to install an app on sdcard, just resources.
Add new install path for /asec in installd.
ignore . when checking for apk path since the sdcard packages id'ed
by package name.
Add new -s option to adb shell pm
Refactor fwd locked from scanMode to ApplicationInfo.
Add new flag for sd install
Add new parse flags for fwd locking and installing on sdcard
New mock api's in PackageManagerService to invoke MountService api's. These
will be refactored again and so have been wrapped internally.
Some error codes in PackageManager
Changes in PackageManagerService to use mPath and mScanPath during installation
and switch to using PackageParser.Package.applicationInfo attributes for
source and public resource directories.
Some known issues that will be addressed later
using system_uid for now. needs some tinkering with uid and packagesetting creation to use the actual app uid
error handling from vold not very robust. ignoring lot of things for now
sending a delayed destroy to delete packages. will revisit later
revisit temp file creation later. just copy for now
Prevents sending out ConnectivityManager broadcasts when the connection
is repeating the same major state. This means continued wifi scanning
does not generate CM broadcasts (though wifi state broadcasts continue).
bug: 2265222
Dr No approval: jsh
This class provides an interface to communicate with native daemons using the
libsysutils framework communication code.
Signed-off-by: San Mehat <san@google.com>
NativeDaemonConnector: fix whitespace
Signed-off-by: San Mehat <san@google.com>
NativeDaemonConnector: Fix review comments
Signed-off-by: San Mehat <san@google.com>
NativeDaemonConnector: On an error, re-try the connection
Signed-off-by: San Mehat <san@google.com>
Merge commit 'c93c70c4b849d31ff1b5e5cf2fb9462fd60e5da0' into eclair-plus-aosp
* commit 'c93c70c4b849d31ff1b5e5cf2fb9462fd60e5da0':
framework: storage: Ensure that filesystems are unmounted before shutdown/reboot
Eliminate the per-process 200ms timeout during ANR thread-dumping.
Dump all the threads at once, then wait for the file to stabilize.
Seems to work great and is much, much, much faster.
Don't dump stack traces to traces.txt on app crashes (it isn't very
useful and mostly just clutters up the file).
Tweak the formatting of the dropbox dumpsys a bit, for readability,
and avoid running out of memory when dumping large log files.
Report build & kernel version with kernel log dropbox entries.
Merge commit '3fb4ee4e03351078ceadf3d8dd05789ae152e8dd' into eclair-plus-aosp
* commit '3fb4ee4e03351078ceadf3d8dd05789ae152e8dd':
Immediately close the shade when the status bar is disabled. (DO NOT MERGE)
Some of the native calls were left unsynchronized in the framework. Pre-empted IOCTL call
interrupted by another call from the framework cannot be handled in the driver.
Bug: 2310455
Shutdown handlers run too early for unmounting, so don't handle it.
Shutdown will be called explicitly from the ShutdownThread
Signed-off-by: San Mehat <san@google.com>
Adds android.os.RecoverySystem (marked as pending) to replace the
(hidden) com.android.internal.os.RecoverySystem. RecoverySystem
contains methods for:
- verifying the signature of an update package
- rebooting to install a package
- rebooting to wipe user data
(The reboot functions require "android.permission.REBOOT" and
"android.permission.ACCESS_CACHE_FILESYSTEM".) Providing these
simplifies implementation of OTA update for device builders.
Change-Id: I63ce743b156e7a1a0327fd395b0e4a82c0eda79a
The ActivityManager.restartPackage() API is now deprecated, and no longer
allows applications to mess up the state of other applications. This was
being abused by task killers, causing users to think their other applications
had bugs.
A new API is introduced for task killers,
ActivityManager.killBackgroundProcesses(), which allows these applications
to kill processes but only the same amount that the out of memory
killer does, thus causing no permanent damage. The old restartPackage()
API is now a wrapper for calling this new API.
There is also a new private forceStopPackage() API that is used for the
system's force stop UI which does what the old restartPackage() API did.
Squash of the following:
services: MountService: Rework the way volume states are handled
MountService: Add new API for directly getting volume state via a mount point
Environment: Switch from using system property for external storage state.
MountService: Add support for UMS
MountService: Fix a few bugs
services: MountService: Add support for mount-on-insertion
services: MountService: Add some debugging around UMS
services: MountService: Fix some UMS bugs and clean-up startup mount code
Signed-off-by: San Mehat <san@google.com>
LocationManagerService was just checking if the string of (comma-separated)
Location Providers contained the provider we were interested in. This works
fine in normal cases, but breaks if we add a provider such as test_network.
Enabling test_network causes LocationManagerService to think that the network
provider is also enabled.
The code in Settings.Secure.isLocationProviderEnabled() checks for the commas
in the string as well, to make sure that a provider name which is a substring
of another provider name won't cause problems. It also centralizes the code
which reads the string.
Signed-off-by: Brad Larson <brad.larson@garmin.com>
This provides a quick fix for http://b/2298803 (if the
in-call UI comes up while the user is holding the
windowshade, the shade sticks in place and the display
becomes unresponsive).
Merged from master (I92a824fe).
TODO: safely restore code to animate the shade closed rather
than simply causing it to disappear.
Change-Id: I53dc76fdfe86636f30f37a223a88d27a5766d3eb
This provides a quick fix for http://b/2298803 (if the
in-call UI comes up while the user is holding the
windowshade, the shade sticks in place and the display
becomes unresponsive).
(NB: The shade disappears immediately when the status bar is
disabled. It might be better for the shade to animate
closed; indeed, this seems to be what the code was trying to
do, but an unfortunate interaction of the animation and
touch-handling code is what caused this bug in the first
place.)
Change-Id: I92a824fe385c8d6e51993be1657b492e580f1e7b
