Expose KEY_ADDITIONAL_SETTINGS_CALL_WAITING_VISIBILITY_BOOL and
KEY_ADDITIONAL_SETTINGS_CALLER_ID_VISIBILITY_BOOL so carriers
that don't support the feature can override them to false
to hide the feature in calling addtional settings.

Bug: 266479292
Test: m -j
Change-Id: I9ae635c051c0f9e180c36aa1725d0f058f1fd70d
Merged-In: I9ae635c051c0f9e180c36aa1725d0f058f1fd70d
(cherry picked from commit c86c69b1)

* changes:
  Add an install flag to force non-staged APEX update
  Add --non-staged flag

Fix minor typos in Extension classes

Test: m
Change-Id: Ieae5ea038d895167126f88cc00d09a9ceafff917

Removed the incomplete APIs and re-directed to the
similar APIs.

Fix: 262591408
Test: atest SubscriptionManagerTest
Merged-In: I1e0dd566114550099de3e0d0eedfcab79c4e54b6
Change-Id: I1e0dd566114550099de3e0d0eedfcab79c4e54b6

This CL introduces a hidden utility method for system components to
get the subset of carrier config.

Many system components has been update to use the new
CarrierConfigManager API to get subset of config. But the callers
side have repetitive private methods to do the same thing. The
utility method in CarrierConfigManager will remove the duplication.

Bug: b/263267340
Test: FrameworksTelephonyTests
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:96727486197e8c0602ffb8f27787aae1b256fa44)
Merged-In: Ifbd84c67525fd9ea2925e5b735788e6c3c22dd34

Change-Id: Iac88c03f2c9b4ac80dd2ed0b72a200cfc1631d21

- Replace CarrierConfigChange receiver with listener
- Get subset of carrier configs with new API instead of private method

Bug: 265961787
Test: atest TelephonySubscriptionTrackerTest
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:628147b7a0e1f6126245ae702f31bb2cf5bd1d4d)
Merged-In: I93f14421624ac76bdf84c06cfd4c067bf2ccde7a
Change-Id: I93f14421624ac76bdf84c06cfd4c067bf2ccde7a

Add RemoteAuthService that implements a logic of D2DA

Design doc: go/remote-auth-manager-fishfood-design

Test: loaded on device.
Bug: 290092977
Change-Id: I2fdc8b87241439c051c4706e9970639c7c83c763

Merge "Make sure to send HOVER_EXIT when starting touch exploration if we haven't sent it already." into main

Using ResourceProcessorBusyBox to speed up resource processing
reuires using the correct package to reference resources from
android_library dependencies.  Fix them, and set use_resource_processor: true.
This saves around 47 seconds when compiling the SystemUI module, and
a similar amount when compiling the SystemUI-core module.

Bug: 284023594
Test: m SystemUI
Change-Id: I8f825ef549a5e0bd67e10d260ecdc3353a526fdb
Merged-In: I8f825ef549a5e0bd67e10d260ecdc3353a526fdb

Cleaned up the old APIs and replaced with the new APIs.

Bug: 270094276
Test: Boot up + Basic phone funcationality tests
Change-Id: Id372018a437e72c1b5874dec620b5158337bf43c
Merged-In: Id372018a437e72c1b5874dec620b5158337bf43c

This is a development only feature to speed up development workflow for
teams that have their code package in an APEX.

Bug: 290750901
Test: m
Test: atest ApexManagerTest
Test: see other change in the topic
Merged-In: Ic0abb9e97d529910805af50208baf42b9b28b171
Change-Id: Ib452dcdbdd30991f60205f94be25389379153183

The child change will change the behaviour of --force-non-staged flag to
perform a rebootless APEX update even for APEXes that don't support
rebootless updates. This behaviour will be used to speed up development
cycle for teams that have their code packaged in an APEX.

However, some developers (and some tests) assume that adb install
--force-non-staged will fail for APEXes that don't support rebootless
updates. In order to preserve such behaviour, this change introduces a
new --non-staged flag.

Bug: 290750901
Test: adb install --non-staged apex-supporting-rebootless-update
Test: verify install was successful
Test: adb install --non-staged com.android.virt.apex
Test: verify install failed with "does not support non-staged update"
Merged-In: I27ac000207e1b6ec39890bd382b3751fbb62e265
Change-Id: I27ac000207e1b6ec39890bd382b3751fbb62e265
(cherry picked from commit d2d672e3)

Fix: 290719741
Test: Manual. Repeatedly interrupt and continue touch exploration to trigger sending ACTION_HOVER_ENTER. This would previously cause a crash.
Change-Id: I7c35172c82c29289cac565a704bc5d4fbd579773

After restoring the data, we need to update the subscription
database cache.

