The selection and composing spans are not parcelable. This patch makes
Content Capture can pass those metadata to the service.
Bug: 184311217
Test: atest CtsContentCaptureServiceTestCases
Test: atest FrameworksCoreTests:ContentCaptureEventTest
Change-Id: I455ee61998bfb19476f84c69b1a671113b6b6a98
Stored state should be cleared here.
This fixes a bug where the views cannot be re-translated after
UiTranslationManager#finishTranslation was called. Some of the state was
cleared, so the translations cannot be re-shown. And that state cannot
be recreated because we find the same ViewTranslationResponse stored and
do not recreate the other state.
Fix: 189892928
Test: manual - translate an app, go to the home screen (triggers
finish), go back to the app and re-translate
Test: atest CtsTranslationTestCases
Change-Id: Ic81fbce7ee09eb72352071c8bdd94a2731826009
Because the tooltip is part of suggestions on keyboard, it should be the
same as keyboard that is not magnified by the accessibility magnifier
when the magnifier is full screen mode.
Bug: 183462449
Test: Manual
Test: atest CtsAutoFillServiceTestCases
Change-Id: I3cfd84357b556a85c2788776d3966cf94d4a981d
The spans were removed by ag/13819626. Adds them back into the
ContentCaptureEvent. And removes unnecessary Map data because we
can get it via travel events. The total of events is not large.
Bug: 190035037
Bug: 184311217
Test: atest CtsContentCaptureServiceTestCases
Test: atest FrameworksCoreTests:ContentCaptureEventTest
Change-Id: I5fa1f7a6b51fd050a63731543cca98972c294b22
If the frames of insets source window and the frame of an app window
are changed in the same layout pass, we should dispatch the insets
calculated from the new insets source frames and the new app window
frame. So when the app frame is not up-to-date, we defer dispatching
insets to the app.
Fix: 160732586
Test: Add logs to check if we dispatch the correct insets to apps while
rotating the display.
Change-Id: Ic9e71d8eabde51067b740b35404ae7c9b68228d7
These traces get expensive in aggregate due to an increased number of
notifyContentCapture events. Remove them.
Bug 188836620
Test: Confirmed no notifyContentCapture trace points in systrace
Change-Id: I5d77bd9048b2e8cd267187f866dc095792a3f455
There is no transition for app launch from recent in live tile so we
have to add a method for Launcher to trigger the nav bar animation for
this case.
Bug: 189278432
Test: manual - launch app from recent in live tile mode
Change-Id: Ia63964491449e6f634350cbc8599cdec42d76d06
When bugreport dumps Activity information, it will dump translation
related information. But it adds -v parameter, the bugreport will
only dump visible task's top Activity. If message apps go to the
background, we cannot find our expected information. We try to let
translation dumpsys to dump the last Activity information if the
Activity isn't destroyed yet, e.g. go to background.
Bug: 187889345
Test: open chat app and back home then dumpsys translation. Let
Activity destroyed and make sure no Activity dump.
Change-Id: Ie5b935696a74ca3f298c5e6e5994fa659389e198
All input events coming into inputfilter from
InputDispatcher will now have FLAG_INPUTFILTER_TRUSTED set by default.
It is now up to the inputfilter (a11y) to remove this flag when appropriate.
One such case is where an unknown a11y service is injecting the events.
From auditing the a11y code, the only such place is MotionEventInjector.
Therefore, label all events coming out of MotionEventInjector as
"untrusted" by removing this new flag INPUTFILTER_TRUSTED.
Separately, a11y should be distinguishable from regularly injected input
events. To tell apart the events injected by a11y, add a new flag,
FLAG_INJECTED_FROM_ACCESSIBILITY. This flag requests InputDispatcher to
apply the device id = -2 for the injected events.
Test: atest CtsInputTestCases:android.input.cts.GamepadWithAccessibilityTest
Test: atest MotionEventInjectorTest
Bug: 175069843
Change-Id: I12d4a7bd6fbab8af202f5ae88b6be97ff9e1754c
Whenever an app makes a Text toast, this gets routed through systemUI
starting from R. Two A11yManagerClients are registered with
A11yManagerService for every toast. Since clients aren't removed until
the process is killed, there are a bunch of clients from systemui
hanging around and potentially cause the system to crash.
We add a way here to explicitly unregister a client
Test: build and flash,
atest android.view.accessibility.AccessibilityManagerTest
Bug: 189010828
Change-Id: I3f71bac475b4539c7be5b328adeb0be35a0b5543
This code was used by ActivityView (which was deleted) only.
