Adding carrier config key to support ability to locally disconnect an
empty IMS conference. The network normally does this, but there are some
carriers which rely on the device to do it.
Test: Add unit test to verify local disconnect when participant count is
zero.
Test: Use conference event package test cmd/intents to inject test CEP to
simulate a conference dropped to 0 participants. Verify IMS conference is
disconnected.
Bug: 151707520
Fixes: 154245549
Merged-In: If6fb25b669a74e7e586736c04101a42ab6adcffe
Change-Id: Id7238bc668e6d6d2d7a04a8696263b260d6b6fd8
Adding new carrier config option to indicate if conference event package
data is supported on the conference peer. We already have an option for
the conference host. This new option is used to indicate when the peer
can or cannot support the CEP data. When not supported, even if the
network provides CEP data, it is ignored by the device.
Test: Manual testing using test intents.
Test: Added unit tests for these cases.
Bug: 150860282
Fixes: 154245604
Merged-In: I47b5e49acf08e085ea23e21a5f6b2d5a5d00bc09
Change-Id: Ib8d8f171699f21f45b0ce2eec2dbe75d7466b76a
1. Let restricted notification that can be dismissed.
2. Only put up restricted notification when any of tethering is
activating.
Bug: 154214549
Test: atest TetheringTests
Change-Id: Ib980aca154036828abdab35e3bb11d42f85ff610
Most of this was previously hidden; these last stragglers were missed.
Test: make update-api ; verify hidden
Bug: 152394802
Fixes: 153376610
Merged-In: I90e05379d3c779bdc6b64e3ff68513bb3f302456
Change-Id: I41bda5b8ad368e1c88e4dd9e45d978a111a22e53
SubscriptionManager#addOnSubscriptionsChangedListener can fail to
register a listener if the TELEPHONY_REGISTRY system service is not up.
Currently this is just silently ignored.
Adding a callback method on the listener to notify the registrant that the
listener failed to be registered, and adding exponential backoff code in
TelecomAccountRegistry to retry registration.
Test: Manual; edited code in Subscriptionmanager to fail the first attempts
to add a listener for TelecomAccountRegistry. Verified the backoff took
place and registration still occurred for the listener.
Test: Tried to write a mockito test but gave up because this code has
far too many intertwined dependencies and is not inherently testable
Bug: 152217039
Fixes: 153376310
Change-Id: Icf3133cdeca526979cb621f29659b880127b680a
Merged-In: Ib032072dcfd2d3d5e700e380af08b7b3142ebb8e
Significantly rework Emergency Affordance Service.
-Enable country-by-country behavior based on ISO not MCC.
-Rely on LocaleTracker for the cellular network country.
-Remove all locking/synchronization.
Bug: 130187110
Test: com.android.server.emergency.EmergencyAffordanceServiceTest
Test: Feature regression test located at go/EAS-regression
Merged-In: I66315988b14d7cbdb5278a1006327818dc5ab11e
Change-Id: I66315988b14d7cbdb5278a1006327818dc5ab11e
(cherry picked from commit 30e7471e5d392fb2693bc825216855104b3e2d6c)
Modules contributing mainline modules (APK/APEX) should set
min_sdk_version as well as apex_available.
For now setting min_sdk_version doesn't change build outputs.
But build-time checks will be added soon.
Bug: 152655956
Test: m
Change-Id: I80b7dcb0bf2ff88e22b65149d957dbbbb0c95b42
Preconditions & Inet4AddressUtils are statically
linked into both framework-wifi and service-wifi.
This creates duplicate classes, which should be
avoided. Instead, service-wifi should simply refer
to the copy of Preconditions/Inet4AddressUtils in
framework-wifi.
(dirty cherry pick from internal branch)
Bug: 153596226
Test: device boots up and connects to wifi
Change-Id: Ie78e5cf0f6ff7bc8a2fa8d7264b742f912601143
Merged-In: Ie78e5cf0f6ff7bc8a2fa8d7264b742f912601143
android_app_import now supports apex_available property. Fill the
property for the CtsShim*Prebuilt apks so that they can be correctly
included in the test APEXes.
Bug: 154488897
Test: m
Change-Id: I9cf2e30bb38abbb33bf76a0eab0dc2a10a5647c9
Expose these constants since they're referred to by a public carrier
config key.
Fixes: 145663507
Test: compile
Change-Id: I276761170b3cf5d740f7dbe945834934082a1428
Merged-In: I276761170b3cf5d740f7dbe945834934082a1428
Volume Group with associated stream type does not have their volume
restore after restart / crash due even if correctly stored in
Settings DB.
A guard prevent to restore it. However, when using volume per attributes,
Settings DB is used, so it is safe to restore volumes for all group.
This CL fixes this issue by removing this guard.
Test: set Music volume & reboot & dumpsys audio
Ensure indexes matches
Signed-off-by: Francois Gaffie <francois.gaffie@renault.com>
Change-Id: I00aa8c17bffca21a128086fbbbeafffc26521a3b
Merged-In: I239145d8bc50237ff902026594423c6d1fa04ad4
When the focus is owned by an external audio policy, and when
a client's request fails, the FocusRequester dedicated Map is not
cleared.
As the client app may not call AudioManager#abandonAudioFocus (since
the requestAudioFocus failed), it leads to not only dumpsys side effects but
also it triggers upon client's focus binder died event, the call
of notifyExtFocusPolicyFocusAbandon_syncAf.
This CL fixes this issue by removing the FocusRequester when the request
fails, the same way it is already done on permanent loss of focus.
Test: build & boot
Signed-off-by: Francois Gaffie <francois.gaffie@renault.com>
Change-Id: I4d1b216b606c8f73c9b68353723284d9e178297d
The CL that moved the initialization of the tethering offload
config HAL from C++ to Java caused the code not to retry
fetching the service if it is not ready when tethering is
started. This is because the C++ version of getService() retries,
but the Java version only retries if getService(true) is called.
Make the new code retry as well.
b/152430668#comment4 asserts that the fetch will be retried only
if the service is installed on the device, so the retries should
be attempted (and thus should not have any startup time impact)
on devices that do not support tethering offload.
Bug: 152430668
Test: builds, boots, tethering offload works
Change-Id: I093f127d90b2aa1b13eb0748378a24726d419472
These would cause these actions (and any that depended upon them) to
re-run during every incremental build. These now produce warnings:
ninja: output file missing after successful execution: out/soong/.intermediates/frameworks/base/ds-ref-navtree-docs/android_common/ds-ref-navtree-docs-proofrerad.txt
It turns out that the -proofread argument to doclava is ignored with the
following argument combinations:
* -static -devsite
* -navtreeonly
Test: m docs; m docs (nothing to do)
Change-Id: I21b939951f5a2f33914e9fc9414df8206091d262
1) Refactor ParcelableConference - too many optional params
2) Add a new call direction parameter for external multiparty
call scenarios, where a conference is assigned a direction.
Bug: 153595501
Fixes: 154246448
Test: atest TelecomUnitTests TeleServiceTests
Merged-In: Icb576a07b9a7a81e228babfea53474be83da7593
Change-Id: Icb576a07b9a7a81e228babfea53474be83da7593
R_EXTENSION_INT is 0 now, but if bumped we don't want to bump
the other return values of this method.
Bug: 154486201
Test: SdkExtensionsTest
Change-Id: I36855ba1881da4056276f4a5626ca64d2129353b
Merged-In: I36855ba1881da4056276f4a5626ca64d2129353b
(cherry picked from commit 11111482211f0ff59f54f71cf0145ab572af3fae)
