This change makes it so that PointerController does not ask its Policy
to load any resources for any displays until a DisplayViewport is set,
and verifies this with unit tests.
Bug: 145699789
Bug: 146385350
Test: atest libinputservice_test
Change-Id: I2e48e7ac4700e6f9fdf939a7bd0e6639b051ade6
Merged-In: I2e48e7ac4700e6f9fdf939a7bd0e6639b051ade6
InputReader is responsible to associate device and display so it makes
sense to allow it set display viewport for pointer controller.
Bug: 146385350
Test: Cursor can be associated with external freeform displays as
expected.
Change-Id: I00d664dd180f1e693b1900582feea8f7ff02f93c
Merged-In: I00d664dd180f1e693b1900582feea8f7ff02f93c
In the test below it only expects MockSprite#setIcon() being called
once, but with animation it may be called multiple times.
Bug: 144947344
Test: atest PoitnerControllerTest#updatePointerIcon
Change-Id: I5c4908fb8301cae144fd637c831f2012a35862e3
Merged-In: I5c4908fb8301cae144fd637c831f2012a35862e3
Also bootstrap unit tests for PointerController. Need to mark 3
functions of SpriteController virtual so their behaviors can be
overridden.
Bug: 130822623
Test: SurfaceFlinger can get cursor type and hotspot.
Change-Id: I739cd03214364144bb4e22a166ecc7abfd3492fe
Merged-In: I739cd03214364144bb4e22a166ecc7abfd3492fe
The names of the individual modules do not quite follow the pattern
that java_sdk_library uses so this temporarily sets the following:
naming_scheme: "frameworks-modules"
That causes java_sdk_library to use a naming scheme that matches the
one used by the individual modules of this. It will be cleaned up
later.
Part of the purpose of the java_sdk_library is to hide the
implementation code and force users of the library to depend on stubs
for a well defined API. Ideally, it would allow access to the
implementation in those cases where it is safe, e.g. from within the
same APEX, or from tests for the implementation. Unfortunately, due to
limitations in the build it does not yet have enough information to
make that decision correctly which means that any code that needs to
compile against the implementation is broken which would prevent us
from converting the module to java_sdk_library.
However, the only way to provide the additional information to allow
the implementation to be correctly exposed is to convert the modules
to java_sdk_library; a cycle.
In order to break that cycle the java_sdk_library creates a special
<module>.impl target which is used directly by tests and any other code
that needs it. Once all the modules have been converted to a
java_sdk_library then we can resolve the limitations in the build and
remove the direct references to <module>.impl.
Test: m update-api
Bug: 155164730
Merged-In: If5c115f482751f9f4b5f047e9e401a18e36799ef
Merged-In: Id1c2e848430c49a2da7402244814cd084f5da77c
Change-Id: Id1c2e848430c49a2da7402244814cd084f5da77c
This should no longer be needed after all the cleanups.
Test: atest NetworkStatsTest NetworkStatsFactoryTest
Bug: 150738490
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I289d935f84b616ed857ef4c5a7427d57c282d00c
No clat app uid 0 rx stats on base iface due to ip6tables raw prerouting drop
(so just copy over the rx stats from the before file)
Test: atest NetworkStatsTest NetworkStatsFactoryTest
Bug: 150738490
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ie73ba0586dafde67bc2726db26e07f7268c18be3
No clat app uid 0 rx stats on base iface due to ip6tables raw prerouting drop
Test: atest NetworkStatsFactoryTest
Bug: 150738490
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I4393afcbe8e1bb886ecf4f5c1573f6ac8a0e29f5
No clat app uid 0 rx stats on base iface due to ip6tables raw prerouting drop
Test: atest NetworkStatsFactoryTest
Bug: 150738490
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ifebd946e06b5fa2eeb7ff4fc0b1c423097021692
This is due to the previous addition of the ip6tables raw PREROUTING
drop rules for incoming ipv6 clat traffic pre-translation to ipv4.
Since we no longer double account, we no longer need these fixups.
Test: atest bpf_module_test clatd_test libbpf_android_test libnetdbpf_test
netd_integration_test netd_unit_test netdutils_test
Bug: 150738490
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ia171b7797cdc99367064d0649bf1293c71579941
Extend the timeout to lower the EthernetTetheringTest flaky.
Besides, also explicitly exempt entitlement check in
EtetherntTetheringTest because it do not test tethering upstream
currently. Thus, the tests would not be interrupted by entitlement
check if test SIM is entitlement required.
Bug: 156713866
Test: TetheringCoverageTests, CtsTehteringTest
Change-Id: I45e8e8d737486def9d0de8943ec7f09ca0942a0b
Exempt-From-Owner-Approval: dianne approved, but from her android.com account instead of google.com account
Test: compile
Bug: 156757711
Merged-In: I330b8cc5aeb14ebcc15fb88c605ac72770497040
Change-Id: I330b8cc5aeb14ebcc15fb88c605ac72770497040
Bug: 151454839
Test: See the other CL in the topic
Merged-in: I4e1d6c419c79663a47341cf26d5229ec17b9ca01
Change-Id: I4e1d6c419c79663a47341cf26d5229ec17b9ca01
When notifying the framework of an incoming call, a lock is held
in both the MmTelFeature and MmTelFeatureConnection. This can cause
a deadlock if the MmTelFeatureConnection is also processing an
event, such as sending/acknowleging an SMS message.
Remove the lock around the listener methods to the framework, since
it is not needed.
Bug: 155083563
Fixes: 155819733
Test: atest CtsTelephonyTestCases:ImsServiceTest FrameworksTelephonyTests
Change-Id: I39f7192c6f79e215ef989797870f5f501197cd08
Merged-In: I39f7192c6f79e215ef989797870f5f501197cd08
Avoids direct access to the stub sources and instead access them via
tags passed to the java_sdk_library.
Bug: 155164730
Test: m checkapi
Change-Id: I93dfb8bfac60bd7ab628292667c8d7a9d5e8eb82
Add checks during boot in case the persist.sys.timezone property is set
to a bad ID.
This can happen in the rare case of a mainline rollback: i.e. if a device has
been set to a new ID and then the update is rolled back. Using GMT as a
fallback probably works without this change (it does in java.util.TimeZone),
but relies on all code, including native code that uses
persist.sys.timezone directly, knowing to interpret a bad ID as "GMT".
This commit makes that choice more explicit and defensive.
This change also removes the possibility of IOException, which is never
thrown, from some ZoneInfoDb methods.
Bug: 155738410
Test: boot with a valid id, verify persist.sys.timezone is unchanged
Test: boot with an invalid id set, verify persist.sys.timezone is "GMT"
Merged-In: I6dc0f4f81848efbbaec6a11a62014471a0ef01fd
Change-Id: I6dc0f4f81848efbbaec6a11a62014471a0ef01fd
Exempt-From-Owner-Approval: Approved / landed internally
Defaults shared library to false as the majority of framework modules
should not be usable as a shared library. The exception is ike but it
does not currently use the defaults and will override this setting
when it does.
(cherry picked from f657ae3ee5)
Bug: 156723295
Bug: 155164730
Test: m droid and flash
Change-Id: I3e78994a9855110b02c3d71944ff38b986208e9a
Merged-In: I3e78994a9855110b02c3d71944ff38b986208e9a