Test: manual (seahawk)
Bug: 179161778
Change-Id: I65544daac3bc2a280239bdcbd54dc726d604165f
If the surface position is changed during the insets animation,
the surface position of the controls in the animation should be updated.
Otherwise, the cached position in the animation will overwrite the
newly-updated leash position.
Fix: 177925500
Test: Steps in the bug.
Test: 1. Open Window Insets Tests -> Window Insets Controller.
2. Click on the [IME INVISIBLE] button.
3. Click on the [STATUS BARS VISIBLE] button during IME animation.
See if the position of IME (Gborad) is correct.
Change-Id: I7441426a41019996651220e86ca90dd416ef9498
TaskSurfaceHelper(Controller) will be used as an interface to communicate
with the SurfaceControl of the underlying Task. This change adds a function
to set the game mode metadata for the task layer.
Bug: 186025682
Test: atest TaskSurfaceControllerTest
Change-Id: I136b65636b98e1883eaf9e4f4f0b34c61350d4e4
Apply a simple text alpha animation when toggling between original and
translated text. The text is fully faded out, then swapped to the new
text, then the fading is reversed.
Quick toggles are handled by ending the previous animation (which resets
the alpha value) when starting a new one. If the toggle is extremely
fast (<250ms for the currently defined animation duration), the text
stays in it's original state instead of swapping back-and-forth. This is
arguably not ideal, but anyway not worth the complexity for fixing.
There is an unhandled edge case where if the color is changed by the app
during the animation, the app's change would get overridden. It should
be rare and doesn't seem very important to fix.
Bug: 178651829
Test: atest CtsTranslationTestCases
Test: manual - across several apps, while scrolling, concurrently
interacting with the app, concurrently closing the app, selecting text
during the animation (with animation speed slowed down to see the
effect).
Change-Id: I08a26de2253bb345f01186a6748b2d0ff6c2a419
This can happen if UiTranslationManager#startTranslation is called more
than once for the same view. If the same TranslationResponse is
returned, it can simply be ignored.
This is preferable to ignoring duplicate ViewTranslationRequests, since
the Response is already cached. The TranslationService can cache
translations, so duplicate requests isn't an issue.
Ideally, the platform intelligence shouldn't issue multiple requests for
the same view.
Bug: 144292180
Test: atest CtsTranslationTestCases
Change-Id: I0bba4a19284a61a9c7a5b79c71b4f42ddffe4e32
Bundle doesn't override equals(). In TranslationResponseValue, the
Extras is used as a simple value type, so contents can be compared.
Bundle.kindofEquals probably works here, but this change just makes it
work for empty bundles for now, to fix a bug in animation.
Bug: 144292180
Test: atest CtsTranslationTestCases
Change-Id: I68daac8cafceb1aa446ee89ba1405595b22e6213
- Disallow callers from concurrently modifying the exclusion rects
list after it's set by making a copy.
Bug: 166594759
Test: atest android.view.cts.SystemGestureExclusionRectsTest
Change-Id: I80183cf85d16ac669320cfda9ec9deed9e348e33
Some recycled views cached its layout and a relayout is unnecessary.
In this case, system still needs to notify content capture the view
appeared.
Bug: 177687849
Test: atest CtsContentCaptureServiceTestCases
Test: atest ContentCapturePerfTests
Change-Id: I480436c4fc3508b9eed485421a1cc778d55d5dde
In performTraversals, "params" may be null. In this case we were
not passing params to the WM indicating no changes (an obscure
artifact of the protocol). In R the WM would simply ignore this
however our updateOpacity logic in S is treating null as "translucent"
it seems easiest to just pass mWindowAttributes directly avoiding
the null case. While the opacity was broken, it seems an animation
was added to StartingSurfaceDrawer which depends on having a translucent
window. We also update this code to request a transparent pixel format
as would have been required in the R API.
Bug: 186555833
Test: Existing tests pass
Change-Id: I180bf1796bbb4b7049ae9811f674f31014c2a400
The issue was that: if navigation bar is requested to show while a
controller is playing an animation of hiding both status bar and
navigation bar, the animation will be canceled and will stay in an
intermediate state, and then, another controller will play an animation
of showing navigation bar. At the end, status bar will stay visible but
it should be invisible.
This CL fast-forwards the insets animation to the end state if it is
canceled.
Fix: 185459811
Test: Steps in the bug.
Change-Id: Id6ae321b6c153ed1d7155afd6111b193ac9c045c
