Implementing the new API for A11y services to take the screenshot
of the specified display.
Bitmap takeScreenshot(int displayId)
1. The main codes is moved from the UiAutomationConnection class and
let the UiAutomation change to use this API.
2. Add a capability at metadata to check whether the A11y services could use this
API or not.
3. This API is a async one for A11y services, but is a sync one for
UiAutomation.
Bug: 10931661
Test: a11y CTS & unit tests
Change-Id: I478bd93c60d4742bef20ae0e423ca9de1bda55d2
To support embedded hierarchies (leash):
1. Add new APIs for leashed child in AccessibilityNodeInfo.
2. SurfaceView needs to use this API to add such child when
initializing the AccessibilityNodeInfo.
Bug: 137593247
Test: a11y CTS & unit tests
Change-Id: Ieaa57c6ad65f4c199175d929c3bcba7fb7e6ef26
Input now tells the app when it has focus.
Touch mode information will now also be sent to the apps through input.
When a key targeting a non-focused display is processed, run the
'moveDisplayToTop' synchronously to ensure that setInputWindows with the
correct window focus information is completed. This would ensure that
the focus event is enqueued before the key event.
Bug: 70668286
Test: atest WindowFocusTests
Change-Id: Iff0b88a05441b29834741aa3dfae31d55871ddd6
As otherwise they may be out of sync, leading to shifting when the
IME disappears.
Bug: 111084606
Test: InsetsSourceTest, InsetsStateTest, InsetsSourceProviderTest
Change-Id: Ifd7dfa6694efccf8693fd46bec1a9dea879790ff
See WindowInsetsAnimationCallback.onPrepare for detailed
description of new behavior.
Also rename InsetsSourceConsumer.setVisible to setRequestedVisible
to communicate that this is client intent vs. current state.
Test: windowinsetstest application
Bug: 111084606
Change-Id: Id35c60e0f59a8aa4f0d300220391d1e2b96a91fd
A first take at the TaskOrganizer API for allowing SysUI to control task presentation.
In this CL we introduce the first two primitives:
1. The interface itself for implementation by SysUI
2. Support for organizing a given windowing mode (but atm really only PIP)
We include a sample app that manages the PIP from an APPLICATION_OVERLAY window.
Bug: 139371701
Test: wmtests/TaskOrganizerTests. TaskOrganizerPipTest
Change-Id: I44a8ed311bc5f06285bba2c6ff3b37a7d19a9190
CL[1] introduced starting input with CompletableFuture for improving IME
performance.
Since CompletableFuture is not recommanded way to use in system server side for
memory usage concern, we use java ExecutorService and Future to remove
this dependency.
[1]: I6aa4a664cfd0c86f75cee2457715317194bbe5e2
e0172102b9
Fix: 147331480
Test: atest CtsInputMethodTestCases
Change-Id: I10391b834c33458c3e0ac846ab738e805d235c26
Update the leash of InsetsSourceControl and send it to the client while
the insets source window is moved.
This CL also guarantees that the array of InsetsSourceControl send to
the client side doesn't contain any null element.
Bug: 118118435
Test: atest InsetsSourceProviderTest InsetsStateControllerTest
InsetsPolicyTest WindowStateTests CommandQueueTest
RegisterStatusBarResultTest InsetsFlagsTest
LightBarControllerTest RegisterStatusBarResultTest
ViewRootImplTest DisplayPolicyLayoutTests
DisplayPolicyInsetsTests DisplayPolicyTests
TaskSnapshotSurfaceTest NavigationBarColorTest
Test: Rotate Camera in the new insets mode
Change-Id: I96d7ac5c37abbb472dab4cd2c24644e136d23947
Re-route ime-control to SystemUI when it isn't driven by an app.
This allows multi-window system-ui components to synchronize
with the ime (eg. adjusting for split-screen).
Ime control goes through a new interface IDisplayWindowInsetsController.
This gets set on WM and there is only 1 per display. All of this
is currently handled in DisplayImeController which will also
drive the animation of the IME and dispatch to ImePositionProcessors.
On the server-side, InputMethodControlTarget is separated from
InputMethodTarget so that IME can be controlled by a different
client than the IME target.
