Method handlePackagesChanged modifies the policy data without holding the
lock. It also calls methods which expect a lock to be held -
validatePasswordOwnerLocked, syncDeviceCapabilitiesLocked, saveSettingsLocked.
Bug: 18262633
Change-Id: Ia4bdc6edc66bccf8a4288e048171ff20858ae2f6
This helps us in majority of the scenarios where the sound model doesn't
change across start/stop calls.
Bug: 17954633
Change-Id: Ibff817bb69bc69d2bb3a2603460fed596688b892
The current implementation uses a whitelist of package names. Use a
system|signature permission instead of rolling our own security and
add that permission to the existing set of whitelisted packages
(SystemUI and VpnDialogs).
In addition to being less of a security risk (using well-known methods
like Context.enforceCallingPermission rather than manually querying
PackageManager and checking UIDs for package names), this enables
other system-privileged apps to control VPN as needed per the below
bug.
Bug: 18327583
Change-Id: I38617965c40d62cf1ac28e3cb382c0877fb1275d
We allow TYPE_INPUT_METHOD windows to show on the lock screen.
These windows can attached other types of windows (For this
case the Swype KB was attaching a PopupWinow which will be
of TYPE_APPLICATION). This causes the popup window app token
to be added to the list of apps to be hidden on the lock
screen, thereby preventing the lock screen from been hidden,
which then causes a layout cycle(s). Now, we remove app
tokens for non-app windows from the hidden list in case it
was added by any attached widow they migth have.
Also, when we are updating window animations, set the
hideWhenLocked flag for windows that are IME tragets so
they don't go through unneeded cycles of having their
visibility policy set to hide/show/hide/show/...
Bug: 18021493
Change-Id: I3680256d41793f62def42fda00e26db1dcc990cc
- Add a timeout so if WindowManager "forgets" to tell that the
activity has drawn, we still unlock after 3 seconds, so the user
is not completely stuck.
- Use the screen height instead of the window height for the
translation animation.
- Don't run the animation if the attached window is not null. The
animation from the attached window will influence the transformation
as well, so there is no need to run an additional animation in this
case (apps with SurfaceView's had broken unlock transitions because
of this).
- If the starting window needs to go away while the unlock transition
is running, modify the existing animation such that it fades out in
the same transition.
Bug: 15991916
Change-Id: Ia5dfa31e1bc0d5745fe228e1daf08e268733b6f1
In SysUI, make sure not to dismiss Keyguard multiple times when just
waiting for a deferred dismissal, so WindowManager doesn't get
multiple calls to keyguardGoingAway.
Change heuristics how notifying Keyguard about activity drawn works.
Always notify Keyguard after executing an app transition, and notify
it also when not doing a transition after a startActivity call.
For that to work, update AppWindowToken.startingDisplayed also when
the window is displayed, but force hidden because of Keyguard.
Further, handle the case correctly when a window gets added during
the Keyguard exit animation by overriding the start time for the
animation of that new window. Also don't apply a transition animation
for a window when executing keyguard exit animation, so by removing
a starting window we don't break this animation.
Last but not least, tell Keyguard to start exiting immediately if
animations for exiting are disabled, like when going to phone/camera
on lockscreen. Before, we always had a delay of 1 second because we
waited for the timeout.
Bug: 1599196
Bug: 18272544
Change-Id: I596b2489f814b934abd256e16079d3d3f326e209
It is disabled dead code already and not useful anymore
with the new caching in LockSettingsService.
Bug: 18163444
Change-Id: Icc184e923e0fbeab31ed128336c01f835b24c6f2
Also fixes a potential issue where refreshing agents
for a user that no longer exists would result in a crash.
Bug: 18318629
Change-Id: I3589ea7e0f2e63fca02daeecf3ca964a8a8e4b3b
This makes the MediaSessionService check if the suggested stream is active
when dispatching a volume event and prefer it to a remote session if it is.
This mostly affects apps that use Activity.setVolumeControlStream.
bug:18318480
Change-Id: Ib8825394bf7a223fbd92bcb2c3ecb7c596ce7c29
Error dialogs absorb all input to ensure that they are not missed.
This can cause the screen to lock up if they are not displayed but
are still absorbing touches. This was what was happening when there
was an error dialog up at the same time as a phone call came in as
in b/17648830.
This fix recognizes when an app is dismissing the keyguard and
forces any error dialogs to be shown over such an app.
This also removes the private flags from the input system as they
are no longer needed.
Fixes bug 17648830.
Change-Id: I5c98b8265a1448b445fdb2f745fc78892f8656a4
This allows work profile MDM to enable unknown sources
even if the user doesn't have UI for it. Installing an
app from an unknown source will still prompt the user
with the package installer dialog, so it's not like the
MDM can now quietly install apps from non-market sources.
