KeyphraseEnrollmentInfo is used for testing the system enrollment
application support within GTS.
Bug: 153264697
Test: gts-tradefed run gts-dev -m GtsAssistIntentTestCases \
-t com.google.android.assist.gts.KeyphraseModelManagerTest# \
testAllSystemKeyphrasesAndLocales
Change-Id: I66558cf849f2815f2c917d128f9f9cc0673750ef
Merged-In: I66558cf849f2815f2c917d128f9f9cc0673750ef
While debugging SurfaceView transaction usage I noticed this
code was unconditionally dead. Notice the else statement
only contains the single if statement within, so it could
be written as else-if. But then notice the condition of this
rewritten else-if is contained within the original if and
so it can never be entered.
Bug: 153120755
Test: Existing tests pass
Change-Id: I032b17911a9c71a1e3d09e3933e6f6a3d86dffb6
There's no need to keep it synchronous
Bug: 154304487
Test: atest NotificationShadeWindowViewTest
Test: atest WallpaperControllerTests
Test: manual
Change-Id: I547e05fb4e2a5f2d30cfc5de354c36cd9070e65b
Below are some test runs that show 100 iterations of getInstalledPackages ran
at activity startup of touchlatency app to demonstrate the perf cost.
R Baseline:
https://screenshot.googleplex.com/PnVi5APraJohttps://pprof.corp.google.com/?id=dc105ca9592f8367c381e2dcf7f1847d
R Optimized:
https://pprof.corp.google.com/?id=2d1d9fadc20bcc61d41bd52116bd94e3https://screenshot.googleplex.com/OAwJnvwPqPH
Savings: 8% of system_server binder response time when the call is executed.
In terms total binder transaction time. Savings are 2% of the total
binder call time.
Exempt-From-Owner-Approval: CP to correct branch for R
Bug: 153656459
Test: atest FrameworksServicesTests:PackageParserTest
Test: atest FrameworksServicesTests:PackageParserLegacyCoreTest
Test: atest FrameworksServicesTests:ScanTests
Test: atest FrameworksServicesTests:ParallelPackageParserTest
Merged-In: I2de9cf1f754a505239d4416e1fc70bf77932c5db
Change-Id: I2de9cf1f754a505239d4416e1fc70bf77932c5db
(cherry picked from commit 7cb9e5f671)
In Q, app developer would've to wait for InputConnection on Editor before
showSoftInput() could be called. This made API very un-intuitive. While
WIC.show(ime) took care of some of the scenarios, it didn't handle
showing IME from onCreate() i.e. when control is not yet available.
This CL introduces a pending IME show request that waits until IME
control is available.
Test: atest WindowInsetsControllerTests#testShowImeOnCreate
Bug: 153014086
Change-Id: I93dd5790e1fb3d7841ad1323a45cd3bae2d6e9b8
Use an explicit intent and set PendingIntent.FLAG_IMMUTABLE to prevent
someone from modifying the intent from PendingIntent.send(...).
Bug: 153356209
Test: atest AdbNotificationsTest
Test: In bug, install and launch the PoC apk and give it notification
permissions. Then, with USB/Wifi debugging enabled, disconnect and connect
the device to create the adb notification. the PoC apk should not have
permission to display information from
content://com.android.settings.files/my_cache/NOTICE.html.
Change-Id: Ie49aa3cf9b33168cf1435fc2427e95aac7f4609b
(cherry picked from commit 2c03881459)
Exempt-From-Owner-Approval: approved in master
To have a way to verify behavior of more directories, mark more
directory getter APIs @TestApi.
Bug: 148664742
Test: atest SysConfigTest#testStageWhitelist
Change-Id: I2d22412935b24ec4c33d8d1b3a7fbb7df4a08841
AutoCorker addresses the situation where big invalidation storms kill
performance but we don't have a way to insert a manual cork around
these update storms.
