Problem:
The context from getPackageContext() does not have the theme set.
Solution:
Use the SystemUI theme to inflate the icon.
And we probably want to have a unified theme for all action chips anyway.
Test: Send myself a message "Code 12345", Verify that the
icon works in both light and dark theme.
Test: Send myself a message "www.android.com", verify that the
app icon is shown properly in both light and dark theme
Fixes: 142972040
Merged-In: I60f5fbea802fe1fdf498e93dfa4fdcebf38af78c
Change-Id: I60f5fbea802fe1fdf498e93dfa4fdcebf38af78c
When we received the wallpaper service disconnected, we could not
immediately know the reason for the disconnection, so if the service
was disconnected because the package was updating, and we received
the update broadcast after 1 second, in original design we will only
revert to the built-in wallpaper. However, it will be too late to bind
the original service after we receiving the package update broadcast.
In this CL, if currently wallpaper service is accidentally disconnected,
we will try to re-bind it for 10 seconds.
Fix: 138973259
Test: Manual install wallpaper service and delay the package update
broadcast on purpose, verify the service can be bind back.
Test: atest WallpaperManagerTest
Change-Id: I11891098164372002b36c558f670621d8c852af8
(cherry picked from commit b87530296b)
This change adds null checks before accessing outInfo in
deleteSystemPackageLI.
Bug: 142083996
Test: manual; remove static dependency on eng build and reboot
Change-Id: If0fd48343e89cbb77ccd25826656194195d5b0cd
(cherry picked from commit 1747101650)
If volume dialog is not expanded, then we just replace the old default
CarVolumeItem with the new group.
Bug: 141752233
Fixes: 141752233
Test: Manual
Change-Id: Idfa6ad3a1129e069a297f1c1cfd47b40a61773cc
Currently when application calls ImageWriter.newInstance(Surface, int,
int), the default format of UNKNOWN is overridden to surface format in
Java side before being overridden again to user specified format (from
RGBA to IMPLEMENTATION_DEFINED). This breaks backward compatibility with
certain gralloc implementations.
Do not override UNKNOWN format in Java side.
Test: testWriterFormatOverride
Bug: 143637345
Change-Id: Ie61157bca11f623eff2b8c0ed2a988c86da07cb2
In #clearValuesAndCheck, values are placed in the value map based on a
key that depends on the current user. When these values are checked
in #call, it uses default user 0 (unless CALL_METHOD_USER_KEY is
populated).
Considering we don't even use getStringForUser when checking the values,
it seems to make sense to check against the userId used to set the
values in #clearValuesAndCheck.
Bug: 143529292
Bug: 143480221
Test: atest VolumeDialogControllerImplTest before and after the patch
Merged-In: I0ee9313b1394515d9405344482dc1779d15c0819
Change-Id: I0ee9313b1394515d9405344482dc1779d15c0819
to be configurable.
Bug: 143372960
Test: manual - took before and after videos
Merged-In: Ia0b6a56859889d538e21f807a88bc5217d89e64a
Change-Id: Ia0b6a56859889d538e21f807a88bc5217d89e64a
GPS is normally disabled when the device is Dozing. This means that GPS
is turned off in Battery Saver when quick doze is enabled. This
interferes with navigation and possibly other use cases. This changes
the behavior so that GPS is only disabled during Dozing if the device is
stationary. This will not interfere with regular Doze since the device
has to be stationary before it will enter regular Doze, and also allows
GPS to continue to work when Quick Doze activates while the user is
still carrying the device around.
Bug: 140162457
Test: atest BatterySaverLocationTest
Test: atest com.android.server.DeviceIdleControllerTest
Test: [manual] Shorten motion timeout, step into idle, check gps
provider dump after short timeout to check mDisableGpsForPowerManager=true
Merged-In: I7bc68c8a5e6744aac3a160ac14fc3375f9374d82
Change-Id: I7bc68c8a5e6744aac3a160ac14fc3375f9374d82
(cherry picked from commit 810c77d5a0)
Also fixed shell cmd so it does not require name for pre-created users
Bug: 111451156
Bug: 132111956
Test: adb shell pm create-user --pre-create-only --guest # couple times
Change-Id: Ie256d55bdfe685f040d16223fcc70f2f2b7969da
(cherry picked from commit e35f5b01f8)
This change is to avoid issues in assigning signed integer
feature (WIFI_FEATURE_DPP) to long featureSet. The
assignement was resulting in large, signed integer.
Bug: 143363031
Test: atest com.android.wifi.server
Test: Wifi ON/OFF, Connect and data transfer
Test: Added logs and verified feature set saved in iface.
Change-Id: Ic3828e675e49a8c1628368a7e6cf7f01e420bd96
(cherry picked from commit 26ac48bcd7)
- createCar(context) will lead into crash when car service is crashed.
- Migrate all createCar to new style which can handle car service crash
while removing unnecessary delay for connect.
Bug: 142543455
Test: kill car service and check if systemui survives
Change-Id: I69298f763a2a2e63c0cf8deae13f747ea19bbf9d
Merged-In: I69298f763a2a2e63c0cf8deae13f747ea19bbf9d
(cherry picked from commit c83d7b93ae0620bac34af2a937d5855b05058213)
No change to logic, only docs.
This clarifies the docs for onShowCustomView. This @links to
FLAG_FULLSCREEN, reminds the developer they must override both
onShowCustomView and onHideCustomView, and provides guidance for
CustomViewCallback.
Bug: 143247282
Test: make -j4 docs
Change-Id: I64de3723674da5c138438921cc8232c4bf2a3d98
In cases where the source of the CellInfo fails
but does not generate a Throwable, the detail info
will not be provided and should be null. Inside the
user process, we need to check for null before
dereferencing the detail object. Oops. :-(
Bug: 142706765
Bug: 142421696
Test: manual
Merged-In: I1307d1d7aacef0bbd33bfc73700ee774e12e64c1
Change-Id: I1307d1d7aacef0bbd33bfc73700ee774e12e64c1
(cherry picked from commit fff052067c)
This reverts commit 9449cfc4a6.
This reverts commit 22fa97577f.
This reverts commit bde48202e7.
Bug: 141649119
Bug: 143195895 is also fixed on my taimen with the above.
Bug: 143185828
Test: atest SystemUITests
Change-Id: I225b10fef2f88d3436ef3a683c09717467b071ad
