For applications that remove the names of resources from the string
pool, check that the attribute has a name before attempting to print it.
Test: manual
Bug: 130553900
Change-Id: I05e5d59f01b2c02c8a024d06fd896074d6bf465b
is SYSTEM_UID and it happens to pass foregroundness check
at the time of sending
It could lead to abuses like directing the user to Settings app
and then giving a PI through any API, so that app could open
activity/trampoline due to Settings app (SYSTEM_UID) being fg.
PI-based starts where sender is SYSTEM_UID will no longer be
supported based on foregroundness check. Any such starts need
to now be explicitly whitelisted (see b/124858756).
Also, ignore windows of type TYPE_APPLICATION_STARTING in
isAnyNonToastWindowVisibleForUid() check.
Bug: 129563343
Test: atest WmTests:ActivityStarterTests
Test: atest BackgroundActivityLaunchTest
Test: atest CtsActivityManagerDeviceTestCases:ActivityStarterTests
Test: atest WmTests:RootWindowContainerTests
Change-Id: I563f71f0b1c7922d8c675d2e4feed909f43446be
Add copy text underneath image to make it clear what the button
means. Make text clickable. Align text colors with spec.
Bug: 130350131
Test: Manual inspection
Change-Id: Ia2916fd48ef8b2d1862d9baf5f9f695f54187d55
InsetsAnimationControlImplTest sets new insets mode to full mode when setUp(), which
leads some tests in ViewRootImplTest fail. Introduce InsetsModeSession
to fix it.
Change-Id: Id159db1395d1b105d9d4d3d13cdba778015e9989
Fix: 130609028
Test: atest ViewRootImplTest InsetsAnimationControlImplTest
In the process of revising libnativehelper interface, the
initialization of the fArray field was dropped. This change fixes
this and adds regression tests.
In writing the tests, it became apparent that the AutoBufferPointer
destructor would crash if the underlying buffer was heap array backed
and the buffer position was non-zero. ReleasePrimitiveArrayCritical
was called with the element corresponding to the current position
rather than the first element.
Bug: b/130390512
Test: atest FrameworksCoreTests:android.graphics.BitmapTest
(cherry picked from commit e522c99430)
Merged-In: I319512aaede7ba8af5d013c9281417695abf9099
Change-Id: I319512aaede7ba8af5d013c9281417695abf9099
In some cases, the supplier of ContentProviderOperation is okay
if certain operations fail, and they'd like ContentProviderResult
to tell them about the failures instead of aborting the remainder
of the transaction.
Start using this for ModernMediaScanner, where we probably raced
with someone when building an UPSERT-style operation. We'll
pick up any changes to those files during the next scan.
Bug: 128494336
Test: atest --test-mapping packages/providers/MediaProvider
Change-Id: Ida8230ff2bbb3bab56eb83928e49e7097bfbc9fd
