We switched the package installer app to always abbreviate all labels
loaded from foreign packages. This is done so that a foreign app cannot
crorrupt the UI via super-long strings.
The default abbreviation is very aggressive and abbreviates to ~5 cm.
When an app with a targetSDK <= M gets installed, all permissions will
get auto-granted. Hence when a user side-loads such an app, we show the
permissions during install so that the user is aware of this.
Unfortunately most permission labels and restrictions do not fit into a
single line.
This change allows much longer strings for permissions names and
descriptions. It still abbreviates eventually and handles permission
descriptions with weird characters (such as back-space) correctly.
Fixes: 110209410
Test: Side-loaded old app and verified that permission labels and
descriptions are not abbreviated.
Change-Id: Ie1db240883c9f5cb9c1b74ca7bdcb80625e3f7f2
It isn't one, and I should have read the documentation on it properly
before treating it the same as EXTRA_LARGE_ICON (which is indeed a
Parcelable).
This fixes a chronic outpouring of technically-probably-harmless warning
stack traces of the following nature which can be verified by running
the framework tests and grepping logcat for this scary-looking string:
java.lang.ClassCastException: java.lang.Integer cannot be cast to
android.os.Parcelable
at android.os.Bundle.getParcelable(Bundle.java:***)
at android.app.Notification.fixDuplicateExtra
at android.app.Notification.fixDuplicateExtras
at android.app.Notification.readFromParcelImpl
Duplicate Integer objects are harmless compared to the Bitmap so there
is zero sense in recombining the objects after an RPC. We just drop the
line which tries to do that.
Test: make FrameworksCoreTests -j30 && \
Test: adb install -r ${ANDROID_PRODUCT_OUT}/data/app/*/FrameworksCoreTests.apk && \
Test: adb shell am instrument \
Test: -e class android.app.NotificationTest \
Test: -w com.android.frameworks.coretests/android.support.test.runner.AndroidJUnitRunner
Change-Id: I2c88ab4e9d5322c05e2e7f68df85ed5b2f296f84
Fix: 79404137
(cherry picked from commit 62eea67846)
Fixes: 72715022
Test: With two devices, set up one, back up onto the other.
SUW for Assist Gesture does not get skipped
Test: atest SettingsBackupTest
Change-Id: Id7348542a618c0366f89c47833af6f66bf52b18f
If draw() invokes postDraw that sets the next callback, the
callback will never be invoked.
Test: Long-swipe-up from navigation bar
Change-Id: If2e4dfff8a845b73fa77426305d5ece1921857e5
Fixes: 109701914
Doc-only change to note that requestRestore() doesn't behave as expected
and kill the app as part of the restore operation.
Fixed formatting of first paragraph running into deprecation notice.
Also removed dupe versioning info from deprecation notice--see staged version.
Staged at:
http://go/dac-stage/reference/android/app/backup/BackupManager#requestrestore
Bug: 29255593
Test: make ds-docs and staged output. Visual inspection and manual link test.
Change-Id: I8baae2216f23bd63510f8b445fcdd8090354b0df
Although parent view may invoke onDetachedFromWindow incorrectly,
itself should guarantee to unregister receiver only when removing
the last attached view.
Change-Id: I4b701c223c6aa354896a4067f77aad9c98d88dfd
Fixes: 78506836
Test: atest FrameworksCoreTests:DateTimeViewTest
When NFC event transaction intent is sent,
the device shows AID information on logcat.
But such information should not be shown for security reasons.
Test: Send NFC EventTransaction and check logcat
Bug: 80551185
Change-Id: I8d1f9c1946bb1514e3a6bfa70fff03be4525d6b5
(cherry picked from commit 37e7028615)
As part of adding the hint text feature in
I357dd5c74b61d149cf8612d1f52c7118ec70c696 I refactored
the code inside reportExtractedText to avoid nesting but lost one
of the condition checks in the process. Since onDraw calls into
this method on each frame, the missing check was causing calls to be
made into the IME even when no content or selection changes have
happened.
