Bundles can't be read from multiple threads safely. This adds locking
around a read that had been previously overlooked and ensures that
unparcel is called on the metadata before it is even available to
other threads.
bug:17894033
Change-Id: I9a4b86a0d0af05b1dcba28a52df2e7a87c685704
During restore first the provider records and then the applications
are restored. Therefore, we keep the restored provider records in a
zombie state and once we get the package install broadcasts from the
package manager we update the zombie providers to be real ones. The
bug was that while doing this we were not updating the provider id
from one with an undefined UID to the one with the UID under which
its package now runs.
bug:17869323
Change-Id: I20b7cbfd2f90e2f5bf3a6b858d8da99cc1c82841
The feature causes Fugu to turn on TV for unexpected cases.
Disables the feature until we figure out a better way to find
right wake up events.
Bug: 17875468
Change-Id: Icfb04f40e5b13f751b7466ea9789ba8eadcddea6
PackageManager and AppWidgetHostServiceImpl should
be loading the resources of any shared libraries being
used by the app, as they have references in their Widgets
or application icons/labels, etc.
Bug:17668152
Change-Id: I359662334edb125d7570089916727df4eeba02bb
When creating a user via the UI, disallow phone calls and SMS by
default. Primary user must explicitly enable it via Settings.
Bug: 17832802
Change-Id: I18cad4be8493ddc8890b5d90da2df256cb3f1ec9
When performing a restore during initial device setup, we could be
installing hundreds of packages. Currently, we're writing all
metadata (including heavy icons) for every session mutation! Because
we're holding the mSessions lock while writing all this heavy data,
we end up causing ANRs when apps call other PackageInstaller APIs.
This patch mitigates by moving the heavy icon data into separate
per-session PNG files, which we only persist when changed.
Bug: 17881962, 17567794
Change-Id: I4dee15d4a65a8eb65c381e6bb7477728b6cc30d2
Set active source status to false when broadcast routing control
commands indicate other device is about to be the new active source.
Bug: 17840270
Change-Id: I176e21ec8789624e29421b912ba641a21f0f5f21
When a system app has been upgraded, we mark the built-in version as
disabled, and skip it when parsing built-in apps, since we expect
the userdata version to be scanned a few moments later.
However, in very rare cases we can end up missing the expected
userdata version, and we'd end up dropping the package entirely, even
though we have a built-in version to fall back to. This change
handles that case by rescanning and enabling the built-in version when
the userdata version never materializes.
Also include critical log messages in check-in dumpsys to help track
down how we ended up in this state.
Bug: 17805839
Change-Id: I9971f882f9bb8ab7934c20b04e0b72366c4d04d0
Bug 17811327 : teach adb restore about the new widget metadata entries
Bug 14165872 : -nosystem should not act like -onlysystem
Change-Id: I39da0ba80df7c5309a78ec1fa38016cebd80aa5f
Fix bug where we disabled system app, but never turned it back on
when the scanPackageLI() failed.
Bug: 17805839
Change-Id: I73999263aee703af187afd980fa0d0ce8451cf0c
Keep around GET_TASKS as a permission available to apps, so apps still
think they have it and don't get all uppity because they don't.
Add a new REAL_GET_TASKS that is the actual permission now.
Plus some temporary compatibility code until everyone can transition
fromn GET_TASKS to REAL_GET_TASKS.
Change-Id: I12c1102eed24844685dcbd2fa3b612811603958f
If there is accessibilty focus and the user touch explores
location that does not change accessibility focus that is
not in the app window, e.g. system bar, double tap does not
click on the system UI affordance.
bug:17588024
Change-Id: I6c8c0f65b208ae1d3f31d7f06b0721dc223ec19f
The placeholder for disconnected networks was setting it to false, but
this technically means that we know an attempt to connect to that
network will fail (which we don't really now). Some applications use
this an decide not to bother trying - an MMS app for example would
never send an MMS because it thinks the network is never available.
This is a L regression.
bug:17669247
Change-Id: Id6041f226da069c267c56418d42c55978c36c66f
We've been seeing some really funky behavior when upgrading or
downgrading system apps around OTA time. Put more of these one-time
logs into durable storage to help investigate.
Bug: 17805839
Change-Id: If898d7df229c1f71e598b0d965325c272060e5e7
This reverts commit 851a5059a47cbf76e530c9d050a677cb6e3f8657 as
it creates a regression. Let us revert this and correctly fix the
issue the original change was trying to address.
bug:17789608
Change-Id: I8abb1a61d5310430e839e4ef60e7ca5cc0cbdd80
We also refine the order of delivery within any given package. Now,
we identify which apps have wakeup alarms being delivered in the
current pass, and deliver all of that app's alarms before moving
on to alarm delivery to apps who are only receiving non-wakeup alarms
in the current delivery pass. The TIME_TICK alarm is also hoisted to
the start of the current delivery pass if present.
Bug 17778686
Change-Id: I6306a00fe657787a77d0254c0807ac51e810fdcf
