Dream settings are stored per-user, so dream manager operations
must act according to the calling or current user.
Bug:7041514
Change-Id: I4a0bbbd76886e6440b1afd89c61af5f4569b0e18
The window manager now has a facility to provide a full-screen
animation, which the activity manager uses every time a user
switch happens.
The current animation is just a simple dumb slide until we get
a design from UX.
Also some cleanup: moved the portrait task animations to the
default config so we always have an animation for them, and finally
got the java symbol stuff out of public.xml.
Change-Id: I726f77422b2ef5f2d98f961f8da003e045f0ebe8
Otherwise there would be no way to switch users if a user chooses None as lock screen.
Also adjust some layout params for Preference items to align the icons and text.
Change-Id: Ib11bb961edaef90de280bc7137d10b68765aea98
(1) It's okay to write literal null as a settings element value
(2) Properly convey the user handle in the put-for-user variant
Bug 7137201
Bug 7139826
Change-Id: I0ed77d65e8377f0e0580a2668f10b7167ad34928
Quick Settings and global actions (aka longpress-Power) toggles are
included in this CL. The Settings app manipulations are updated in
a different CL because that's a different git package.
Bug 7132230
Change-Id: I50838f5e8d7b25a750d2bcae90bf384b09816dbb
Bug: 7073422
Create the plumbing to use an annotation to allow access to
inherited methods of jsinterface objects. The default webview
behavior has not changed yet. However internally an a flag is
introduced to restrict javascript access to methods that have an annotation.
Change-Id: I41927248e6bc4b09d17c0707c60fe5e6ab681e66
Bug: 6961494
Don't call clearHelpers in destroy(). It is already called
in onDetach, which is where it needs to be. It can't be
called by the finalizer
Switch onCancel listener to onDismiss listener. Fixes an issue
where destroying can live the webcore thread hung because onCancel
won't be called, but onDismiss is
Change-Id: I669c389c8c9a81d92c27f88d83bcb711c6c0a5ee
This effectively makes the pattern view slippery, which enables the
user to start their pattern outside the view and cross the view.
Fixes bug 7118762
Change-Id: I085c5ec8f7ccd2fc37a10606a8925078dc44be13
Now we default to the current user instead of user 0 for most commands
(except where we can do the command for all users).
Many more commands take a user argument: force-stop, kill, profile,
dumpheap.
Improved help text.
Change-Id: I719a13b4d31b668f57ca21e51d7043ac3e0d4e1b
This CL extends the HELO and APNM packets to include the user id
of the application. This allows ddms users to differentiate between
the same app running for multiple users.
Bug 7110696.
Change-Id: I490d0e3781b8fb9db65cf81188677a0955650511
Trackball and generic motion events now pass through the IME in case
it would like to handle them before passing them on to the view
hierarchy.
While I was at it, I also...
...fixed the documentation on InputMethodService.onKeyUp()
...added documentation to InputMethodService.onTrackballEvent()
...added trackball and generic motion events to the "input" command
...fixed input consistency verification involving ACTION_OUTSIDE
Bug: 7050005
Change-Id: I40ab68df4a9542af6df25de6ec2ec500e4c02902
Previously, clone() on an Animator with only one value would mistakenly
think that the clone had a real starting value (which would end up being 0 in the
int and float cases). Fix is to set the 'mHasFirstValue' flag appropriately for the
clone, based on the state of the cloned animator.
Issue #7106442 ObjectAnimator.clone() does not work properly for single parameter
Change-Id: I08bf03b7687a65eb613c1671a58e4cbfae66a30e
Instead of unpacking libraries into /data/data/<appname>/lib, unpack
them into /data/app-lib so that multi-user applications can use the same
libraries.
Change-Id: I636dafb5a9eea8cf5a41f1360698e7695455f385
Some camera apps may wish to replace the system camera shutter sound
with their own, especially if they are taking rapid bursts of
images. Add method to allow this when possible.
Hidden for now.
Change-Id: I6520f5441d28675626fafab48c6609c589fc6f7e
This change makes the SurfaceTexture JNI updateTexImage call the native
SurfaceTexture's doGLFenceWait method to perform the needed synchronization.
Change-Id: Ie70a1fe6b44d439d1ffe7b97689a421ff8c02fda
In order to make sure we have unpacking and deletion working, add a fake
JNI library to every APK.
This can be expanded in the future to make sure replacement is working
correctly as well.
Change-Id: I1ad04c61df9518429aa17795b8a9d3608791cd8e
Make sure that native library paths have the correct permissions and
owners. Also make it easier to do a "String#startsWith" test so we can
get better errors when it fails.
Change the observer to have more sane semantics with a CountDownLatch
instead of weird synchronization on a local variable.
Change-Id: I2437e5ea886d6e6cb8b4edeab80d6053b79857d4
Change all the test method headers to throw Exception so we don't have
to try to catch anything during the tests. Catching an exception might
mask an error during the tests.
Also fix a lot of formatting errors that have built up over the years.
This includes 3-space indentations instead of 4-space, missing spaces
between braces and parens, too-long lines, wrapping when it isn't
needed, etc.
Change-Id: Ifccea5f0706b6f9235f37974a5f80bee76c654dd
After a display is removed, there may be a brief time
when a Context still exists that is bound to it.
We need to provide metrics in this case.
Bug: 7131637
Change-Id: I11b264a000653adbf0f3da399eaab66c4b40fb2a