Test: CtsWidgetTestCases:.TextViewTest
Bug: 73613936
Change-Id: If56e3f1d45e64dccd052e4cff4d742f0cbecc07c
The BinderProxy class is not thread-safe, hence all calls into it
must be serialized. This was achieved by holding the gProxyLock in
JNI code. However, a recent change added calls into BinderProxy
from ActivityManagerService without holding that lock, causing
ConcurrentModificationExceptions.
Instead of dumping debug info from AMS, make the call directly
from JNI, so we can make sure gProxyLock is held correctly.
Also, only dump on debug builds.
Bug: 71353150
Bug: 109701487
Test: sailfish builds, boots, info gets dumped with lowered limits.
Change-Id: I446a71ce4115b9936a01a170401ef98ba3818c0b
This reverts commit 44c4ca41d8.
Reason for revert: Wanted to go with a different option and this was mistakenly Dr. No'ed.
Bug: 80444486
Test: clean revert
Change-Id: I7fdd13a5bc0f4d9fabf91f3e5497c4065008b531
This CL fixes the problem that vendor public libraries (libs that are
listed in /vendor/etc/public.libraries.txt) are not loadable via
System.loadLibrary(). (Note that the libs were accessible via dlopen()
though.)
The problem was happening because when System.loadLibary() is called,
the classloader first checks whether the lib is found and accessible in
its own native lib search paths. The native lib search paths basically
come from the java.library.path property, which in turn is from
namespace.default.search.path in /system/etc/ld.config.txt. When VNDK is
enforced starting from P, namespace.default.search.path does not have
paths other than /system/lib because otherwise system process can have
unlimited access to libs in vendor partition. The linker namespace is
dynamically configured by the libnativeloader so that only the public
vendor libs are accessible. However, as a side effect of removing
/vendor/lib from namespace.default.search.path, the classloader always
fails to find any lib under /vendor/lib even if the lib is a public one.
In order to solve the problem, while keeping rest of the non-public
vendor libs from apps, /vendor/lib (and /odm/lib and /product/lib as
well) is added to the classloader AFTER native loader is created for the
classloader.
Bug: 93333337
Test: m -j
Test: System.loadLibrary("adsprpc") is successful in Pixel (because
libadsprpc.so is in Pixel's vendor public lib list)
Test: atest cts/tests/tests/jni
Change-Id: Iac12384548cbdd51234568082d02eeba466c160c
setSynchronousMode is now applicable to any journal mode.
This can be useful for apps that want to use stricter durability
settings for their database.
Test: android.database.sqlite.cts.SQLiteDatabaseTest
Bug: 86953240
Change-Id: I7dbaa6d0bc0975da8067e87b402f63f51b4add10
If an app crashes during Keyguard transition, make sure to keep
Keyguard transition
Test: AppTransitionTests
Test: go/wm-smoke
Change-Id: I80b80952f93d2b5611754f05a3dc333905cd1c86
Fixes: 80132133
Update the logic in PackageParser to avoid unistalled system app
can't be found.
Test: follow steps on b/80444486 and can add account normally
Test: PackageManagerTest
Bug: 80444486
Change-Id: Ifafac34111d883e51563018be793655c25d09fba
Th wait() call in AMS.getContentProviderImpl() can cause a deadlock when it
starts a new process for a target provider and all system server binder threads
are reaching out to the same provider at the same time, because AMS won't be
able to receive a callback from the provider process.
Let's serialize access to AM.getContentProvider() to avoid this scenario.
Longer team, we should stop waiting in the system server and instead do so in
the client side.
Bug: 74523247
Test: Boot, add a google account, and let syncs run and finish.
Test: Watch free video on Coursera
Change-Id: If5be2dc4b6f22f72d1bb3aa1b5b4d49f20c8d94d
