The query WindowState.isDisplayed did not take into account being
displayed due to app animations.
When an existing input method target was animating away the logic
for detecting if it was still on screen was faulty. This led to
assigning the input method to a layer below its target and obscuring
the input method until the animation was complete.
Bug: 7296703 fixed.
Change-Id: Ib00db4f21b726ed57d25d6a1e796b65a7d45ee97
Bug: 7086881
Load resources for the correct user.
Also clean up package monitoring and locking.
Added dump method to SearchManagerService.
Sneaking in a change to make crash dialogs visible to current user.
Change-Id: Id56dd15428d66084de995e28be242db27c15fda3
Always use the ElectronBeam now, even when we are only animating
the backlight so that we will have a black surface remaining
on the screen after the screen turns off.
When turning on the screen, keep the black surface showing until
we unblock screen on then dismiss it as usual.
This change eliminates the flashing of old display content when
the screen is turned on. It also helps to conceal some of the
latency of turning the screen on. We always turn the screen on
immediately (even when screen on has nominally been blocked) and
rely on the black surface to hide the screen contents until the
last moment. Dismissing the black surface is practically
instantaneous compared to turning the screen on.
Bug: 7299370
Bug: 7139924
Change-Id: I57d13287acd05bd0a48811095bb02dc7bc7cbeb6
A couple problems:
- We need to clear app preferences later, now that we have encrypted apps.
- The multi-user implementation of this would allow different preferred
apps from different users to potentially interefere with each other.
They are not completely separate data structures.
Change-Id: Id4f1ebb6414fdf30ff1049adaa1efe83dabac01a
During app freezes resized windows were being dropped if the freeze
window timed out. This fix adds windows to the list of resized
windows but does not notify the clients of the resize until freezing
is completed.
Bug: 7094175 fixed.
Change-Id: Iee1f5f532a0e661fbf900e4540146ae4b645d68e
The new attribute allows an Activity such as the alarm to appear
on all users screens.
Bug: 7213805 fixed.
Change-Id: If7866b13d88c04af07debc69e0e875d0adc6050a
Defines the lower end of the allowable screen brightness range
as the lesser of the dim level, the user brightness setting
minimum value and the lowest auto-brightness level.
Bug: 7295909
Change-Id: I7a72b4611631f9e51578205ff12898c5bae02b1b
7296314 Crashing dreams are stuck
7296510 Transition from lock screen to dreaming is really bad
The window layer for dreams is now moved down below the keyguard,
so that some of the expected stuff like crash and ANR dialogs can
be seen on top of them. While doing this, I reorganized how we
define the layers so the constants are just in the switch statement,
so it is much less crazy-making trying to read how things go
together.
We now have some special cases for when a dream is being shown
to turn off its animation if the keyguard is currently shown.
Since we know it will be hiding the keyguard we need it to be
shown immediately so that you don't see whatever is behind it.
Cleaned up some handling of when the lock screen is displayed
while a FLAG_SHOW_WHEN_LOCKED window is displayed, so that the
lockscreen doesn't transiently get shown and mess up the fullscreen
or system UI state. This also fixes problems with any normal
activity that is doing this.
Hid the methods on DreamService for setting lights out mode. It
doesn't make sense to have such methods on DreamService, because
you can just as well do that on your own View that is showing the
dream content, and when you can do that you can fully participate
in the (required) interactions about it such as being told when
the mode goes away.
The DreamService method for going fullscreen now uses the window
flag for doing this, which is what you want, because you want this
state to persistent on that window and not get knocked out if
something above the window tickles the system UI state.
Also fixed the problem where dreams that hid the status bar would
have a jerky animation when going away, since they were causing the
activity behind them to be layed out without the lock screen. This
is a kind-of ugly special case in the window manager right now to
just not layout windows that are behind a dream. Good enough for MR1.
Change-Id: Ied2ab86ae068b1db0ff5973882f6d17b515edbcd
1. The reason is that the screen magnifier computes that the whole
screen is not magnifiable. The miscalculation was caused due to
an incorrect assumption that the non-magnified area is only at
the bottom. In fact, on a phone in landscape the non-magnified
area is both on the right and at the bottom. This change adds
a correct algorithm for computing the magnified region.
2. Increasing the delay for computing the magnified area when the
keyguard goes away to allow all windows hidden by the keyguard
to be shown. In rare occasions the previous delay was not long
enough resulting in a state where the keyboard is considered
a part of the magnified region.
3. Removed some dead code.
bug:7293097
Change-Id: Ic5ff91977df8bcf4afd77071685c3eb20555d4f3
This can happen due to a race when the input device is
opened or if the kernel happens to be missing the required
Android patches to set the timestamp correctly.
Bug: 7291243
Change-Id: If4319440eaff2889147c86296abd39efc5664346
1. The active window is the one the user is touching or the one
that has input focus. It has to be made current immediately
after the user has stopped touching the screen because if the
user types with the IME he should get a feedback for the
letter typed in the text view which is in the input focused
window. Note that we always deliver hover accessibility events
(they are a result of user touching the screen) so change of
the active window before all hover accessibility events from
the touched window are delivered is fine.
bug:7296890
Change-Id: I1ae87c8419e2f19bd8eb68de084c7117c66894bc
Bug: 7247911
Bug: 7294899
Also did some cleanup of unused code that resulted from refactoring
the app widget service. Fixed a few more ambiguous calls that weren't
using the correct user id.
Added some logging and improved the dump() formatting.
Change-Id: I27abb5c6341458e1e50a2cc9ab67e8de573ab283
The PowerManager may call into the BatteryService while
holding its locks. We need to be careful that the BatteryService
doesn't call into other services, particularly the ActivityManager
while holding its own locks.
Bug: 7298218
Change-Id: Ibf8ab13224f204a9857825265e864f93583bce8e
Created a new flag that indicates that a window should be shown
to all users. For the flag to be valid the owner of the window
must have system permissions.
Also separated system window types into those that show to all
users (e.g. StatusBar, Keyguard, ....) and those that appear only
to the owning users (e.g. Drag, ANR, TOAST, ...). Those that appear
only to their owner can override their default behavior using
the new flag (e.g. LowBattery).
Fixes bug 7211965.
Change-Id: I1fdca25d57b7b523f0c7f8bceb819af656c388d4
Fixed a race between the UiModeManagerService and PowerManagerService
both of which are trying to wake the device when docked / powered.
Bug: 7281240
Change-Id: Ia41fef48f17f2a2eb56549437d295f9a86c95af2
Reduce reliance on Service interface overrides, instead
steering clients to the DreamService-specific lifecycle
hooks:
onAttachedToWindow .. onDreamingStarted ..
onDreamingStopped .. onDetachedFromWindow
The old Dream.java is finally gone now too.
Bug: 7281802
Change-Id: Ib7802c3397fde60ad1132fa49831da182eef4d7a
Create a BLUETOOTH_STACK permission to share between all users of
Bluetooth apk. Bluetooth Manager binds to apk of current user
bug 6925422
Change-Id: I2df5546e977c2d77bc66d97d12479d933826922a
This fixes an issue where the device would not lock immediately
when the user quickly pressed power off / power on even if
configured to do so. We were suppressing the screen off
and wake up broadcast in this case.
Now we make sure to always send at least one broadcast to
indicate the transition. We still collapse back-to-back
full cycle transitions though so as not to end up enqueuing
useless broadcasts.
Bug: 7061116
Change-Id: I7211c5fd963c271c2b0aceb4d2f746063c629079
