According to traceview, most of the time spent in filterByLanguage
is calling getLocaleObject to create java.util.Locale objects.
The solution is to cache the object to save the cost. Note that both
mSubtypeLangugageTag and mSubtypeLocale are final, so it is fine to
cache the locale object.
Test: bit FrameworksCoreTests:android.view.inputmethod.InputMethodSubtypeTest
Test: bit FrameworksCoreTests:com.android.internal.inputmethod.LocaleUtilsTest
Test: Try to switch user back and forth, and tap a textview to show IME.
Bug: 37647204
Fix: 37647213
Change-Id: Ib6ff7e97b17fc547c8109af8177f05fc3ea41b08
Bug: 64269689
Test: m checkbuild
Test: bit CtsGraphicsTestCases:*
Test: bit CtsViewTestCases:*
Test: bit CtsTextTestCases:*
Change-Id: Idc9f2b4612741be949204cd06efc41e1738d6587
Add black background behind letterboxed apps on displays with tall
aspect ratio. This is needed to make sure that other windows won't
bleed through status/navigation bar area, because letterboxed
windows are not considered fullscreen.
Bug: 63156733
Bug: 62893418
Test: go/wm-smoke
Test: Trigger animations to/from recents with different rotations.
Change-Id: If0eac9a3801518185d560a87d9f88ee0aafa3a9c
Consider this VirtualDisplay (VD) scenario:
HostActivity creates a VD which holds SettingsActivity. When EditText
on SettingsActivity is tapped, it gains focus.
On eventual taps, it loses focus i.e. the Window in VD loses focus and
the host activity in primary display gets the focus instead. This
happens because WM's TaskTapPointerEventListener.onPointerEvent()
is called on the default display only.
Root cause:
1. Tap detector isn't registered for non-default display.
2. Tap detector has no info on which displayId touch was received.
3. InputFlinger doesn't deliver InputMonitor events for
non-default displays (fixed in a separate CL)
Fixing above results in onPointerEvent(MotionEvent) to deliver the
Touch events successfully to VD. We restrict these changes to physical
multi-displays and VR VirtualDisplays (which uses virtual touch device).
[VrManagerService calls WMInternal.setVr2dDisplayId(int)]
In future, displayId should be part of InputEvent. Bug: 64258305
Bug: 62033391
Test: bit FrameworksServicesTests:com.android.server.wm.DisplayContentTests
Change-Id: I3626f4de5aa9bcf905da9abd39f3ab1baefc4c48
Instead of ANR in an infinite loop, this will detect
a cycle and return null.
Bug: 62960867
Test: Added in CTS FocusFinderTest#testChainVisibility
Change-Id: Ifb4638e3350642f065c40b3bea112cc12efff447
We need to omit it from screenshots, and screen mirroring, so SurfaceFlinger
will have to have some knowledge. Eventually we'd like to do this by
having everything else have a parent, and mirroring and screenshotting will
operate on that parent. Since we aren't comfortable doing that for DR though, here's
a quick and dirty hack to tide us over with low risk.
Bug: 63311708
Test: screencap, screenrecord, android.view.cts.SurfaceViewSyncTest
Change-Id: I2b43be00567ee19cf27aeef344e5415f4a69297f
Following four steps with RecyclerView - CardView - ImageView:
a. fade in imageview in onBind:
calling imageView.setTransientState(true)
b. RecyclerView animate the item: calling
cardView.setTransientState(true)
c. fade-in finishes, calling imageView.setTransientState(false)
d. RecyclerView animation finishes: calling
cardView.setTransientState(false)
After these four steps, RecyclerView unexpectedly has transient
state.
The problem is in step b, when calling cardView.setTransientState()
it incorrectly calls parent.childHasTransientStateChanged(this, true)
which causes RecyclerView's mChildCountWithTransientState increased
to 2. And it's decreased to 1 in step d and stay as 1 forever.
The child should only call childHasTransientStateChanged() when
actual hasTransientState() changed.
Bug: 64235615
Test: ViewTransientStateTest
Change-Id: I99ed35cc9c49e54d36590d8f1d206501fd3288f2
FYI: We have plans to reintroduce this in some form in the near future.
Fixes: 62571798
Fixes: 64096360
Test: bit FrameworksCoreTests:android.view.textclassifier.TextClassificationManagerTest
Change-Id: I864d6c1143af6a2485cabdfe85116b5922a3e315
This CL logically reverts a compatibility hack I introduced [1] for
apps that had relied on @hide method SurfaceView#setWindowType(int).
[1]: I5217f6417a73690ae8a978754218b7b089070fdd
3b5011afc9
Fixes: 62054282
Test: Manually verified that Bug 36345857 is still not reproducible
Change-Id: Icee198c554de558cfa4ffe0b264064969839654e
The presence of these new flags leads to issues with application that
do not expect their presence. Since these flags can appear at
critical times, such as on orientation change, these issues are
brought to the surface often.
This CL remedies this problem by first removing the rotation
property. It is not used and the original issue of orientation and
Configuration alignment has been addressed. For app bounds, the CL
reverts the behavior back to identifying diffs as a screen size
change.
Fixes: 64004417
Test: bit FrameworksServicesTests:com.android.server.wm.AppBoundsTests
Test: go/wm-smoke
Change-Id: I1fabb564dfb5c13d897336708523cf7cd5099fa0
Allow developers to easily add their own debugging info
when developing custom ViewGroups.
Bug: 30117235
Test: tests still pass
Change-Id: Ia83005b4dc727c7b959a50e7b0130582f6e89762
Bug: 37858096
Test: Adding a CTS test to verify an exception when using a
node after recycling it. Ran a11y CTS and unit tests.
Change-Id: I5486a7abec681cb6fbb315ff87f0915d5505b8f3
