Bug 33053923
Also added an early error check when the same shared element is
used for multiple shared elements in the source or target fragment.
Test: If5dfbac66cfd9e60660439ce5c519547de3ef1d0
Change-Id: Iac3ab8c1b8ea0208a0c76e043654184dd634948d
We want to create the classloader for the WebView in advance in the
zygote so that it can preload Java and native code for its children, but
the zygote can't talk to the package manager (so doesn't have a
PackageInfo for the APK) and also doesn't have an ActivityThread, so
constructing a LoadedApk is difficult.
Instead, we use the fact that ApplicationLoaders contains a
process-global cache of classloaders for APKs, and prepopulate a cache
entry without constructing a LoadedApk. This requires making
ApplicationLoaders public. To calculate the correct library paths from
the information the zygote has, we reuse the logic in LoadedApk (which
is already public, and just needs a small change to allow a null
ActivityThread when checking for instrumentation).
The other parameters for classloader creation (target SDK, bundled app,
etc) are hardcoded to usable values for the WebView's case. WebView
never needs to use any system libraries that aren't public so claiming
it's not bundled is fine even when that isn't actually true, and WebView
will always target the current platform API level.
Once the classloader is created, look up the factory class and call
preloadInZygote on it to give it a chance to preload the native library
and do other shared initialisation.
Bug: 21643067
Test: enable multiprocess WebView, examine librank output to see sharing
Change-Id: I696ead637e3f7382bcc58cfaf61eac5921862015
Resource creation is going to get much more nuanced
than just passing in a static set of paths, so move
the creation of Resources out to where the context
of the creation is more easily understood.
Test: Existing CTS tests pass
Change-Id: If5d79e0f84a111790c15a7b44db3f44c75e7141c
The default implementation uses the main looper to handle service
operations. This is generally okay until we have to make network
calls to retrieve resolution data.
Bug: 25119046
Test: manual
Change-Id: I22b0f279e6c1fde21f207578d2930ab4858bbc78
Unless the app is not targeting O and the user hasn't set
a preference in Settings.
Also apps no longer need the vibration permission to vibrate with
a notification.
Test: runtest systemui-notification
Change-Id: Ia84b1c97d13cb479db07035ca622b1fe52261329
This CL adds bookkeeping to DPMS which will allow us to tell the user
in the Settings UI whether/when the admin requested a bug report or
retrieved logs from the device.
Bug: 32692748
Test: Full DevicePolicyManagerTest unit test coverage; end-to-end tests
will follow as Settings CTS verifier tests
Change-Id: I89728fce4b7e0ff061b354c73caf3742e95a3a3e
Settings needs to access a variant of getInstalledApplications() which
takes a |userId| argument. Since this is not currently exposed by
PackageManager, Settings calls into PackageManagerService directly. This
is ugly and breaks the regular abstraction layer hierarchy.
The CL fixes the problem by exposing the required variant of
getInstalledApplications() as a hidden API, analogously to what was done
before with getInstalledPackages().
Bug: 32692748
Test: Will be CTS-verifier-tested together with Settings
Change-Id: Id9c4e8e18524d312159821f1a4d5527263c7e950
For phase-2, instead of taking a hash prefix, we will send the
hostname to the resolver.
Bug: 25119046
Test: build & install the sample resolver and run 'adb shell am start -a android.intent.action.VIEW -c android.intent.category.BROWSABLE -d "https://www.tripadvisor.com/Tourism-g33020-San_Jose_California-Vacations.html"'
Change-Id: I6e621139e182f9a9cfcf608c26d4763f1379bdb2
For an app to either send or receive content change notifications,
require that they have some level of access to the underlying
provider.
Without these checks, a malicious app could sniff sensitive user data
from the notifications of otherwise private providers.
Test: builds, boots, PoC app now fails
Bug: 32555637
Change-Id: If2dcd45cb0a9f1fb3b93e39fc7b8ae9c34c2fdef
Framework edition
* Ensure that a fragment that has had its view destroyed has a chance
to recreate it.
* Clear mInLayout after views are destroyed
* Make sure that when a fragment is restored from the back stack and
reinflates its view that child fragments from fragment tags are
reconnected properly.
Bug: 32691935
Test: cts
Change-Id: I370d6db46e72fd8ac35f716e2130b9e10b93b1da
Bug 32472451
It is important to remove an OnPreDrawListener from the correct
ViewTreeObserver. When a View is added to the view hierarchy, the
initial ViewTreeObserver is not active. The listener must then be
removed from the current OnPreDrawListener. When a View has been
removed from the hierarchy, it is important to remove the listener
from the orignal ViewTreeObserver.
Test: cts-tradefed run singleCommand cts -d --skip-preconditions
--skip-connectivity-check -m CtsUsageStatsTestCases
Test: cts-tradefed run singleCommand cts -d --skip-preconditions
--skip-connectivity-check -m CtsFragmentTestCases
Change-Id: I735f71d2d9c84e86ef846aab0088a8651300fbe8
displayId can be specified as a part of ActivityOptions.
We will look for a valid stack on the specified display, starting
from the topmost one. If no valid stack is found, new dynamic stack
can be created on external display.
Launch stack and display id can't be set at the same time, otherwise
exception will be thrown.
Test: ActivityManagerDisplayTests
Test: #testLaunchActivityOnSecondaryDisplay
Test: #testConsequentLaunchActivity
Test: #testConsequentLaunchActivityFromSecondaryDisplay
Test: #testLaunchActivityFromAppToSecondaryDisplay
Change-Id: I8a202bc076319e23948d81b99a2c68d7b733b5e4
Bug 32610133
A change in behavior was introduced in which the dispatch* methods
moved the state of the manager's fragments even when the manager's
state doesn't change.
Test: I88f5a3313ad05429167c417b840f888f86fd1c11
Change-Id: Ie260b1c57fae505b2636ff7572a493a92449dff4
The code would previously interpret "null" (meaning the user
hasn't touched the setting in the Settings app) as meaning
"use 12 hour format". The actual meaning of "null" is
generally accepted in the platform as
"use the locale default". For example, the Settings app does
this, as does the code in libcore.icu.LocaleData. For
locales where the default is "use 12 hour format" anyway
this wasn't obvious. However, locales like Germany should
default to 24 hour, and then the setting is displayed
incorrectly in the Settings app.
Also, German devices boot for the first time with 12 hour SHORT
and MEDIUM time formats too. After the user had explicitly set
the format the setting would be correct.
There appear to be no tests for this behavior, presumably
because of the difficulty in forcing the setting to "unset"
which would have to have to happen before the app was launched,
and the fact that CTS tests run in the US locale.
Bug: 32868036
Test: Manual testing with a device set to German
Change-Id: Ifd2e8d345f6afee467d7525d6793fc8fda37c900