Bug: 18316350
Change-Id: Ia8f4fe36f12a258aa888e085acc0b358925f4817
Introduced a buffer for messages to be kept for later processing
when they arrive before the service is ready to handle them.
<Active Source>/<Initiate ARC> are handled in this CL to address
the reported bug.
Bug: 18277839
Change-Id: I1e73dd83dc9dd4a822fbcf4fda4c6221f09631cb
It may take an attach to move a task into the recents list. If the
timing is right a task may not be in recents by the time we are
removing it from the stacks due to a finish. In that case we should
look in the stacks for the task as well as looking in recents.
Fixes bug 18017409.
Change-Id: Idcfe2e263c9d0fe9a063fdf22515ac4e7fe89ecb
installd can then clear the ".booting" marker from the dalvik-cache
(owned by root). This marker is used to detect boot loops.
bug: 18280671
Change-Id: I878f1463c7f523892605c17b980a51ac3b6645e2
Currently Nat464Xlat reads the clat IPv4 address and updates the
clat LinkProperties when the interface is created. This causes a
race condition: because clatd only sets the IPv4 address after
creating the interface, it's possible that Nat464Xlat will read
the address before clatd has set it, causing the framework to
think that the clat IPv4 address is 0.0.0.0/32.
This seems to be happening more frequently now, perhaps because
clatd takes a bit longer to configure the IPv4 address now that
it needs to check that the address is free before using it.
Fix this by making Nat464Xlat listen for the interface coming up
instead of listening for the interface being added.
Bug: 12111730
Change-Id: Ic1c59b5b6dbb851b7431d1b06885f67803373bb9
This expands the use of EXTRA_REFERRER to be relevant anywhere,
allowing apps to supply referrer information if they want. However,
if they don't explicitly supply it, then the platform now keeps
track of package names that go with Intents when delivering them
to apps, which it can be returned as the default value.
The new method Activity.getReferrer() is used to retrieve this
referrer information. It knows about EXTRA_REFERRER, it can return
the default package name tracked internally, and it also can return
a new EXTRA_REFERRER_NAME if that exists. The latter is needed
because we can't use EXTRA_REFERRER in some cases since it is a Uri,
and things like #Intent; URI extras can only generate primitive type
extras. We really need to support this syntax for referrers, so we
need to have this additional extra field as an option.
When a referrer is to a native app, we are adopting the android-app
scheme. Since we are doing this, Intent's URI creation and parsing
now supports this scheme, and we improve its syntax to be able to build
intents with custom actions and stuff, instead of being all hung up
on custom schemes.
While doing this, fixed a problem when parsing both intent: and new
android-app: schemes with a selector portion, where we were not
respecting any scheme that was specified.
Change-Id: I06e55221e21a8156c1d6ac755a254fea386917a2
Child windows are also removed when WMS.removeWindowInnerLocked()
is called to removed a window. This causes the number of windows
to decrement by more than 1 which causes an out of bounds exception
in AppWindowToken.removeAllWindows() which was expecting a decrement
of 1. Changed code to only continue looping if the size of the
Windows array is still greater than 0.
Bug: 18202119
Change-Id: I6124717272c552ec98e89cbacaadcd964fdba02e
Windows with FLAG_DRAWS_SYSTEM_BAR_BACKGROUNDS were
getting an incorrect content inset hint, because the
hinting didn't see the adjusted systemUiVisibility.
Also adds hinting for the stable insets.
Bug: 17508238
Change-Id: If9647277feb6811b15665b801accd896c51dbd12
This was broken when NetID selection was moved to after
NetworkAgentInfo construction time. NetworkMonitor inherits
from StateMachine whose log tag is immutable after construction
making it impossible to include the NetID. Instead define a
simple log function in NetworkMonitor that includes correct
NetID. This makes debugging via logs much easier and avoids
ambiguities that can make logs useless.
Change-Id: Ie72fc210b92bd425184d288b35ed476669e2e073
We seem to have renamed a file as checked-in twice, which means
we checked it in twice and created a malformed name with the suffix
"-c-c" instead of the correct suffix "-c".
Bug:18280677
Change-Id: Ie3164010898a333e5d9b97151d285ea376de799e
If the broadcast could not be queued due to a stopped user, the
party trying to send a PendingIntent should be notified right away.
Otherwise, for instance, AlarmManager could be waiting forever to
be called back after the broadcast is delivered.
This is a potential fix for:
Bug: 18290018
Change-Id: I07c0751e80f11e69dfa2be5c96a033aecb298b81
