> This will make it easier to migrate to dagger as it will
increase the number of constructor parameters significantly
Test: presubmit
Bug: 361850561
Flag: EXEMPT dagger migration
Change-Id: Id06246ec74c678d7614f4c6a798c5644b40a0545
Splitting allows easily making a fake that doesn't have to override the implementation details of LauncherPrefs. The fake should not deal with SharedPreferences.
LauncherPrefs could be an interface, but then the companion object has more limitations. The solution there is to have a dedicated object class, e.g. `LauncherPrefItems`. I went with an abs class to make the refactor simpler.
Flag: EXEMPT refactor
Bug: 369641781
Test: go/testedequals
Change-Id: I97a2d495d3b5fa892fa53a11fb3f7a7dfb98515b
This makes the behavior same as instrumentation test
Bug: 370080120
Test: Presubmit
Flag: EXEMPT test fix
Change-Id: Id4183c91f3951748b657b583fde8860abe55d3a8
Using same key for blob store and content provider based layouts
Using a single key for layout provider. This ensures that if one
test doesn't cleanup properly, any followup test which relies on
model initialization overrides the key appropriately
Bug: 370080120
Test: Presubmit
Flag: EXEMPT bugfix
Change-Id: Idef0a89a5d1ec89e24aae8a4549fd122634dad2f
* changes:
Cleaning up unusued handlers
Moving RecentsWindowManager away from the singleton pattern
Limited recents in window introductory cl
abstracting fallback views to support container instead of activity
> This avoids duplicate RPCs when trying to get multiple properties
from AppInfo
> This would be used in a followup cl when AppInfo is added to caching
logic in IconCache
Bug: 366237794
Test: atest ApplicationInfoWrapperTest
Flag: EXEMPT refactor
Change-Id: I55b964d4f8cfa1ff1770e310ac278719495e285d
package validation happens during loader anyway. And using MATCH_INSTALLED
for checking cross user apps may not work all the time
Bug: 363324203
Test: atest GridSizeMigrationUtilTest
Flag: EXEMPT bug fix
Change-Id: Id36e8fdff132242705ef53c85dc6567fbde5d952
Test: Built and tested locally
Flag: com.android.launcher3.enable_fallback_overview_in_window
Bug:292269949
Change-Id: I5352ba0b6c5bc196fbd1322d435a7e27e884f7b5
This CL forcefully resets taskbarModePreference on tearDown() to ensure
the following tests are not affected by the changes made by
`testTaskbarPinningChangeInLockedTaskbarChange`.
Bug: 368322666
Flag: EXEMPT bug fix
Test: nexus_unit_test_multi_device_platform
Change-Id: I8bd8e8ca8b4b832ed17e86fbdfc9bda7bf9ab16e
This CL allows freeform-first devices to lock the taskbar in a pinned
state when Launcher is visible.
Bug: 361419732
Flag: com.android.window.flags.enter_desktop_by_default_on_freeform_displays
Test: DisplayControllerTest
Test: manual
Change-Id: Ic0d2740e33fee19c676d88aa2bcf370eb64366e8
FloatingHeaderviewTest was named with ...Tests (plural), preventing it from
being excluded from launcher-testing-helpers-robo and getting included
in some unintended tests.
Bug: b/341795751
Flag: EXEMPT test renaming.
Test: Running FloatingHeaderViewTest
Change-Id: I8149f0a5548e77674bb2a23a34bd7671e0a3b371
Platform build can now have versions line "eng.himans"
Bug: 187761685
Test: online regex verifier
Flag: NONE test change
Change-Id: I86b9324dfbbc36272215eae58af836e9a3741d70
- All taskViewIds become obsolete after removing and re-adding, so we'll need to reset mTopRowIdSet when removing all taskViews
- OverviewActionsView.updateActionButtonsVisibility can be called before updateDimension (where mDp is set), so a null check is necessary
- Fixed TestStaibilityRule failing when running FallbackRecentsTest alone. FallbackRecentsTest runs with 3P Launcher, in that case launcherPackageName is our test activity, and launcherVersion can be null. This can only occur in a local run, thus return LOCAL in this case.
Fix: 355042336
Test: FallbackRecentsTest and manually testing dismiss
Flag: EXEMPT BUG_FIX
Change-Id: I9125ea38a567deb0d3028e7d7407397d1dc06ade
