This CL only implement the handwriting initiation for milestone 1:
a) handwriting is only initiated for a focused editor.
b) user must start handwriting within the editor's boundary.
Bug: 209854913
Test: atest FrameworksCoreTests:HandwritingInitiatorTest
Change-Id: Iab6d83f58aa74f0af8091731c80bbdb2eb29c024
Some apps or games could use the combo keys such as holding Ctrl or Alt
to trigger its own shortcut or behavior. To prevent the system shortcut
might be handled before dispatching to the app, we move these keys
handling to 'dispatchUnhandledKey' when the app didn't handle.
This will also fix the corresponding state when the modifier key is
released.
Bug: 156619364
Bug: 213096704
Test: atest AppKeyCombinationsTest
Change-Id: I42770045c61ad6396b27132b8a61d34fc121d77e
Since we now support long edge cutout, the description of this API is
out of date.
Bug: 211939507
Test: make
Change-Id: I04c4c34e56c4e8b8d466e05cf60e89da31d6cc9b
Alternative approach to fab15e55446080bdcfc05ba315e8ef914b0a6f65 which
was racy because the disconnect callback is called without the
BQ lock and the client can continue to modify the BQ state after
disconnecting from the queue.
This approach resets the BQ and BBQ states when the SurfaceControl is
updated. This solves one concrete problem of not relying on the old
SurfaceControl to be destroyed in order to release the currently presented
buffer back to BQ.
In addition this change resets the sync state in BBQ with the rationale
the system does not want to sync on buffers presented on an older
SurfaceControl.
Bug: 197269223
Test: atest BLASTBufferQueueTest
Test: labtest ag/16407859 cf-foldable * 3 (b/197269223#comment40)
Change-Id: I137d3284c4e9216963f9c672f192afd334501ee5
FrameCallback sends back the syncStatus and allows the caller to return
a FrameCommitCallback. This is so the caller can evaluate the results of
the sync and determine if it wants to continue waiting for a commit
callback. In cases, where the sync has failed, there will never be a
commit callback so the caller can avoid waiting.
This is partically helpful for VRI because it wants to sync with a frame
that will draw, but also doesn't want to get stuck indefinitely waiting
on the callback. VRI can check the syncStatus and only handle blast sync
and/or reportDraw when its determined that a draw will happen.
This allows VRI to remove frameCompleteCallback since it can now wait
until a buffer has drawn instead of when the vsync has completed. This makes
sure it waits on FrameDropped since that result indicates HWUI will draw on
the next vsync
Test: BlastSync
Bug: 200284684
Change-Id: Ic6e2f08ea21ac8a1634a3389c927fcb68ede3f7b
The default value for DISPLAY_DECORATION is false. If there's a new
SurfaceControl, we only need to set the flag if it's true. This avoids
applying an unnecesary transaction.
It also prevents a boot failure which is independently fixed by
Ib11d46439db57b90486bad07dd90f2cf0822182a.
Bug: 211797674
Bug: 211835607
Bug: 212402133
Test: boots
Change-Id: Iccb662f9a6312b547b3d28fa1f5c1e9ff5dce9eb
The first time this topic landed, it resulted in b/212402133. We avoid
running into this bug with Ib11d46439db57b90486bad07dd90f2cf0822182a.
Original commit message:
When passed to native, this flag will tell a Layer that it should use
Composition.DISPLAY_DECORATION.
Bug: 193170859
Test: manual
Change-Id: I7f1685eb7dc57271f532065dcd1d4dcc449c5cb0
When work profile disable by quick settings, the
user would be locked then we cannot query installed
input method lists for work user by existings API
InputMethodManager#getInputMethodListAsUser().
Adding a new API to query input method services
regardless of the user state. This API currently
only used by Settings, it shouldn't affect the
original behavior for other usages.
Bug: 210083408
Test: verify we can get input methods list after
disable work apps by quick settings
Test: atest CtsInputMethodTestCases
Change-Id: I54d5dbec7e76d6a68935007ed3af0641f717a7c5
We convert the InputFeatures flags into AIDL enums that will have values
generated in both native and java code so that the value only has to be
defined once.
Bug: 162194035
Test: build, presubmit
Change-Id: I8176265079292ee969c409ddd52676a52f208601
A stylus interceptor is a window that receives all stylus events within
its touchable bounds, while letting all other events be dispatched to
windows behind it. This makes it possible for the framework to create a
stylus interceptor on top of another app to implement handwriting
recognition.
The feature has no effect when the window flag FLAG_NOT_TOUCHABLE is
not set.
The feature has no effect when the window is not a trusted overlay.
Test: manual
Change-Id: Ia481ee8d831643bc99445b47659a7f5e9875c388
PositionUpdateListener callbacks maybe replaced before they
are applied. We need to merge with the existing transaction
so we don't drop any destination frame updates. But accessing
the previous callback transaction is unsafe since we might fight
with render thread.
Fix this by locking access to the transaction object.
This also fixes a potential SurfaceControl access issue where
we may release the SurfaceControl that is used by render thread
in the UI thread.
Fixes: 211090247
Test: atest SurfaceViewSyncTest
Change-Id: I28ed344754601169c6cefd919668e76ef5a467c3
Unless we are looking at stack traces (e.g. from strict mode) it's not
possible to identify which type of object is not being closed (most
methods are 'close' or 'release). Change the logged text to clarify.
Change-Id: Ib90eac716f43c2c2caf8d8c6fb64a7bd90562da9
Test: manual
This is a step to move the layout logic to the client side. We won't
have DisplayPolicy#layoutWindowLw then.
This CL also renames computeWindowFrames to computeFrames because we
already have "Window" in the class name.
Bug: 161810301
Test: atest DisplayPolicyLayoutTests WindowLayoutTests
Change-Id: I4162e39f0d2aa35141fbbc72e43decd2f3f491f0
When passed to native, this flag will tell a Layer that it should use
Composition.DISPLAY_DECORATION.
Bug: 193170859
Test: manual
Change-Id: I7f22dfd03d86786afee078a95f0af590be0bbdfd
Revert "Add tests to ensure hardware only for RuntimeShader API"
Revert submission 16475336-RuntimeShader_softwareOnly
Reason for revert: Breaking tests as system is dying
Error: Software rendering doesn't support RuntimeShader
Bug: b/211066630
Reverted Changes:
Id0ee4d1f0:Enforce that RuntimeShader is only hardware accele...
Ib4c2e54bd:Add tests to ensure hardware only for RuntimeShade...
Change-Id: Id7bbf2609a06c1cd69f6a457dab44f9478d95742
Changes:
- Listens to changes from the client coming through IActivityClientController#requestCompatCameraControl to ActivityRecord#updateCameraCompatState
- ActivityRecord#updateCameraCompatState sends updated state via TaskInfo to WM Shell
- ITaskOrganizerController#updateCameraCompatControlState to dispatch the user interactions with the control from WM Shell triggers callback to ActivityRecord#updateCameraCompatStateFromUser
- ActivityRecord#updateCameraCompatStateFromUser remembers the user's choice and asks client to apply treatment through ICompatCameraControlCallback
Feature is guarded with config_isCameraCompatControlForStretchedIssuesEnabled
Test: atest WMShellUnitTests:ShellTaskOrganizerTests, atest WmTests:ActivityRecordTests
Bug: 206602997
Change-Id: I083aa6718bd67456bedd9444e9b78740c041f870