Fix: 269062965
Test: atest android.telephony.cts.SubscriptionManagerTest#testRestoreAllSimSpecificSettingsFromBackup
Test: Boot up + Basic phone functionality tests
Merged-In: I2f8cb6418ee661a3cd8c02e3db6733cb12817d01
Change-Id: I2f8cb6418ee661a3cd8c02e3db6733cb12817d01

Context:
  The list of Preloaded drawables is defined in
  /system/framework/framework-res.apk. It describes a set of
  drawables to be loaded as the cache. The cache includes the
  drawables states for each preloaded drawable.

  For example: we have preloadSet(
          res/drawable/a.png
          res/drawable/b.png
          res/drawable/c_vector_drawable.xml).
  Because both of loading png and decoding xml to be a vector
  drawable take I/O and parsing time, in order to prevent app
  processes from doing again, the drawable state describes where is
  the decoded image of Bitmap in memory and the data structure of
  VectorDrawable.

  When some process wants to load R.drawable.a, it intends to load
  res/drawable/a.png, it's already in the cache because Zygote has
  preloaded as the drawable state for res/drawable/a.png. The
  Resources clones another drawable state and creates a new Drawable
  instance back to the app.

Solution:
  ResourcesPerfWorkloads sets a special version of the framework
  resources package but it doesn't rebuild the drawable state cache
  according to the special framework resources package.

  When the drawable state cache belonging to the old framework
  resources package doesn't match the new framework resources
  package, accessing the preloaded drawables may trigger this kind
  of issue.

  The solution is to clear the static caches in ResourcesImpl and
  load the cache defined in the specified framework resources
  package.

This patch also moves the preload resources tasks from ZygoteInit.java
to Resources.java.

Bug: 193826688
Test: atest ResourcesPerfWorkloads

Change-Id: I0fa25612bf82dd810d1d9443068a9a302a61d53d

Bug: 293151120

All OWNERS of android.database.sqlite are now owners of
android.databse.

Test: N/A
Change-Id: Ia0199e57299aff129b1a1dc8415f7e6908033391

Bug: 289414897
Test: it builds
Change-Id: Iaaa94da72df680406138364e06865c768b79edef

We should not traverse all leaf taskfragments when app create task fragment but is in freeform mode.

Bug:292868273

Change-Id: I99aa05bf814009948e5ff8c58d1ebb26116c3b60
Signed-off-by: tanxiaoyan <tanxiaoyan@xiaomi.com>

Since we added AIDL name tracing globally, it's been a very powerfull
tool to root cause issues in the field. One gap is non-autogenerated
AIDL interfaces like ContentProvider and Cursor. This was explicitly
disabled in aosp/2254719 to workaround some unrealistic perf startup
issue. Given the gap in identifying more important issues in the field
it's worth enabling this so we can fix 'real' problems.

Enabled caching of the non-autogenerated AIDLs and also added the
'AIDL::java' prefix to the name of the non-autogenerated AIDL names
to be consistent with the autogenerated ones.

Bug: 288578300
Bug: 270044534
Test: atest AidlTest
Change-Id: I96633ec976db01991620b43d7a7e9528ea22de1f

Define Thread hardware feature FEATURE_THREADNETWORK = "android.hardware.threadnetwork"
so that apps/users can know whether the device has Thread support.

This flags is defined as public API but not system API because we
defined TRANSPORT_THREAD as public API and general apps can access
Thread network by adding callback for the Thread network. Adding
the FEATURE_THREADNETWORK as public API allows apps to be more
clear/confident of whether they can retrieve the Thread network
with TRANSPORT_THREAD.

FR: b/235016403

Bug: 284296494
Bug: 292948264
Change-Id: I3bce3458fa721a20d8937a0c70eec2dbc7c54bc6

The storage permissions of some system applications have been authorized by default and cannot be modified by users with SYSTEM_FIXED or POLICY_FIXED flags. Once revoked due to an self-upgrade package, the user cannot re-grant.

Bug: 292732635
Test: manual
Change-Id: Ifb2480fa05c281dcdf9129b453f168ed68ebfc5d

Merge "Fix the concurrency issue when 'linkToDeath' in AccountManagerService that leads to JE" into main

Bug: 280547417
Test: boot pixel and cuttlefish
Test: atest SELinuxMMACTest
Change-Id: I0b407ad8b95278453d5b917e0da7024b245062b9
Merged-In: I0b407ad8b95278453d5b917e0da7024b245062b9