Messages sent to the embedder host may be one-way messages with no response
handler. If the host calls FlutterEngineSendPlatformMessageResponse on a
one-way message, then just delete the message response handle.
Also update the documentation to indicate that
FlutterEngineSendPlatformMessageResponse must be called for all messages.
Previously the docs implied that some FlutterPlatformMessage objects may
have a null response_handle. The embedder will now set a response_handle for
every message (even if the sender does not expect a response).
This patch works around Android's limitations on reflection. With it embedded
platform views that use virtual node hierarchy trees should be accessible on all
Android versions to date.
The workarounds in this PR are brittle. Ideally the methods would be made public
in Android instead so we wouldn't need to employ tactics like these to work
around the missing methods.
`AccessibilityNodeInfo#getChildId` is blocked from any type of reflection
access, but the underlying private member that the getter accesses,
`mChildNodeIds`, can still be reflected on. On Android versions where we can't
access the getter, this patch falls back on reflectively accessing the field
instead. Unfortunately this field is a
[`LongArray`](https://android.googlesource.com/platform/frameworks/base/+/master/core/java/android/util/LongArray.java),
a utility data class private to Android. So in this case we're reflecting to
both access the member and actually read data from it, since we need to use
reflection to call `LongArray.get(index)`.
`AccessibilityNodeInfo#getParent()` doesn't have any lucky available underlying
members. However, `AccessibilityNodeInfo` itself is `Parcelable`, and
`mParentNodeId` is one of the pieces of data that's written to a parcel via
`AccessibilityNodeInfo#writeToParcel`. So the fallback for that is to write the
node to a parcel and then read the parcel for the ID in question. This will
break if the implementation details of `AccessibilityNodeInfo#writeToParcel`
ever change. The details have already changed enough in the past to require two
sets of logic for reading from the parcel.
The engine RunBundleAndSnapshotFromLibrary API expects a bundle path directory
containing the application's assets. If the Android embedding is using
AOT ELF library packaging and does not need to extract assets, then create an
empty directory at the bundle path.
Fixes https://github.com/flutter/flutter/issues/34287
Currently, all our host unit-tests that have rendering concerns use the software backend because of OpenGL ES availability and stability issues on the various platforms where we run host tests. Unfortunately, entire subsystems are disabled (and not tested) when rendering with the software backend. This patch pulls in SwiftShader and via pending patches in the buildroot, configures the host unit-tests to optionally use OpenGL ES in a stable manner without relying on the OpenGL drivers being present (and functional).
I have wired up the embedder test fixture in this patch to use the SwiftShader based OpenGL ES driver. I will update the shell and runtime unittests in a subsequent patch as well. The on and offscreen surfaces are configured as 1x1 pbuffer surface because we should be able to write pixel tests using OpenGL directly wihout having to deal with surfaces.
This should satisfy the low-latency need of DevTools.
Test added:
* ReportTimingsIsCalledSoonerInNonReleaseMode
* ReportTimingsIsCalledLaterInReleaseMode
In some cases, the window needs to be redrawn without a resize. This
adds a callback for that case to trigger a repaint.
Since there's no embedding API for repainting, trigger it with a window
metrics event using the current window size.
Fixes https://github.com/flutter/flutter/issues/30731
Previously AOT compiled Dart code would be packaged as a group of assets
within the APK. This has been replaced by a single ELF library containing
the same data.
For https://github.com/flutter/flutter/issues/33807
We still need to make layers' children immutable for full immutability.
That will require us to change the SceneBuilder API to build the layer
bottom up instead of top down (post-order traversal instead of pre-order
traversal).
#9203 broke the keyboard_resize integration test(see more details in flutter/flutter#34085 (comment)).
This re-lands @9203 and fixes the issue the integration test uncovered by always allowing to hide the keyboard.
The difference from the original change is 07d2598
Generally what this PR is doing is setting up a delegation mechanism
for Android's onCreateInputConnection.
It works by letting the framework know when an embedded view gets loses
focus(within the virtual display), the framework maintains a focus node
for each Android view that is kept in sync with the focus state of the
embedded view.
The TextInputPlugin is extended to allow for 2 type of text clients a
"framework client"(what we had before) and a "platform view client".
When the AndroidView's focus node in the framework is focused the
framework sets a "platform view text client" for the TextInputPlugin,
which will result in the TextInputPlugin delegating
createInputConnection to the platform view.
When a platform view is resized, we are detaching it from a virtual
display and attaching it to a new one, as a side affect a platform view
might lose an active input connection, to workaround that we "lock" the
connection when resizing(by caching it and forcing the cached copy until
the resize is done).
Additional things worth calling out in this PR:
To properly answer which views are allowed for input connection
proxying we compare a candidate view's root view to the set of root
views of all virtual displays.
We also preserve a view's focus state across resizes.
Note that this PR only wires text for the io.flutter.view.FlutterView
For the new Android embedding some additional plumbing is necessary.
Corresponding framework PR: flutter/flutter#33901flutter/flutter#19718
Using it, a Flutter app can monitor missing frames in the release mode, and a custom Flutter runner (e.g., Fuchsia) can add a custom FrameRasterizedCallback.
Related issues:
https://github.com/flutter/flutter/issues/26154https://github.com/flutter/flutter/issues/31444https://github.com/flutter/flutter/issues/32447
Need review as soon as possible so we can merge this before the end of May to catch the milestone.
Tests added:
* NoNeedToReportTimingsByDefault
* NeedsReportTimingsIsSetWithCallback
* ReportTimingsIsCalled
* FrameRasterizedCallbackIsCalled
* FrameTimingSetsAndGetsProperly
* onReportTimings preserves callback zone
* FrameTiming.toString has the correct format
This will need a manual engine roll as the TestWindow defined in the framework needs to implement onReportTimings.
This adds the key event source from Android so that the framework can differentiate between keyboard events and game controller events. Also added vendor and product ids so that Flutter code can decide to do special processing based on the kind of device, if desired.
This regression was introduced in #7847.
The PlatformViewsChannel method call handler was always setting the result to `notImplemented` even after handling a result, this resulted in a "Reply already submitted" exception being thrown.
Note that the method channel code is swallowing this exception and logging an error, so we didn't crash instead we were logging an error(this is why the integration test didn't fail).
Filed https://github.com/flutter/flutter/issues/33863 to make sure tests fail when such exceptions are thrown.
This PR also cleans up an unused `NoSuchPlatformViewException` that was introduced in #7847.
https://github.com/flutter/flutter/issues/33866
The podspec must be copied to the build output root, otherwise
--local-engine won't work in projects containing plugins.
Mirrors the iOS podspec copy rule.
FlutterMain.findAppBundlePath is using the presence of the flutter_assets
directory as a signal that the app is initialized. For consistency,
FlutterMain will run the resource extractor at startup and create
flutter_assets even if no resources need to be extracted.
