* Create FlutterEngine to manage a Shell to support maintaining execution state across instances of FlutterViewControllers
* Refactor PlatformViewIOS to support adding or removing a FlutterViewController dynamically
* Refactor FlutterDartHeadlessCodeRunner to implement FlutterEngine
* Refactor FlutterViewController to accept a FlutterEngine at initialization or to create one for backwards compatibility; any Shell related responsibilities are now passed through to the Engine instance
* Remove FlutterNavigationController (unused)
* Update all public Objective C doc comments to be consistent and formatable
* Add public documentation to all public headers
Adds a FlutterPlatformViewFactory protocol - a simple factory protocol to be implemented by plugins
that exposes a UIView for embedding in Flutter apps.
* Adds a FlutterPlatformView protocol, which is used to associate a
dispose callback with a `UIView` created by a FlutterPlatformViewFactory.
* Exposes a registerViewFactory method in FlutterPluginRegistrar.
* Implements the `flutter/platform_views` system channel on iOS, allowing
Dart code to ask for creation/destruction of UIViews.
TL;DR: Offscreen surface is created on the render thread and device to host
transfer performed there before task completion on the UI thread.
While attempting to snapshot layer trees, the engine was attempting to use the
IO thread context. The reasoning was that this would be safe to do because any
textures uploaded to the GPU as a result of async texture upload would have
originated from this context and hence the handles would be valid in either
context. As it turns out, while the handles are valid, Skia does not support
this use-case because cross-context images transfer ownership of the image from
one context to another. So, when we made the hop from the UI thread to the IO
thread (for snapshotting), if either the UI or GPU threads released the last
reference to the texture backed image, the image would be invalid. This led to
such images being absent from the layer tree snapshot.
Simply referencing the images as they are being used on the IO thread is not
sufficient because accessing images on one context after their ownership has
already been transferred to another is not safe behavior (from Skia's
perspective, the handles are still valid in the sharegroup).
To work around these issues, it was decided that an offscreen render target
would be created on the render thread. The color attachment of this render
target could then be transferred as a cross context image to the IO thread for
the device to host tranfer.
Again, this is currently not quite possible because the only way to create
cross context images is from encoded data. Till Skia exposes the functionality
to create cross-context images from textures in one context, we do a device to
host transfer on the GPU thread. The side effect of this is that this is now
part of the frame workload (image compression, which dominate the wall time,
is still done of the IO thread).
A minor side effect of this patch is that the GPU latch needs to be waited on
before the UI thread tasks can be completed before shell initialization.
According to the iOS docs, implementing
`- (id)insertDictationResultPlaceholder`
```Implementation of this method is optional but can be done when you want to provide a specific rectangle for the placeholder animation while the dictation results are being processed. ```
If you do not implement this method, UIKit will insert a default placeholder of 10 whitespace characters. By overriding this, no placeholder text will be inserted. If you implement the `insertDictationResultPlaceholder`, you must implement
`- (void)removeDictationResultPlaceholder:(id)placeholder willInsertResult:(BOOL)willInsertResult`
Add a `-[FlutterViewController setFlutterViewDidRenderCallback:]`
method on iOS so client applications can be notified when the Flutter
view has rendered. This can be used for add2app cases to determine
when to take an initial screenshot for comparisons in automated
tests.
The callback is expected to be an Objective-C block (or Swift
closure). I chose to support only a single callback because it's
much simpler (especially since it does not require a separate method
to unregister the callback), and it's not clear that there are use
cases that would justify additional complexity. Clients have the
flexibility to make their callback invoke other callbacks anyway.
I alternatively considered adding a `-[FlutterViewController
viewDidRenderFirstFlutterFrame]` method that clients could override
in a subclass, but using an Objective-C block seems more flexible and
less of a burden.
Fixes https://github.com/flutter/flutter/issues/20665
The touch mapper was a relic of a time when the pointer IDs had to be less
than 16. Not respecting the touch phase is getting in the way of clients that
fake their own touches. Turns out the AppDelegate also like to fake touches to
simulate status bar events. Now, except in cases where there is a specific
override in place, the UI touch phase is respected.
Also adds //flutter/benchmarking which, similar to //flutter/testing, allows for the creation of a benchmarking executable. This is also the target that contains benchmarking utilities.
Use a screen height ratio to determine if the keyboard is onscreen or not. This is used because Android does not provide a deterministic API to detect this.
We remove the padding when the keyboard is closed and the inset is due to the hidden navigation bar, otherwise, we apply the full keyboard inset.
