Framework resources have their mResDir set to null, since the framework
resources are implicitly always included. Guard against null
when checking to see if the Resources mResDir matches the asset path
WebView wants to inject itself into.
Bug: 36953234
Test: manual
Change-Id: Ie3dc0cc1240441a2466585224cdc7c15555c66bf
Moving tests for accessibility manager into the a11y
directory and getting them to use some of the newer
test infrastructure. Minor tweaks to
AccessibilityManager to make it testable.
Deleting the AccessibilityManagerService tests entirely.
This class is difficult to unit test with the current
infrastructure, but is covered in CTS tests to a large
degree. I wasn't able to get these working quickly, and
they rely on timeouts that make them run much slower than
the other tests. They also throw exceptions when their
mock accessibility services receive accessibility events,
which causes the system to blacklist those services, causing
the tests to fail differently until the system is reset.
I think it's better to focus on writing new, more robust
tests than on rewriting these tests to pass.
Bug: 36614219
Test: A11y unit tests now work.
Change-Id: I8effe44f22110d55ebb700dc46f59b7e7a1aa946
(cherry picked from commit 9213d23178)
Bug 36740474
When onNewIntent() is called, it calls noteStateNotSaved on
the FragmentManager. It wasn't propagating that to child
FragmentManagers. This CL properly propagates the
noteStateNotSaved().
Test: I7cf360a4ed82ee34797896dc110477731c8206c0
Change-Id: Icfcbc385ef68657145b05cb59d5b2a70fe020992
Currently, authenticate FillResponses do not support partition and follows
the "Highlander approach" (There can be just only one), which causes the
authentication UI to show on all views.
This CL overloads FillResponse.setAuthentication() so it requires the
AutofillIds of the autofillable fields, although behind the scenes it
calls the old method - once clients use the new method, the old method
will be removed and the underlying implementation changed.
The new behavior will be tested by testFillResponseAuthJustOneField()
on LoginActivityTest, although currently it's testing the old behavior.
Test: LoginActivityTest.testFillResponseAuthJustOneField
Bug: 35707731
Bug: 36855717
Change-Id: I601f3e4776aa8763415a06d8d802901a930728d2
Turns out we run this code during early boot, before the device idle
service has even been constructed yet. Find another way to achieve
the needed service execution.
Bug 36865930
Test: manual
Change-Id: I8e3304f37c3a5ee125b73aef2b7d7c7b387aa200
We have seen issues where we fail restarting a process
because there are so many services with so many pending
start arguments that we hit IPC limits. So instead of
doing an IPC for each service start, collect them together
in a list that is sent once, and send it inside of a
ParcelledListSlice to control how much data is written
inline in the IPC.
Test: boot and ran
Change-Id: Ifed26ccdf535871e577fc02c7ef1d09060ab06ca
Add null checks to ScrollView and HorizontalScrollView for checking
the revealOnFocusHint. This should never happen in code called by
the framework, but some apps were hitting it.
Bug: 36379645
Test: none
Change-Id: I220eb88d82126ff08f47a7c2a7fbdddebf07de81
As per CDD: The "android.*" namespace for intent constants is reserved
for public
Android API in AOSP. (Whether public to the full SDK, @SystemApi or
defined in AOSP support libraries.)
ACTION_SERVICE_STATE intent is generally useful for system/oem
apps thus move to system api
Bug: 33679956
Test: Manual
Change-Id: Ie38b53f077e8a013351d35387f9133e0ebb26cc9
This allows more fine-grained control than getting all installed
providers for a user, when you might only want to check for a particular
package. For instance, Launcher can use this API to surface widgets per
app without having to ask for all the widgets.
Test: Unit test on AppWidgetServiceImplTest
$ runtest --path=services/tests/servicestests/src/com/android/server/appwidget/AppWidgetServiceImplTest.java
Bug: 34940468
Change-Id: I182bf1c012d31182024422fc4a63f57f151c3ee5
Instead of trying to be clever by poking at underlying flash part
sizes, rely on the fact that device storage printed on retail
packaging is a power-of-two value.
For a typical device with a 23GiB data partition, this will return
a value of "32GB" which matches the retail packaging.
Test: builds, boots
Bug: 34827187
Change-Id: Ib4cf7f637dffc9238252e1fedcd86dc8b5cf656d
Platform is now providing autofill feature. Disable WebView's simple
form data save feature for platform O and above.
Test: Removing the functionality and the test
Bug: 36869838
Change-Id: If6b9fc12edbe4146fca99d9c6ef8fde36d61f852