Bug: 140788621
Test: m
Change-Id: If07d693886fca340c7a18d5a607a4f235aa7107d
Instead, the click listener is set on the app opp container since
clicking on the app opp container results in the same action as clickng on the
individual app icons. As per a11y guidance, only set the click listener
on the container.
Test: manual
Fixes: 153281363
Change-Id: Ic1c2bdb2f77fc459a3a3e997ed494b4bab36a273
In ImeFocusController#onPostWindowFocus -> onViewFocusChanged,
it implies that the view has focus after onPostWindowFocus, this logics
we keep leveraged from IMM#onPostWindowFocus -> focusInLocked() long
ago, which may not always be true as SearchView layout may call
View#clearFocus when size changed after IME switcher shown.
When IME switched and back to app activity, in the above call path,
even activity window focused, and set the fallback focus view as
activity's root view, but the root view is not actually focused,
if we set this view as the next served view, then calling
View#onCreateInputConnection will return null, because it's not an editor.
Use correct view focus state when calling onViewFocusChanged.
Fix: 153612876
Test: manual as below:
1) Launch Files app, taps SearchView EditText
2) switch IMEs with IME switcher dialog
3) see if Password keyboard shown.
4) keeps 2) and 3) several time.
Expected: There is no password keyboard and
the keyboard input is still workable.
Change-Id: I68bb95fc3cbfe1f5992ccf87694b34c3e52bb31f
- First touch on TextView blocks secondary touches on handles.
- First touch on handles blocks later touches on TextView but doesn't block secondary touches on other handles.
Bug: 150995597
Test: manually tested & automation tests:
atest FrameworksCoreTests:EditorCursorDragTest
atest FrameworksCoreTests:TextViewActivityTest
Change-Id: I7717fc061fc81514fc1dad0d3acbc73e683516cf
RemoteViews is public API used out of scope of widget. The correct place
to call noteAppWidgetTapped is in AppWidgetHostView.
Fix: 153676411
Test: manual test, tap a widget, "adb shell dumpsys usagestats | grep
USER_INTERACTION" to oberserve USER_INTERACTION event sent to UsageStas, "adb shell dumpsys appops | grep appWidgetVisible" to observer appWidgetVisible flag.
Change-Id: Ic473211b91fd952dbb81b09b1e1568d6f69a0dd8
Map the camera2 CONTROL_ZOOM_RATIO to legacy zoomRatio in Camera.Parameters.
Test: Force Pixel 2 HAL to HAL1, and run testZoomRatioTest
Bug: 149052011
Change-Id: I888c919ae10d3e2882f3df797524a06e2b3be259
Bug: 153103051
A binder cache recompute() function cannot compute a result based on any
data that is not known to the binder server. If the code depends on
local data that can change, then invalidation will not work properly.
The original getState() method returned OFF if the bluetooth service was
unavailable. This computation now occurs in the getStateInternal()
method, outside of the binder cache query() method.
The recompute() method converts RemoteExceptions to RuntimeExceptions.
Then, the conversion is reversed in getStateInternal(). This double
conversion is needed because the cache recompute() method has no throw
spec.
Test: Create a debug image that enables binder cache VERIFY. Run the
following tests:
* atest BluetoothInstrumentationTests
* atest PtsChreTestCases
* atest UserLifecycleTests
* manual testing connecting to bluetooth devices and toggling airplane
mode.
No cache inconsistencies found. No test failures seen.
Change-Id: I93b9742587c4eb695d9a11fc6ab145f6a40a0ece
documentation to refer to BlendMode.SRC_IN
instead of PorterDuff.Mode.SRC_IN
Fixes: 151952140
Test: N/A
Change-Id: Ie4b9a585c2dae49e633f6741180b6749493d069a
- Allow targeting task display area in ActivityOptions.
- Change LaunchParamsController to select preferred task display
area instead of display.
Bug: 152116619
Test: WM CTS and unit tests
Change-Id: I16ab96008e1ce03422f89a0609d3121051261fda