Bug: 133381284
Test: Existing IME tests pass
Change-Id: I8e8ed2e09c45998c228df72e52a671fa327308f2
- IME insets always get reported through WindowInsets.getInsets(ime())
- However, getSystemWindowInsets will not report it if ADJUST_RESIZE
isn't set.
- Fix bitmask check for ADJUST_RESIZE
Test: InsetsStateTest
Test: WindowInsetsTest
Fixes: 146465040
Bug: 111084606
Change-Id: Ib19c89050af4f669ddda14d2bf2415aa3b5092c1
Setup for P/H experiment by switching to WindowManagerGlobal.
Follow up patch will set the flag based on DeviceConfig
Bug: 147096935
Test: build, boot, manual
Change-Id: If062fc16a90e0f3069064b969e1a63a6142421df
Existing annotations in libcore/ and frameworks/ will deleted after the migration. This also means that any java library that compiles @UnsupportedAppUsage requires a direct dependency on "unsupportedappusage" java_library.
Bug: 145132366
Test: m && diff unsupportedappusage_index.csv
Change-Id: I288969b0c22fa3a63bc2e71bb5009fe4a927e154
- Only consider client state if new API is actually turned on.
- We only sent the state the client was actually controlling.
However, we then later check for visibility for state that may not
even exist. Thus, we keep the state but only update the sources
the client is actually controlling.
- Initialize source with default visibility. This prevents
issues where the source may not exist but we still check it.
This fixes issues where a letterbox was accidentally placed on
screen because WM was thinking client requested fullscreen flag
(hiding the status bar), and other places where we checked
requested visibility.
Also renamed client state to requested state.
Test: Open any app, open IME
Bug: 111084606
Change-Id: Ibead561fc5593d8944400320f5e31dbe262612fe
Isolate the usage of SkMatrix within the graphics module and
replace external usages with a stable C API.
Bug: 137655431
Test: CTS presubmit
Change-Id: I058051a6afb4ae6d48ecdcf5cf8bd21aa0995e90
This adds support for a custom bugreport handler. When the option is
enabled instead of immediately triggering the full bugreport, only a
CUSTOM_BUGREPORT_REQUESTED broadcast is sent.
Bug: 146052311
Test: built and tested locally
Change-Id: Ifea9906c15e0bc1281192a90b4dbdb0ad1ecfbdd
1. Only print log when debug is on.
2. Callback be notified when input does not start.
3. Avoid using hard code tag in Log class.
4. onCreateInlineSuggestionsRequest() do nothing in NOP.
5. Add missing javadoc.
Bug: 146525448
Test: manual verification
Change-Id: I41e1de92ffcdb8020aef99acbfec274e0294bad3
Some apis in class, WallpaperManager and Display, are not visible in
tests, make them visible for tests usage.
Bug: 142845271
Test: atest android.display.cts.DisplayTest#testGetSupportedWideColorGamut
--rerun-until-failure 10
Test: atest android.app.cts.WallpaperManagerTest#testWallpaperSupportsWcg
--rerun-until-failure 20
Test: atest android.app.cts.WallpaperManagerTest#testWideGamutWallpaper
--rerun-until-failure 20
Change-Id: I1958df1e53aa7f65a2eb9ee7fe2a6b098ac2ab39
When the SurfaceView is invalidated (e.g. geometry has changed) we ask
the ViewRootImpl to render the next buffer in to a BLAST transaction which
the SurfaceView will also use from it's RT callbacks in place of deferTransactionUntil
Test: Flip USE_BLAST_BUFFERQUEUE flag. Run SurfaceViewSyncTests.
Bug: 146598493
Change-Id: I43d301f25101afae32ad1c43a3a3210c5aeadd0f
Should be a little more clear. The internal IWindowSession implementation
keeps the WindowlessWindowManager name, but the API section definitely
needs a more usable name. Instead of passing in a root Surface we want it
to be owned by the SurfaceControlViewHost itself. This way we can ensure
we only return a SurfacePackage. In follow up CLs we will add an
accessibility ID to the SurfacePackage so we can ensure the Control
is passed around with the accessibility ID.
Test: Builds
Bug: 134365580
Change-Id: I30520e8b169d3744a66ccedf61cc4515cfef4dee
