* Re-create texture from pixel buffer onGrContextCreate
OnGrContextDestroy we destroy the texture, this is because
we can not access it from the potentially new context that
we get on bringing the app back to foreground.
To show a valid texture on fg, we need to preserve the pixel
buffer, using which we will create the new texture.
* ensure texture cache exists
* only reset pixel buffer if not null.
Class.newInstance() propagates any exception thrown by the nullary
constructor, including a checked exception. This effectively bypasses
the compile-time exception checking that would otherwise be performed by
the compiler. The Constructor.newInstance method avoids this problem by
wrapping any exception thrown by the constructor in a (checked)
InvocationTargetException.
If the mapping callback is not set or it the callback returns invalid data, ICU initialization will be embedder responsibility.
This affects all embedders and the following have been audited:
* Android: Via a symbol mapping.
* iOS: Via a file mapping.
* Embedder: Via a file mapping.
* Fuchsia: Via a VMO mapping
* Test shells and Flutter tester: Via file mapping with ICU data needing to be next to the executable.
The GetMapping calls removed in this patch had the same code and had to be repeated across different test harnesses as well as in dart_snapshot.cc. Just make this a factory method so the code is less verbose.
Creates a minimal FLEPluginRegistry protocol, which is a subset of the
FlutterPluginRegistry.
This is a small step toward eventually merging the APIs, but allows
changing the example project structure to better reflect what a future
template will look like.
This is the first step in making Flutter aware of and responsive to Z
bound overflow. On its own this patch shouldn't result in any changes in
behavior. This will need to be followed up with a patch in Fuchsia's
Flutter runner to set the Z bounds after this lands, and another patch
in the engine actually implementing the desired overflow behavior.
This Z bound info is routed through the engine itself to make sure the
bounds in flow are truly consistent from the Fuchsia runner. However
these values should eventually be exposed to the framework as well.
Adds StandardMethodCodec support to the C++ client wrapper. This makes it
substantially easier to add Windows and Linux support for existing plugins, as
StandardMethodCodec is the default plugin protocol.
Fixesflutter/flutter#30670
Does not include extensibility for the codec, which will be added later.
Adds a type that can hold any of the types corresponding to the Dart types
that are supported by the standard message channel codec. This provides
the foundation for adding standard message codec support for the C++
desktop shells (flutter/flutter#30670).
Adds a guard on `_activeClientID` in `insertNewline:`. The conditional
around the `insertText:replacementRange:` call already catches this, but
we then unconditionally pack `_activeClientID` into an `NSArray`, which
disallows nil.
FlutterStandardReader's `readValueOfType:` method returns nil when
called with `FlutterStandardFieldNil`. By extension, `readValue` can
also return nil values.
Since the GLFW shell is likely to continue to exist in at least the
medium term, it is useful for maintenance to be able to build and use it
on any desktop platform.
Includes a buildroot roll to pick up the macOS build suppport for
third_party/glfw.
When flutter/synchronization was first authored, we did not own fml (it was called fxl then). Now we do, so use a single spot for such utilities. The pipeline was meant to be a general purpose utility that was only ever used by the animator (it even has animator specific tracing), so move that to shell instead (where the animator resides).
Some components in the Flutter engine were derived from the forked blink codebase. While the forked components have either been removed or rewritten, the use of the blink namespace has mostly (and inconsistently) remained. This renames the blink namesapce to flutter for consistency. There are no functional changes in this patch.
When running on Fuchsia, "fml/trace_event.h" sends trace events towards
Fuchsia's tracing system, rather than its own |fml::tracing::*| one. Do
the same thing for |FlutterEventTracer|, in order to unlock category
"skia" trace events when running on Fuchsia.
The reason this was not working before is because |FlutterEventTracer|
was directly using |fml::tracing::*| functions, rather than the
TRACE_EVENT_* macros.
In the future, we can look into having the skia event tracer enabled by
default on Fuchsia, controlling the presence of Skia events by
enabling/disabling the "skia" category, once we measure the performance
cost of doing so.
