We've found local evidence that on fast hardware that the
MediaProvider.updateVolumes() method can be invoked while the
device is still unlocking. This can cause a race condition due to
the "reportUnmounted" logic in StorageManagerService.getVolumeList()
which attempts to hide all volumes from apps until after the user
has been unlocked.
Since MediaProvider itself is the one hosting the FUSE daemon, it
always needs to know the real underlying state of volumes, so this
change adds special treatment to never augment the mount state
when the caller is the FUSE daemon.
Bug: 159189865
Test: manual
Change-Id: Id86e86aa942a697877860961d20ff0d8458dbc10
Immediately on boot (before a user is unlocked), PackageManager only
returns components that are direct boot aware. If we don't specify that
direct boot unaware components be returned as well, then ASC will think
all system apps are headless.
Also switching the cache from an ArrayMap to an ArraySet because there's
no need for the boolean value.
Bug: 155761007
Test: manually verify that apps with launcher activities aren't in the headless app list
Test: atest CtsUsageStatsTestCases:UsageStatsTest
Test: atest FrameworksServicesTests:AppIdleHistoryTests
Test: atest FrameworksServicesTests:AppStandbyControllerTests
Change-Id: Ieec648226d0d5a32197647af286c2cd4c167dd8b
- The old calls appear to account for the display cutout now which causes
the landscape/seascape comparison to fail
Bug: 157878422
Test: atest TaskSnapshotControllerTest
Test: TaskSnapshotSurfaceTest
Test: DecorViewTest
Test: Open an app with displaycutout layout shortEdge or always, and
rotate to landscape in non-gestural mode, and ensure that the
divider and background are drawn in the right place
Change-Id: Ib51964a76e25f902858aad4e8aeb810dfca2edc0
Merged-In: Ib51964a76e25f902858aad4e8aeb810dfca2edc0
To do this reliably, this CL changes the motion helper logic such that all dragging and animating changes the 'temporary bounds', which are applied to PIP via userResize. When a drag/animation ends, we schedule a finishResizePip to 'actually' move PIP to its final location.
This also fixes an issue where the resize gesture could steal events during a drag.
Bug: 158247953
Fixes: 157264677
Test: manual
Change-Id: I0129a4c77d06fb9cd186eacd4a35975f54432e9d
This is for migrating keystore logging to use statsd.
Test: Adding tests for logging is yet to be decided.
Change-Id: I66fbc2b8ed98a34f243387f8550b6523b13cb9a2
Bug: 157664923
I accidentally omitted logging for the event of banning/unbanning
notifications for the entire app, when I added notification
channel ban logging in ag/10450892.
Change-Id: Ibf2be1367bf41ad2b431cfcd2fef820ff270fed9
Test: atest PreferencesHelperTest
Bug: 156398451
On Android different interfaces usually use different routing tables.
As a result, a change in interface should not be treated as route
update, but rather a remove and an add.
This change fixes a bug in VPN seamless handover where routes
failed to be updated when a new tunnel interface replaces the existing
one within the same network.
Bug: 158696878
Test: atest com.android.cts.net.HostsideVpnTests
Test: atest NetworkStackTests
Test: atest CtsNetTestCases
Test: atest FrameworksNetTests
Original-Change: https://android-review.googlesource.com/1331916
Merged-In: I57987233d42a0253eaee2e1ca5f28728c2354620
Change-Id: I57987233d42a0253eaee2e1ca5f28728c2354620
Before we would only check the active tab for available
targets and finish the activity if there are none.
However, with the work tab, there can be cases where
the active tab has no targets, but the inactive tab does.
Fixes: 159202588
Test: manual
Change-Id: Ie03ebd1b022362681df432ec29299df86dceb490
