Pointed out by Wei: a later setAudioSessionId() could override the
previous one whose task has not completed.
Bug: 112549970
Test: MediaPlayer2Test
Change-Id: Ib5862d6dfd1ea4ddcbccb421906510ab4a3bae87
While we are initializing the policy file from disk, if there is no
such file we still have to allow the notifications to be shown.
Previously we were assuming that it's false if there's no such tag
in the xml.
Test: existing tests pass
Change-Id: Ia76a8f3e0af1b176de6141bec1db4babd4f97e00
Fixes: 119833914
Commit d8bf22e8c6a added support to the package manager shell command
for the new --apex flag but did not mention it in the usage text.
Test: adb shell cmd package | grep -e '--apex'
Change-Id: I009bbc979b3a0729fe567f0ef74db18e087d14f5
This change adds RoleManager.setRoleNamesFromController() to allow
RoleControllerService to initialize the roles for a user. This change
also fixes the persistence of roles by calling writeAsyncLocked().
Bug: 110557011
Test: build
Change-Id: I921b6aa691478ca4c0dd1a75fc929a96ce1e7df5
See build/soong/README.md for more information.
Bug: 119809391
Test: atest FrameworksServicesTests FrameworksMockingServicesTests FrameworksUiServicesTests (mostly passes, those that don't aren't run in APCT)
Change-Id: Id28d25cc36df47e2f533ae878fb0d2745164fbfb
We still had some legacy code which would make sure that the alpha of
dragged views would remain the same, but that isn't just unnecessary
but could lead to bugs where the view itself was stuck with alpha 0.
Test: runtest systemui
Change-Id: I36768bebcb2831aa422bafe7b18c92a4834e7745
Fixes: 80525283
This change modifies the logic in
NotificationStackScrollLayout.updateFirstAndLastBackgroundViews() which
determines if the first or last visible notification row has changed.
The "old" values were being read after the changes were applied, meaning
that the change was never detected, so an animation was never started.
Test: manually
Change-Id: I29869f5b0f78fd81a17d2c3dbe18c067f94fdb5e
Fixes: 119792278
Test: Manually on Thermal HAL 2.0 device
Test: Manually on Thermal HAL 1.1 device
Test: Manually on no Thermal HAL emulator
Test: atest $ANDROID_BUILD_TOP/frameworks/base/services/tests/servicestests/src/com/android/server/power/ThermalManagerServiceTest.java
Bug: 111086696
Bug: 119413961
Change-Id: I6723406123d12339e82e9e87eec14b7f9a301897
It wasn't used anymore. Additionally are we also removing the
drag animation events that were originally created for the shadow
alpha animations.
Test: runtest systemui
Change-Id: Ia5cbf7ecb2728fc1298c1b22f3f37bac6b531803
Fixes: 119791828
When the binder dies, it makes no sense to send an error since the client
is already dead. Instead, just notify the driver to cancel the current
operation.
Bug: 119592051
Test: manual
Change-Id: I60e515bfff933df5e672ac2797d7d8620e6ece09
Previously, all transferred alerts were removed when the inflation was
aborted. This could cancel transfers when the child was simply updating.
Now, this only occurs on remove. On update, the previous heads up/
ambient inflation flag is carried over and the helper determines if
the content should actually heads up after inflation.
Bug: 111809944
Test: runtest systemui, manual
Change-Id: Ic7c5414849e2bfa1978bb1c57ed314c4997b8ea2
This CL fixes lazy content inflation for alerting notifications in
groups by inflating them on the fly if we would transfer to a
notification that does not have its content inflated. We introduce a
helper class here to explicitly deal with the transfer logic,
refactoring a lot of the GroupManager code out.
This removes the previous workaround to always inflate heads up +
ambient views.
Bug: 111809944
Fixes: 111809944
Fixes: 117933032
Fixes: 117894786
Test: runtest systemui
Test: manual, posted groups that trigger a transfer and a transfer back.
Group alerts behaved as expected.
Change-Id: I9b5ec4c8bdeea20707874d90213dcd1d22d8b503
- Proxy requested assist data from swipe up through the the
IntelligenceServiceManager to AiAi for each activity that we receive
assist data for (the AssistDataReceiver to be removed later once AiAi
implements the new IntelligenceService interface).
Bug: 117268952
Test: Build dummy intelligence service, ensure that we get capture event
on swipe up.
Change-Id: Iec29792932d30e94a702bd5079711c6615d0738a
Moves some windowing-mode change handling into config update
First, this moves the initiation of mode configuration-changes into
the ATM hierarchy (from WM). It restructures some more of the bounds
calculations in TaskStack so that they can be used via the controller
and with more data coming from a parent configuration instead of
hard-coded to the DisplayContent. These interfaces are transient as we
transition more towards a "unified" hierarchy set-up.
The main part of this change is in ActivityStack.onConfigurationChanged
where we actually process which changes effect the different modes
we are in. For example, if rotation/density/screen-size change, we
need to recalculate split-mode bounds to match. Along with moving work
to onParentChanged, this should actually handle more display changes
than before (ie. it should be possible to reparent a split stack into
another display and have it resize appropriately).
Bug: 113900640
Test: go/wm-smoke
Change-Id: Ifd4def135b62a54259dd0f9f030e839c29bebebf
