- Don't reply from the unregistered address.
- Use "unrecognized opcode" as the default reason.
Bug: 16799466, Bug: 16798785
Change-Id: I7c2ece6436f7ebd59986d2baf4f45cd86e6622d9
1. Apply masking to feature abort of SetXXXTimer
2. Announce error when failed to send SetXXXTimer.
3. Remove TimerRecordingAction when deivce is disabled.
4. Report clear timer record error thru onClearTimerRecordingResult.
Change-Id: Id243aa25481dad57509b4229ca214fa9b6b00fdc
Loosens the restriction for reporting an issue with an app from "same
profile" to "same user".
Fixes bug 16792315.
Change-Id: I25100a3f7030b2a74e9fe40894103677e8db6cdd
When the maximum number of retries that has been exceeded is not for the primary profile of the user, wipe only the profile that set that policy (e.g. the managed profile) rather than the entire user. At the moment the whole device is wiped if the max number of incorrect passwords for a managed profile is reached, as the password is shared with the USER_OWNER.
Bug: 14453697
Change-Id: I5746de104133c0ea0a51d75b9c92e1516d365d8c
Use 400dp as the visual width of the shade across all devices
and orientations, except small device portrait.
Since the width is now dynamic across orientations, update the
necessary views in SystemUI on rotate.
Remove obsolete styles + resources.
Set max QS columns to 3 on all devices, due to the newer shade
width.
Bug:16825515
Change-Id: Iff171549a79bc318a4d585a2d9ea2af9bc686922
Now remote views contain the application info for the app on whose
behalf they should show content. The application info contains the
user id and the package name which makes it unique in the multi
user environment. If the app which creates the remote views does
not have the same package as the one passed in the contructor of
the remote views, we lookup the app info for the target package.
There is a special hidden overload that allows creating a remote
views instance that represents an app in given package and user
(requires interact across users) permission. The latter is used
by SystemUi when creating remote views for apps that posted a
notification.
bug:16838651
Change-Id: I2daa974629f89a2446848e59300d994dadcf4439
This change introduces a class that handles launching the secure camera even
when other apps may be using the camera hardware. It does this by sending out
a broadcast telling apps to stop using the camera as soon as the user begins to
swipe to use the secure camera. It then waits until either the camera is
available or a timeout period has expired before it fires off the intent to
start the secure camera.
Secure camera apps can indicate (via their meta-data) that they will instead
handle the responsibility of waiting for the camera to be available. In this
case, the new class still handles sending out the broadcast telling other apps
to close the camera, but it fires off the intent to start the secure camera
before the camera hardware is available. This is optimal because some of the
time spent closing the camera can happen in parallel with the time spent
starting the secure camera app. Since we can't depend on third-party apps to
wait for camera availability, the default is for us to wait for the camera to
be available before firing off the intent. However, our first party app is
able to take the optimized approach.
Change-Id: I1eaed94875e8f50018cffab25aa1e9b21d85f25b
Since all service-internal checks use stream alias, and the alias
for NOTIFICATION is RING, even on tablets.
This is a followup to ag/506532.
Ensure that #getMasterStreamType always follows the stream_system
alias, and update the system alias for tablets per the new policy.
Update the associate javadoc comment in AudioManager to make this
clear.
Note: there is now no difference in the alias maps between phones
and tablets.
Also fix a condition affected by this in VolumePanel.
Bug:16626879
Change-Id: I2700a48d2a3e7703607f771a1ab7f325596fd789
Putting the deprecated at the beginning of a class comment seems
to cause problems with JavaDoc so I've moved it to the end.
Bug: 16799480
Change-Id: Ia23dab2ad8ceb8e107e047480b5cb324d11a1457
This changes makes the initial, default language set and
the #setLanguage call with the default language as input
exactly the same.
Previously, the requests made after the initial default language
set were missing the name of the default voice.
+ Some tests clean-up. Some of them fail due to issues with
Locale#toLanguageTag upper-casing the variant field.
Change-Id: I5470617007fe45462b7198bf62a03eb5fe47a590
