Re-initiate the DismissCircleView upon theme change to make sure the PiP
dismiss target inherits the right theme color.
Bug: 199359063
Video: http://recall/-/aaaaaabFQoRHlzixHdtY/bIMe3fAFoW6Nw47Ye894MZ
Test: manually, see video
Change-Id: Ia65c633fd636114ddbdc2ff497245e2cb73f7745
Root Cause: ACCESSIBILITY_BUTTON_MODE only be updated in navigationbar, did not be updated in new-design taskbar.
Solution: Move updated function up to NavigationBarController, which
controlls NavigationBar and Taskbar.
Bug: 203173063
Test: manual test
Change-Id: Ib18fcd7062bf4e7627f4ffaeefd150c3d1055a47
Fix issue where attempts to render overscroll content
into the alpha 8 mask would crash if there are WebView
related Functors in the scene.
Provide implementations for getTypeName in SkDrawable instances
of FunctorDrawable.
Fixes: 203960959
Test: manual
Change-Id: Idf321d9bfcbe2b8ba1eb205eadcec95e55865305
If the unlock animation should run, we should run it to make
sure that the device's state actually updates to KEYGUARD
when running the unlock animation.
Test: manual
Bug: 205587594
Change-Id: I50a948f3bb199f52d81dbd6066a603bea67a16bc
If a profile owner is defined for a specific user, do not delete usage
stats for a package on package deletion.
Bug: 197399948
Test: atest UserUsageStatsServiceTest
Test: atest UsageStatsTest [all]
Change-Id: I94a8e3dfca8ef4c7616f77944d61726e06043b85
Merged-In: I94a8e3dfca8ef4c7616f77944d61726e06043b85
- Update lightRevealAmount to the dozing amount immediately
when the lightRevealEffect is set in StatusBar (since
the PowerReveal and LiftReveal rely on dozing amount,
while the CircleReveal relies on timing)
- After the CircleReveal finishes, set the light reveal back
to the default LiftReveal
- Dump the light reveal effect to dumpsys for easier debugging
Test: manual (force AuthRippleController's KeyguardUpdateMonitorCallback
to be called before the BiometricUnlockController's callback)
Fixes: 201605121
Change-Id: Iedd31ec5aa75bda6998a0c942fd5a5eab953e74c
In case there are delays in onEnrollmentChanged, let's make
sure that KeyguardUpdateMonitor uses the correct logic for
whether to run UDFPS or not.
Test: atest LockIconViewControllerTest
Bug: 205670174
Change-Id: Ia2b4fd290b4d78819fabf7bea744faed62fa0d3b
The meaningful contents of FeatureFlags only relates to debug
builds and has been moved into the debug version of
FeatureFlagManager.
Bug: 203548872
Test: atest SystemUITests && manual
Change-Id: Ifcef0cfe4e2d07daec34dfc7a6cbe2d305b312b5
The case is like: (A,B,C are activities from top to bottom)
A orientation=unspecified, dialog style(!fillsParent)
B orientation=landscape, entering pip mode
C orientation=portrait, the activity that decides screen orientation
The non-opaque activity won't decide screen orientation, that causes
to miss the proper timing to trigger fixed rotation when the previous
top is switching from fullscreen to pip with screen orientation change.
This makes the callback order of PipTaskOrganizer still follows first
onFixedRotationStarted and then onTaskAppeared. So the PiP won't stay
in a stale state with wrong bounds or alpha.
Bug: 204844178
Test: 1. Launch Chrome and play video in fullscreen landscape.
2. Enable auto-rotation, put device in landscape.
3. adb shell am start -a android.intent.action.VIEW
Chrome should enter PiP correctly.
Change-Id: I34250bb879898059568bc3b3b6b4da88fea75f57
The window type of the Taskbar is navigation_bar_paenel, which
is unmagnifiable. However, we used to exclude these windows from
magnification region computation to avoid cutout.
When the foldables is unfolded, the taskbar wouldn't cause the cutout.
To fix this problem, so we add a flag to distinguish the windows
that cause the cutout.
Bug: 196510717
Test: manual test
atest AccessibilityManagerServiceTest
Change-Id: I882bf48585f646cee5827044c88da37250041c0f
Merged-In: I882bf48585f646cee5827044c88da37250041c0f
(cherry picked from commit dc9b90f46a)
