This reverts commit ac84d3ba81f08036308b17e1ab919e43987a3df5.
There seems to be a problem with this API change. Reverting for now to
fix the build.
Change-Id: Ifa7426b080651b59afbcec2d3ede09a3ec49644c
Not yet working, so turned off.
Also fix a bug where the display size configuration became inconsistent
after a configuration change -- we now figure out everything about the
display size when computing a new configuration.
1. Added an Input Filter that interprets the touch screen motion
events to perfrom accessibility exploration. One finger explores.
Tapping within a given time and distance slop on the last exlopred
location does click and long press, respectively. Two fingers close
and in the same diretion drag. Multiple finglers or two fingers in
different directions or two fingers too far away are delegated to
the view hierarchy. Non moving fingers "accidentally grabbed the
device for the scrren" are ignored.
2. Added accessibility events for hover enter, hover exit, touch
exoloration gesture start, and end. Accessibility hover events
are fired by the hover pipeline. An accessibility event is
dispatched up the view tree and the topmost view fires it.
Thus predecessors can augment the fired event. An accessibility
event has several records and a predecessor can optionally
modify, delete, and add such to the event.
3. Added onPopulateAccessibilityEvent and refactored the existing
accessibility code to use it.
4. Added API for querying the currently enabled accessibility services
by feedback type.
Change-Id: Iec03c6c3fe298de3f14cb6efdbb9b198cd531a0c
Applications now get the display size from the window manager. No
behavior should be changed yet, this is just prep for some real
changes.
Change-Id: I2958a6660895c1cba2b670509600014e55ee9273
The mapper sends a DOWN when the finger is released then starts
a short timer. If the finger is not pressed again before the timer
elapses, then the mapper sends an UP to complete the tap. If the
finger is pressed again then the mapper starts sending MOVEs
as part of a drag.
Double/triple taps work as intended because we also start watching
for a new tap when the finger is pressed again. If a new tap
occurs the old tap is also finished. So each individual finger
short press/release cycle constitutes a distinct tap.
Change-Id: Iaa51182edee4378162a66adfd746088e93af2a30
Added a new PointerIcon API (hidden for now) for loading
pointer icons.
Fixed a starvation problem in the native Looper's sendMessage
implementation which caused new messages to be posted ahead
of old messages sent with sendMessageDelayed.
Redesigned the touch pad gestures to be defined in terms of
more fluid finger / spot movements. The objective is to reinforce
the natural mapping between fingers and spots which means there
must not be any discontinuities in spot motion relative to
the fingers.
Removed the SpotController stub and folded its responsibilities
into PointerController.
Change-Id: I5126b1e69d95252fda7f2a684c9287e239a57163
To avoid blowing past the Binder IPC limit, change the
PackageManagerService to have a DB-like interaction where the client
tells the service the last "row" that it read.
The fact that we use a HashMap instead of a TreeMap makes this
problematic. For now we're just making a new ArrayList for the keys and
then sorting them for each call. This can make the API slower for callers
of this, but it's probably greatly overshadowed by the cost of the data
transfer itself.
Bug: 4064282
Change-Id: Ic370fd148d4c3813ae4f2daffa1a7c28d63d5a09
- Create /data/user directory and symlink /data/user/0 -> /data/data for
backward compatibility
- Create data directories for all packages for new user
- Remove data directories when removing a user
- Create data directories for all users when a package is created
- Clear / Remove data for multiple users
- Fixed a bug in verifying the location of a system app
- pm commands for createUser and removeUser (will be disabled later)
- symlink duplicate lib directories to the original lib directory
Change-Id: Id9fdfcf0e62406a8896aa811314dfc08d5f6ed95
Without that lock, there is a chance of race condition
where while composing a specific index, requestBuf with
the same index can be executed and touch the
same data that is being used in initEglImage.
(e.g. dirty flag in texture)
We now write battery history directly into a buffer, instead of
creating objects. This allows for more efficient storage; later
it can be even better because we can only write deltas.
The old code is still there temporarily for validation.
Change-Id: I9707d4d8ff30855be8ebdc93bc078911040d8e0b
Sometimes the virtual keyboard was not hidden when switching between
applications. An example of this was when launching the browser from
the Google Search widget:
1) Tap the Google Search widget and enter some text, e.g. "google"
2) Select one search items, e.g. "google maps"
3) Browser opens. Press back button.
4) Select an item again, e.g. "google maps" - Keyboard does not
close.
When switching application, the virtual keyboard needs to find a new
Z position (window index) among the other windows. Normally it is
placed on top of the first window that is visible and can get focus
(canBeImeTarget()).
With a new application being launched, there is
an exception: a special "starting window" is placed on top of the
Activity window while the application is starting up. Since this
window should not get input, we need to look below that window.
When doing this, the previous implementation assumed that the
first window below always was focusable. If it wasn't, the
input method was placed above the "starting window", which
caused confusion that led to the keyboard not being closed
automatically.
In the case of the Browser, it sometimes has a "fake TitleBar"
window that can not get focus and that is placed above the
Activity window.
With this fix, we now keep looking through the windows below
the "starting window" until we find a window that can receive
input.
Change-Id: I1117846eb0f57603e64329bd955e28182f98f226
You can remove sub-tasks inside of a task, or an entire task.
When removing an entire task, you can have its process killed
as well.
When the process is killed, any running services will get an
onTaskRemoved() callback for them to do cleanup before their
process is killed (and the service possibly restarted).
Or they can set a new android:stopWithTask attribute to just
have the service automatically (cleanly) stopped at this point.
Change-Id: I1891bc2da006fa53b99c52f9040f1145650e6808
Don't keep unused global references to classes, don't throw exceptions
when an exception is already pending, and fix a (harmless) misunderstanding
about how GetStringChars works.
Change-Id: Ie445036f057daa8a1c76aceb7bad2a84fb81d820
We can help you with that.
Note also that getParcelFileDescriptorFD did no such thing. All its callers
were passing in a regular java.io.FileDescriptor and expecting the int. No
ParcelFileDescriptors involved.
Change-Id: Idc233626f20c092e719f152562601f406cc1b64a
Refactored PointerController to extract the surface management
code to a new component called a SpriteController so that it can
be used to move finger tracking spots around the screen as well.
The SpriteController is designed to fully decouple the client from
any latency introduced by surface transactions and drawing.
All sprite updates are performed asynchronously on the Looper using
a copy of the sprite state.
Added a stub SpotController implementation for touch pad UX.
It will be implemented in a subsequent patch.
Fixed a little bug in pointer orientation changes when entering
DISPLAY_ORIENTATION_90 the x offset was clobbered.
Change-Id: Ib25d162d577c9b354cb74d5d761c3c9f9f438d42
Some API stubs for managing users and storing their details.
List of users is stored in an xml file.
Each user's properties are stored in a separate xml file.
Some unit tests for modifying the XML files.
Change-Id: If2ce2420723111bd426f6762def3c2afc19a0ae5
This leak was intentional, it was there to deal with the fact that
some gralloc implementations don't track buffer handles with
file-descriptors so buffers needed to stay alive until there were
registered, which is not guaranteed by binder transactions.
In this new implementation, we use a small BBinder holding a
reference to the buffer, which with tuck into the parcel. This forces
the reference to stay alive until the parcel is destroyed, which
is guaranteed (by construction) to happen after the buffer is
registered.
this allows the public facing API to not expose the previous hack.
Change-Id: I1dd6cd83679a2b7457ad628169e2851acc027143
