The public API for getting at KeyphraseSoundModels is based on the user,
locale, and keyphrase ID. However, the sound model database would allow
multiple models to be registered using the same values, as it has a
primary key on the model UUID. This can potentially lead to a scenario
where multiple models are enrolled for the same user, and the data that
is returned when querying models is arbitrary.
V6 of the sound model database now enforces a primary key over the user
ID, locale and keyphrase ID. When inserting a new model with the same
primary key, the old model will be removed and the new model will be
used instead.
When upgrading from V5 to V6, if there are any models that would violate
the primary key constraint, they are all dropped.
Bug: 32174118
Change-Id: I9597c6a994f01a426625c3be2c53e826f26f5156
Currently, if the "ro.boot.vendor.overlay.subdir" property is set, we
only search in the subdir. It seems more powerful to search both in
there and in the standard /vendor/overlay directory.
Also, RRO packages have the "priority" attribute in the <overlay>
attribute in AndroidManifest.xml which serves for disambiguation.
Bug: 32202741
Change-Id: I9ef44b8ca4ac0e92ba02e25b3672c6b274fe4639
Calculate the alpha value for each pixel in the display mask based on
the area of overlap between the circle and that pixel.
This works around a bug in Skia that is fixed in master.
Change-Id: Ieb8df0d6f46f07cff1dbc368d74938ec699e08cf
Resumed and focused activity should be adjusted simultaneously.
This was missing when task was moved back and focus was adjusted only
after previous activity stopped.
Bug: 31983988
Change-Id: Ie6c908b0c8c1adf94680b42ca5a8c9b01308e361
The code detecting whether we should rotate seamlessly (which for
example we do not support for 180 degree rotations) was moved to before
mRotation was updated, so it was always passing the same value for old
and new rotation, and we were failing to perform the 180 degree rotation
check. Correct this.
Test: Rotate the device 180 degrees at once while in Camera. Things look ok.
Bug: 32025371
Change-Id: I0ef08e8cc4e5d87fdff6d839381196c2b52e71d5
If a device isn't voice capable, it shouldn't show the emergency
option.
Test: use tablet which isn't voice capable, insert indian sim and
make sure no emergency option is displayed in global actions.
Bug: 31953703
Change-Id: I351e87320f3ffec76d1c1fc5aac78e5c48c0ac54
The admin can instead use the value of 0 to reset to default.
Test: runtest --path frameworks/base/services/tests/servicestests/src/com/android/server/devicepolicy/DevicePolicyManagerTest.java
Bug: 31430135
Change-Id: I0d6b29ca4eca65d7ca72a8975a0c28c9050a946c
(cherry picked from commit 943aabd11c)
All of the instances of calling into other services with locks held
should be resolved, so this isn't needed anymore.
Test: have 2 rules end or start at the same time; ensure the resulting
state is correct.
Fixes: 28459870
Bug: 31980467
Change-Id: I3f12f1d3b8460361dab6fa9468f0ede55ca8dd5e
(cherry picked from commit 89aeab0e71)
The scheduling mechanism was properly checking for it, but the raw
data-handling engine wasn't. Now it is.
Bug 32054889
Change-Id: Id57ea12be8e2674c4f678f058278b8c14865b6fa
Internal implementation of "adb install" has been changed to
session install. As a result '-s' option doesn't work.
Size info is required when using '-s' option in session install.
PackageHelper.fitsOnExternal() returns false and moving app to SD is
failed when size info is not set.
So set size info when '-s' option is used.
And change identity on some method because shell doesn't have some
permission to access to asec.
Bug: 29932779
Author: Ryuki Nakamura <ryuki.x.nakamura@sonymobile.com>
Change-Id: I65edcc0f01ceecaaba17792ad329791a9253d603
When the playback device is asked to go to Standby state with <Standby>
message, it needs to check whether it can accept the request or
not. This check should be done before asking the system for sleep mode.
Additionally, this change will fix the behavior after screen
timeout. After the timeout, playback device may go into sleep mode, but
canGoToStandby() check will prevent the device from tracking the right
current power status value. So need to set the status properly, and keep
HdmiCecLocalDevices enabled for accepting future CEC commands.
Bug: 30182748
Test: Check fugu screen after turning off/on Samsung TV and
Archer. Also tested the case of screen timeout.
Change-Id: Idbcd7abd5d4e22942845760f77f7f7e622b987c0
Because of course during system shutdown, the wallpaper service
won't actually get relaunched. Sometimes shutdown can take long
enough that this timeout kicks, so we need to avoid clearing the
live wallpaper state spuriously.
For simplicity we just check at "would have timed out now" time
rather than try to distinguish between the shutdown case and a
genuine crash that raced with the shutdown broadcast.
Bug 32020355
Change-Id: I9335b2c0214b4c750ef950fed157d186aa670176
Only show dismissing window if we are about to unlock, and not just
in all cases. For that, set a flag to true during the process, and
reset the flag after the Keyguard has unlocked.
Test: Have a FLAG_SHOW_WHEN_LOCKED activity that starts a
FLAG_DISMISS_KEYGUARD_ACTIVITY. Check whether window is hidden
when toggling the power button. Make sure that the device is
in a trusted state during the whole test.
Change-Id: I9346dd869d1cf50a42c64a5f7c69f8e24f1e1f9b
Fixes: 31810884
