debugging by identifying cases in which
the input event stream is corrupted.
Change-Id: I0a00e52bbe2716be1b3dfc7c02a754492d8e7f1f
This patch adds a mechanism for capturing, filtering, transforming
and injecting input events at a very low level before the input
dispatcher attempts to deliver them to applications. At this time,
the mechanism is only intended to be used by the accessibility
system to implement built-in system-level accessibility affordances.
The accessibility input filter is currently just a stub.
It logs the input events receives and reinjects them unchanged,
except that it transforms KEYCODE_Q into KEYCODE_Z.
Currently, the accessibility input filter is installed whenever
accessibility is enabled. We'll probably want to change that
so it only enables the input filter when a screen reader is
installed and we want touch exploration.
Change-Id: I35764fdf75522b69d09ebd78c9766eb7593c1afe
Added new drag_can_accept and drag_hovered XML attributes and the View
logic to support them. Drawable states are now refreshed automatically
when a drag starts/ends and when a drag crosses the boundary of a
participating view.
Change-Id: I25f8ee02c83b3fa4f27201997d7eabf4be653fd8
("setCorrectionSpan" was added in Id3abc9ea4d11753cd )
Also..
- Added a class java doc for CorrectionSpan
- Removed FLAG_DEFAULT
- Changed the return type of getSuggestions from Array<CharSequence> to String[]
Change-Id: If5eb091e307a7a40c5b4a70ec1fe6059ecd9fb2d
The uniqueness of InputMethodInfo was guaranteed by mId (like InputMethodInfo#equals), but the hashCode was not implemented in the same way.
This change fixes a problem happening when the user gets the hashCode of InputMethodInfo obtained through IPC.
Change-Id: Ib876c5cb0d778481100597ec31202f94fb7b8f37
The input dispatcher sends a HOVER_ENTER to a window before dispatching
it any HOVER_MOVE events. For compatibility reasons, the window will
*also* receive the HOVER_MOVE. When the pointer moves into a different
window or the pointer goes down or when events are canceled for some reason,
the input dispatcher sends a HOVER_EXIT to the previously hovered window.
The view hierarchy behavior is similar. All views under the pointer
receive onHoverEvent with HOVER_ENTER followed by any number of HOVER_MOVE
events. When the pointer leaves a view, the view receives HOVER_EXIT.
Similarly, if a parent view decides to capture hover by returning true
from onHoverEvent, the hovered descendants will receive HOVER_EXIT.
The default behavior of onHoverEvent is to update the view's hovered
state by calling setHovered(true/false). Views can query their current
hovered state using isHovered().
For testing purposes, the hovered state is mapped to the pressed
drawable state. This will change in a subsequent commit with the
introduction of a new hovered drawable state.
Change-Id: Ib76a7a90236c8f2c7336e55773acade6346cacbe
bug:2513822
Text added to accessibility events is truncated to max
length. However, the fromIndex and toIndex properties
are relative to the text before being truncated, thus
potentially our of bound. Removed the max length
limitation because test very rarely is longer that
500 characters and in the cases in which text is longer
than 500 character it will not be a real problem to
pass a bit more data through an IPC (very rarely).
Change-Id: Ie70ac630dfeb56d4f59079abb2f46d07582796b6
WebCore is too slow
Make sure that we can recover properly from a bad gesture with missing
events that never come back from webcore. Lower timeout to 1 second.
Confirm movement on touch event enqueue so that we don't get phantom
taps or long presses when webcore is slow to respond.
Add sanity check in ScaleGestureDetector to end a gesture early on a
bad MotionEvent stream rather than throwing up.
Change-Id: I69690409d7edd6b4320dbcf3b052aba4023360fe
- CorrectionSpan is a span which has suggestions made by IME.
This has a function to change the current IME to other IME specified
in this span. For security reasons, only the current IME
is allowed to use this function through InputConnection.
(IME token is used for checking the validity of it.).
- CorrectionSpan stores following information:
flags, subtype Id, InputMethodInfo Id, suggests, locale, original string
Change-Id: Id3abc9ea4d11753cdc4f483a2bb3128f49ba198a
bug:2133127
If a View is disabled its onTouchEvent handler uses a shortcut
path for efficient handling which does not clear the pressed
flag. As a result if the view is disabled in a long click
handler the view keeps being in pressed state. The fix is to
clear if needed the pressed flag in the shortcut path
of onTouchEvent.
Change-Id: I046a62c2fce751c27fec3cfb756b74431fc003c3
Replaced VelocityTracker with a faster and more accurate
native implementation. This avoids the duplicate maintenance
overhead of having two implementations.
The new algorithm requires that the sample duration be at least
10ms in order to contribute to the velocity calculation. This
ensures that the velocity is not severely overestimated when
samples arrive in bursts.
The new algorithm computes the exponentially weighted moving
average using weights based on the relative duration of successive
sample periods.
The new algorithm is also more careful about how it handles
individual pointers going down or up and their effects on the
collected movement traces. The intent is to preserve the last
known velocity of pointers as they go up while also ensuring
that other motion samples do not count twice in that case.
Bug: 4086785
Change-Id: I2632321232c64d6b8faacdb929e33f60e64dcdd3
Bug #4092053
The problem always existed but was made visible by partial invalidation.
When saving a layer, the renderer would try to postpone glClear()
operations until the next drawing command. This however does not work
since the clip might have changed. The fix is rather simple and
simply gets rid of this "optimization" (that turned out to be
usless anyway given how View issues saveLayer() calls.)
This change also fixes an issue with gradients (color stops where
not properly computed when using a null stops array) and optimizes
display lists rendering (quickly rejects larger portions of the
tree to avoid executing unnecessary code.)
Change-Id: I0f5b5f6e1220d41a09cc2fa84c212b0b4afd9c46
Associate each motion axis with the source from which it comes.
It is possible for multiple sources of the same device to define
the same axis. This fixes new API that was introduced in MR1.
(Bug: 4066146)
Fixed a bug that might cause a segfault when using a trackball.
Only fade out the mouse pointer when touching the touch screen,
ignore other touch pads.
Changed the plural "sources" to "source" in several places in
the InputReader where we intend to refer to a particular source
rather than to a combination of sources.
Improved the batching code to support batching events from different
sources of the same device in parallel. (Bug: 3391564)
Change-Id: I0189e18e464338f126f7bf94370b928e1b1695f2
Report inclusive minimum and maximum ranges for all
axes including X and Y.
Set mouse pointer bounds to 0..width-1, 0..height-1.
Rotate touch and mouse positions more carefully, paying attention
to the maximum bounds when calculating the complement of an axis.
Simplified the InputReader somewhat and removed support for a
couple of poorly defined input device configuration parameters.
We now assume that the touch device provides useful absolute axis
ranges for the X and Y axes since the alternative does not actually
make sense.
Bug: 3413541
Change-Id: I121d28a125c4f9618cb283dc460d33ff1a907023
