Change I11ea3fdf889689d76211506e81420e694238ee68 introduced
Binder#getCallingUidOrWtf.
It should have used Log.wtfStack instead of Log.wtf. The stack trace is
important for diagnosing the origin of a given binder call.
Bug: 252975769
Test: TH
Change-Id: I6946b1aea536f4b662ffffe8024dbc54d3dc36e0
This CL fixes a bug introduced by ag/19712818: when inflating the
parental control dialog, we should use the Quick Settings context (which
is always using dark mode) and not the application context. Otherwise,
if we are in light mode, then the color of the text is the same as the
background and it looks like there is not content in the dialog. In dark
mode, everything was looking fine because the text color from the
application context would be light.
Bug: 260108168
Test: Manual
Change-Id: Ibc9d5374d44955493a41b477482d5585546c7eb1
Currently with shell transition, isVisible() is true once visible
is requested for launching activity. Then the input sink surface
will be created when previous activity completes pause. But because
the launching activity is in transition and the transaction is sync
transaction, its surface operation is also deferred until the
remote applies the sync transaction. In other words, the creation
is too early and leads to lock contention such as blocking preparing
starting window.
With the sync state check, the timing of creating input sink surface
will be similar as legacy transition, which is after the transition
is ready, i.e. sync finished. That timing is usually less sensitive
(contention) to transition latency.
Bug: 260059642
Test: atest ActivityRecordInputSinkTests
Test: Transition latency of OpenAppMicrobenchmark
Change-Id: Id9719510c74296392e988abf8d48fe3af78026d3
When showWhenLocked activity is launched on a second display, it's
possible that a window container for a wallpaper on the default display
and the showWhenLocked activity on the second display are both collected
for app transition.
This is a workaround to prevent NPE for such cases.
Bug: 260997053
Test: atest MultiDisplayLockedKeyguardTests
Change-Id: I9dbd0928aa785a6d547b8e19841748280ba982c9
Include a Note Task shortcut to the Widget Picker with utilities to enable (show) or disable (hide). The Note Task shortcut opens in a Floating Window.
Test: atest SystemUITests:NoteTaskInitializerTest
Test: atest SystemUITests:NoteTaskControllerTest
Fixes: b/251207256
Change-Id: Ic434b2cd912c3568284c3daec6c338e45870c11f
java_sdk_libraries can be part of non-updatable modules. Those modules
might have a need to expose @TestApis that are only used in CTS tests.
This change adds an ability for them to do so by introducing
all-non-updatable-modules-test-stubs, and adding it to the
android_test_stubs_current.
Note that this change doesn't allow updatable modules to contribute to
test API stubs, since the list of the modules that can contribute to
test APIs stubs is hardcoded in the non_updatable_modules constant.
Bug: 261004711
Test: m
Change-Id: I9d17f49702bd64f5718b445f14b203c146bc6794
Fix: 258422365
Test: atest NotificationManagerServiceTest SnoozeHelperTest; Also tested the reproduction steps in b/258422365#comment6 manually to check that the issue doesn't happen anymore.
Change-Id: I5a2823f10053ea8c83c612a567d6d4f1b6af23e7
This attaches the black scrim in the background. All changes to that view have to go though the background handler.
Bug: 254924424
Test: current unfold tests should work unchanged && manually verified things are as expected with a perfetto trace
Change-Id: I224d9084b6f20230810faa6cc6d6b8bd6b46f518
With new ConstraintLayout versions layout_constrainedWidth was causing the view to assume match_parent width.
As MotionScene supports importing external layouts, we're not loading them at runtime anymore.
Also, the transition for large_screen_header_constraint is not used anymore, as initial and final constraint were the same.
Bug: 259933125
Test: CombinedShadeHeaderConstraintsTest and manual
Change-Id: I3961d9f4311c96b5d0663e1700fca78be5e6544b
This test was part of packages/modules/Connectivity but tests a
platform-specific service. Move to the framework tests directory.
Use the AndroidJUnit4 runner.
Bug: 259341026
Test: atest FrameworksServicesTests
Change-Id: I70ba589ac460c769b4890b4e7a05cf02f54f80fb
For reduce jank or flicker when resize ActivityEmbedding in split
case, we should apply shell transition to resize case too.
Fix: 247678744
Test: manual
Test: pass existing tests
Change-Id: I8a1aa500d5928284ca7955042dffb617be15635a
When launching an activity into a new TaskFragment with a different
split layout from the existing split, there can be many situations, such
as different split ratio or layout direction.
Use a jump cut for now since it is not a common use case, and it is hard
to cover all situations.
Bug: 241043533
Fix: 258126611
Test: verify with opening new activity to new split layout.
Change-Id: I6c96518b282567339fb890e7f4af7d9d82c98418
The wake up transition should be used to waiting for ShowWhenLocked
transition, but we cannot know whether there have an activity which
requested device wake up, it can be also be a normal screen on.
For those keyguard-related transition, they can be request in
KeyguardController while updateVisibility, and collect participated
activity at the same time, so the transition won't happen too early.
The empty wake transition can affect tests which use injectInputEvent
while waiting for animation (WMS#waitForAnimationsToComplete), which
lost those input event.
Bug: 259231431
Test: run atest KeyguardTests KeyguardLockedTests KeyguardInputTests
w/o AOD enabled.
Test: atest LockscreenWithInvalidPassword
Change-Id: Ib2db22bb5fe92f79dbd14948dae8d8a2b538f605
Moved PSBV.touchHandler out of NPVC and
into PhoneStatusBarViewController.
Bug: 249275160
Test: existing tests
Change-Id: Id71bb9fab1d889cd1c193595b233cccd9c3f37ef
